Benchmarks preparing for a future article by Michael Larabel.
LAMMPS Molecular Dynamics Simulator LAMMPS is a classical molecular dynamics code, and an acronym for Large-scale Atomic/Molecular Massively Parallel Simulator. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ns/day, More Is Better LAMMPS Molecular Dynamics Simulator 23Jun2022 Model: 20k Atoms Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.177, N = 9 SE +/- 0.120, N = 9 SE +/- 0.174, N = 9 SE +/- 0.342, N = 7 SE +/- 0.297, N = 9 7.579 8.031 8.215 14.471 13.691 1. (CXX) g++ options: -O3 -lm -ldl
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 512 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 22.98 30.41 30.70 39.73 41.02
OpenRadioss OpenRadioss is an open-source AGPL-licensed finite element solver for dynamic event analysis OpenRadioss is based on Altair Radioss and open-sourced in 2022. This open-source finite element solver is benchmarked with various example models available from https://www.openradioss.org/models/ and https://github.com/OpenRadioss/ModelExchange/tree/main/Examples. This test is currently using a reference OpenRadioss binary build offered via GitHub. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: Chrysler Neon 1M Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 17.37, N = 7 SE +/- 2.88, N = 3 SE +/- 10.68, N = 9 SE +/- 5.48, N = 3 SE +/- 4.49, N = 3 983.54 1060.53 1003.59 915.82 891.84
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: INIVOL and Fluid Structure Interaction Drop Container Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 130 260 390 520 650 SE +/- 8.95, N = 9 SE +/- 10.02, N = 7 SE +/- 5.87, N = 9 SE +/- 4.48, N = 9 SE +/- 2.52, N = 3 538.42 582.32 585.75 349.91 322.67
Timed Linux Kernel Compilation This test times how long it takes to build the Linux kernel in a default configuration (defconfig) for the architecture being tested or alternatively an allmodconfig for building all possible kernel modules for the build. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Linux Kernel Compilation 6.8 Build: allmodconfig Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 0.63, N = 3 SE +/- 0.19, N = 3 SE +/- 0.32, N = 3 SE +/- 0.17, N = 3 SE +/- 0.38, N = 3 1508.04 1312.35 1310.96 717.71 714.38
OpenFOAM OpenFOAM is the leading free, open-source software for computational fluid dynamics (CFD). This test profile currently uses the drivaerFastback test case for analyzing automotive aerodynamics or alternatively the older motorBike input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenFOAM 10 Input: drivaerFastback, Medium Mesh Size - Execution Time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 2260.99 2926.98 2827.88 2445.68 2351.51 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfoamToVTK -ldynamicMesh -llagrangian -lgenericPatchFields -lfileFormats -lOpenFOAM -ldl -lm
OpenBenchmarking.org Seconds, Fewer Is Better OpenFOAM 10 Input: drivaerFastback, Medium Mesh Size - Mesh Time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 60 120 180 240 300 225.95 276.71 296.21 245.10 226.01 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfoamToVTK -ldynamicMesh -llagrangian -lgenericPatchFields -lfileFormats -lOpenFOAM -ldl -lm
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 256 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 23.07 30.33 30.63 39.37 40.68
Blender Blender is an open-source 3D creation and modeling software project. This test is of Blender's Cycles performance with various sample files. GPU computing via NVIDIA OptiX and NVIDIA CUDA is currently supported as well as HIP for AMD Radeon GPUs and Intel oneAPI for Intel Graphics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.1 Blend File: Barbershop - Compute: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 4.06, N = 3 SE +/- 0.87, N = 3 SE +/- 0.87, N = 3 SE +/- 0.08, N = 3 SE +/- 0.46, N = 3 1281.54 928.10 925.38 484.46 481.15
OpenRadioss OpenRadioss is an open-source AGPL-licensed finite element solver for dynamic event analysis OpenRadioss is based on Altair Radioss and open-sourced in 2022. This open-source finite element solver is benchmarked with various example models available from https://www.openradioss.org/models/ and https://github.com/OpenRadioss/ModelExchange/tree/main/Examples. This test is currently using a reference OpenRadioss binary build offered via GitHub. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: Bird Strike on Windshield Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 60 120 180 240 300 SE +/- 3.59, N = 9 SE +/- 6.27, N = 9 SE +/- 7.66, N = 9 SE +/- 2.98, N = 12 SE +/- 2.93, N = 12 259.88 276.24 266.65 177.58 183.64
Timed Node.js Compilation This test profile times how long it takes to build/compile Node.js itself from source. Node.js is a JavaScript run-time built from the Chrome V8 JavaScript engine while itself is written in C/C++. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Node.js Compilation 21.7.2 Time To Compile Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 1.08, N = 3 SE +/- 0.37, N = 3 SE +/- 0.38, N = 3 SE +/- 0.94, N = 3 SE +/- 0.68, N = 3 915.94 737.29 739.34 427.64 421.43
OpenVKL OpenVKL is the Intel Open Volume Kernel Library that offers high-performance volume computation kernels and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Items / Sec, More Is Better OpenVKL 2.0.0 Benchmark: vklBenchmarkCPU ISPC Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 130 260 390 520 650 SE +/- 0.33, N = 3 SE +/- 2.52, N = 3 SE +/- 0.33, N = 3 SE +/- 0.58, N = 3 SE +/- 0.67, N = 3 288 327 331 596 603 MIN: 20 / MAX: 4338 MIN: 25 / MAX: 4775 MIN: 25 / MAX: 4820 MIN: 47 / MAX: 8430 MIN: 48 / MAX: 8450
OpenRadioss OpenRadioss is an open-source AGPL-licensed finite element solver for dynamic event analysis OpenRadioss is based on Altair Radioss and open-sourced in 2022. This open-source finite element solver is benchmarked with various example models available from https://www.openradioss.org/models/ and https://github.com/OpenRadioss/ModelExchange/tree/main/Examples. This test is currently using a reference OpenRadioss binary build offered via GitHub. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: Bumper Beam Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 2.49, N = 12 SE +/- 2.19, N = 12 SE +/- 3.91, N = 12 SE +/- 1.80, N = 15 SE +/- 1.49, N = 15 153.90 140.26 145.32 107.17 108.64
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 512 - Model: GoogLeNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 76.13 91.07 91.90 129.06 132.62
Timed LLVM Compilation This test times how long it takes to compile/build the LLVM compiler stack. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed LLVM Compilation 16.0 Build System: Unix Makefiles Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 160 320 480 640 800 SE +/- 0.50, N = 3 SE +/- 0.98, N = 3 SE +/- 0.38, N = 3 SE +/- 1.22, N = 3 SE +/- 1.00, N = 3 719.39 670.27 662.52 395.19 387.01
GPAW GPAW is a density-functional theory (DFT) Python code based on the projector-augmented wave (PAW) method and the atomic simulation environment (ASE). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better GPAW 23.6 Input: Carbon Nanotube Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 8.24, N = 12 SE +/- 4.83, N = 12 SE +/- 2.92, N = 3 SE +/- 2.33, N = 12 SE +/- 1.88, N = 3 228.90 217.50 205.87 166.51 154.99 1. (CC) gcc options: -shared -fwrapv -O2 -lxc -lblas -lmpi
Xmrig Xmrig is an open-source cross-platform CPU/GPU miner for RandomX, KawPow, CryptoNight and AstroBWT. This test profile is setup to measure the Xmrig CPU mining performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org H/s, More Is Better Xmrig 6.21 Variant: GhostRider - Hash Count: 1M Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 700 1400 2100 2800 3500 SE +/- 3.35, N = 3 SE +/- 5.04, N = 3 SE +/- 10.83, N = 3 SE +/- 4.40, N = 3 SE +/- 22.46, N = 3 1742.9 1690.9 1699.3 3239.4 3266.9 1. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
OpenRadioss OpenRadioss is an open-source AGPL-licensed finite element solver for dynamic event analysis OpenRadioss is based on Altair Radioss and open-sourced in 2022. This open-source finite element solver is benchmarked with various example models available from https://www.openradioss.org/models/ and https://github.com/OpenRadioss/ModelExchange/tree/main/Examples. This test is currently using a reference OpenRadioss binary build offered via GitHub. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: Rubber O-Ring Seal Installation Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 5.45, N = 12 SE +/- 9.50, N = 9 SE +/- 8.68, N = 9 SE +/- 3.52, N = 15 SE +/- 1.00, N = 3 177.74 175.46 162.03 101.62 110.05
Llama.cpp Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Tokens Per Second, More Is Better Llama.cpp b1808 Model: llama-2-70b-chat.Q5_0.gguf Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.306 0.612 0.918 1.224 1.53 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 1.36 1.09 1.06 1.16 1.21 1. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OpenRadioss OpenRadioss is an open-source AGPL-licensed finite element solver for dynamic event analysis OpenRadioss is based on Altair Radioss and open-sourced in 2022. This open-source finite element solver is benchmarked with various example models available from https://www.openradioss.org/models/ and https://github.com/OpenRadioss/ModelExchange/tree/main/Examples. This test is currently using a reference OpenRadioss binary build offered via GitHub. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: Cell Phone Drop Test Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 3.30, N = 15 SE +/- 3.05, N = 12 SE +/- 3.06, N = 15 SE +/- 1.27, N = 15 SE +/- 1.68, N = 12 93.26 100.19 95.48 68.68 67.71
Timed Gem5 Compilation This test times how long it takes to compile Gem5. Gem5 is a simulator for computer system architecture research. Gem5 is widely used for computer architecture research within the industry, academia, and more. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Gem5 Compilation 23.0.1 Time To Compile Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 100 200 300 400 500 SE +/- 0.89, N = 3 SE +/- 0.18, N = 3 SE +/- 0.34, N = 3 SE +/- 0.46, N = 3 SE +/- 0.80, N = 3 468.38 436.05 432.22 277.73 273.49
High Performance Conjugate Gradient HPCG is the High Performance Conjugate Gradient and is a new scientific benchmark from Sandia National Lans focused for super-computer testing with modern real-world workloads compared to HPCC. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org GFLOP/s, More Is Better High Performance Conjugate Gradient 3.1 X Y Z: 104 104 104 - RT: 60 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.12782, N = 12 SE +/- 0.06329, N = 12 SE +/- 0.04118, N = 3 SE +/- 0.01035, N = 3 SE +/- 0.05325, N = 3 7.58350 6.67485 6.73761 6.94846 7.22104 1. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -lmpi_cxx -lmpi
Helsing Helsing is an open-source POSIX vampire number generator. This test profile measures the time it takes to generate vampire numbers between varying numbers of digits. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Helsing 1.0-beta Digit Range: 14 digit Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 0.23, N = 3 SE +/- 0.35, N = 3 SE +/- 0.18, N = 3 SE +/- 0.45, N = 3 SE +/- 0.37, N = 3 437.41 372.10 375.92 199.83 200.77 1. (CC) gcc options: -O2 -pthread
Stockfish This is a test of Stockfish, an advanced open-source C++11 chess benchmark that can scale up to 1024 CPU threads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Nodes Per Second, More Is Better Stockfish 16.1 Chess Benchmark Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9M 18M 27M 36M 45M SE +/- 121681.53, N = 15 SE +/- 150504.97, N = 15 SE +/- 160295.89, N = 3 SE +/- 511472.18, N = 4 SE +/- 574948.55, N = 15 16898291 22631087 23143858 42791247 44270858 -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl 1. (CXX) g++ options: -lgcov -m64 -lpthread -fno-exceptions -std=c++17 -fno-peel-loops -fno-tracer -pedantic -O3 -funroll-loops -msse -msse3 -mpopcnt -mavx2 -mbmi -msse4.1 -mssse3 -msse2 -mbmi2 -flto -flto-partition=one -flto=jobserver
Apache IoTDB Apache IotDB is a time series database and this benchmark is facilitated using the IoT Benchmaark [https://github.com/thulab/iot-benchmark/]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 110 220 330 440 550 SE +/- 1.92, N = 3 SE +/- 4.57, N = 3 SE +/- 4.56, N = 3 SE +/- 7.45, N = 3 SE +/- 4.31, N = 3 524.13 460.21 462.14 415.04 399.56 MAX: 26588.63 MAX: 27836.55 MAX: 27920.53 MAX: 27057.75 MAX: 27980.82
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16M 32M 48M 64M 80M SE +/- 207001.08, N = 3 SE +/- 766998.72, N = 3 SE +/- 441409.11, N = 3 SE +/- 669845.56, N = 3 SE +/- 597457.28, N = 3 56494499 64354682 64182563 71135797 73716025
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org verify/s, More Is Better OpenSSL 3.3 Algorithm: RSA4096 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80K 160K 240K 320K 400K SE +/- 1626.68, N = 3 SE +/- 1102.39, N = 3 SE +/- 20.61, N = 3 SE +/- 28.66, N = 3 SE +/- 143.01, N = 3 195618.8 193565.9 194210.3 372775.0 373632.3 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org sign/s, More Is Better OpenSSL 3.3 Algorithm: RSA4096 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3K 6K 9K 12K 15K SE +/- 28.09, N = 3 SE +/- 6.93, N = 3 SE +/- 2.36, N = 3 SE +/- 10.46, N = 3 SE +/- 5.90, N = 3 3000.2 7692.3 7697.3 14222.8 14289.7 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
Apache IoTDB Apache IotDB is a time series database and this benchmark is facilitated using the IoT Benchmaark [https://github.com/thulab/iot-benchmark/]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.51, N = 3 SE +/- 0.94, N = 3 SE +/- 0.48, N = 3 SE +/- 0.16, N = 3 SE +/- 0.21, N = 3 137.92 123.74 122.55 106.72 103.37 MAX: 23917.2 MAX: 23948.92 MAX: 23938.16 MAX: 23939.91 MAX: 23852.59
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16M 32M 48M 64M 80M SE +/- 209117.79, N = 3 SE +/- 503428.59, N = 3 SE +/- 268932.10, N = 3 SE +/- 109473.22, N = 3 SE +/- 205604.00, N = 3 56035278 62420035 62951458 71922519 74075244
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 256 - Model: GoogLeNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.05, N = 3 SE +/- 0.00, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.08, N = 3 76.84 90.98 91.77 128.54 132.13
MariaDB This is a MariaDB MySQL database server benchmark making use of sysbench rather than the existing pts/mysqlslap test profile that uses MariaDB with mysqlslap/mariadb-slap as the benchmark driver. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_read_write - Threads: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40K 80K 120K 160K 200K SE +/- 554.89, N = 3 SE +/- 199.70, N = 3 SE +/- 73.28, N = 3 SE +/- 135.64, N = 3 SE +/- 158.42, N = 3 180020 159448 158499 184015 181519 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
SecureMark SecureMark is an objective, standardized benchmarking framework for measuring the efficiency of cryptographic processing solutions developed by EEMBC. SecureMark-TLS is benchmarking Transport Layer Security performance with a focus on IoT/edge computing. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org marks, More Is Better SecureMark 1.0.4 Benchmark: SecureMark-TLS Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90K 180K 270K 360K 450K SE +/- 739.32, N = 3 SE +/- 1137.99, N = 3 SE +/- 1282.82, N = 3 SE +/- 1244.71, N = 3 SE +/- 1954.24, N = 3 426706 437475 430421 435859 429779 1. (CC) gcc options: -pedantic -O3
MariaDB This is a MariaDB MySQL database server benchmark making use of sysbench rather than the existing pts/mysqlslap test profile that uses MariaDB with mysqlslap/mariadb-slap as the benchmark driver. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_read_write - Threads: 64 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40K 80K 120K 160K 200K SE +/- 562.40, N = 3 SE +/- 60.79, N = 3 SE +/- 211.83, N = 3 SE +/- 108.10, N = 3 SE +/- 389.87, N = 3 186418 168388 168593 201579 201182 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_write_only - Threads: 64 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70K 140K 210K 280K 350K SE +/- 798.38, N = 3 SE +/- 526.67, N = 3 SE +/- 419.19, N = 3 SE +/- 559.43, N = 3 SE +/- 56.85, N = 3 296034 265093 265119 349473 348099 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_update_non_index - Threads: 64 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50K 100K 150K 200K 250K SE +/- 516.32, N = 3 SE +/- 410.68, N = 3 SE +/- 295.88, N = 3 SE +/- 437.08, N = 3 SE +/- 290.99, N = 3 220068 200246 201401 234051 233313 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_update_index - Threads: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50K 100K 150K 200K 250K SE +/- 290.25, N = 3 SE +/- 249.27, N = 3 SE +/- 331.57, N = 3 SE +/- 113.22, N = 3 SE +/- 307.26, N = 3 210974 198305 198779 206240 207212 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_update_index - Threads: 64 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40K 80K 120K 160K 200K SE +/- 250.58, N = 3 SE +/- 125.38, N = 3 SE +/- 308.54, N = 3 SE +/- 130.01, N = 3 SE +/- 282.52, N = 3 187089 177842 177579 200156 198538 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_write_only - Threads: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80K 160K 240K 320K 400K SE +/- 574.53, N = 3 SE +/- 133.55, N = 3 SE +/- 471.57, N = 3 SE +/- 101.24, N = 3 SE +/- 441.31, N = 3 304169 310088 306926 352931 346114 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 11.5 Test: oltp_update_non_index - Threads: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50K 100K 150K 200K 250K SE +/- 223.46, N = 3 SE +/- 181.91, N = 3 SE +/- 399.51, N = 3 SE +/- 466.56, N = 3 SE +/- 861.02, N = 3 253110 224636 224118 239542 239192 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
Timed Godot Game Engine Compilation This test times how long it takes to compile the Godot Game Engine. Godot is a popular, open-source, cross-platform 2D/3D game engine and is built using the SCons build system and targeting the X11 platform. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Godot Game Engine Compilation 4.0 Time To Compile Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80 160 240 320 400 SE +/- 0.34, N = 3 SE +/- 0.20, N = 3 SE +/- 0.15, N = 3 SE +/- 0.56, N = 3 SE +/- 0.40, N = 3 364.42 329.18 326.16 191.08 188.53
Blender Blender is an open-source 3D creation and modeling software project. This test is of Blender's Cycles performance with various sample files. GPU computing via NVIDIA OptiX and NVIDIA CUDA is currently supported as well as HIP for AMD Radeon GPUs and Intel oneAPI for Intel Graphics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.1 Blend File: Pabellon Barcelona - Compute: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 1.05, N = 3 SE +/- 0.09, N = 3 SE +/- 0.40, N = 3 SE +/- 0.09, N = 3 SE +/- 0.18, N = 3 411.15 316.02 315.13 166.82 165.35
ClickHouse ClickHouse is an open-source, high performance OLAP data management system. This test profile uses ClickHouse's standard benchmark recommendations per https://clickhouse.com/docs/en/operations/performance-test/ / https://github.com/ClickHouse/ClickBench/tree/main/clickhouse with the 100 million rows web analytics dataset. The reported value is the query processing time using the geometric mean of all separate queries performed as an aggregate. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Queries Per Minute, Geo Mean, More Is Better ClickHouse 22.12.3.5 100M Rows Hits Dataset, Third Run Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 60 120 180 240 300 SE +/- 1.74, N = 3 SE +/- 1.72, N = 3 SE +/- 4.93, N = 3 SE +/- 1.20, N = 3 SE +/- 2.29, N = 5 249.90 228.79 232.47 241.90 251.80 MIN: 9.25 / MAX: 12000 MIN: 10.54 / MAX: 10000 MIN: 10.67 / MAX: 12000 MIN: 15.08 / MAX: 7500 MIN: 15.75 / MAX: 10000
OpenBenchmarking.org Queries Per Minute, Geo Mean, More Is Better ClickHouse 22.12.3.5 100M Rows Hits Dataset, Second Run Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 3.87, N = 3 SE +/- 1.92, N = 3 SE +/- 1.00, N = 3 SE +/- 1.60, N = 3 SE +/- 1.51, N = 5 249.50 226.35 236.47 240.37 248.89 MIN: 9.27 / MAX: 12000 MIN: 10.69 / MAX: 10000 MIN: 10.66 / MAX: 10000 MIN: 15.16 / MAX: 10000 MIN: 15.92 / MAX: 8571.43
OpenBenchmarking.org Queries Per Minute, Geo Mean, More Is Better ClickHouse 22.12.3.5 100M Rows Hits Dataset, First Run / Cold Cache Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 1.98, N = 3 SE +/- 2.84, N = 3 SE +/- 1.84, N = 3 SE +/- 2.08, N = 3 SE +/- 2.37, N = 5 234.39 211.59 222.47 222.80 227.39 MIN: 9.14 / MAX: 12000 MIN: 10.43 / MAX: 10000 MIN: 10.49 / MAX: 12000 MIN: 12.9 / MAX: 7500 MIN: 13.31 / MAX: 10000
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: LU.C Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10K 20K 30K 40K 50K SE +/- 714.33, N = 15 SE +/- 662.65, N = 12 SE +/- 613.23, N = 15 SE +/- 545.16, N = 15 SE +/- 521.71, N = 15 40491.56 37048.80 37149.88 46710.34 47556.69 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 64 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 23.10 30.22 30.53 39.07 40.24
Blender Blender is an open-source 3D creation and modeling software project. This test is of Blender's Cycles performance with various sample files. GPU computing via NVIDIA OptiX and NVIDIA CUDA is currently supported as well as HIP for AMD Radeon GPUs and Intel oneAPI for Intel Graphics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.1 Blend File: Classroom - Compute: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 1.57, N = 3 SE +/- 0.38, N = 3 SE +/- 0.07, N = 3 SE +/- 0.14, N = 3 SE +/- 0.07, N = 3 336.94 259.85 259.46 136.92 136.11
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 512 - Model: AlexNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 0.02, N = 3 SE +/- 0.11, N = 3 SE +/- 0.55, N = 3 SE +/- 0.05, N = 3 SE +/- 1.74, N = 3 192.28 240.74 240.23 422.88 425.47
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: fcn-resnet101-11 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 11.74, N = 14 SE +/- 8.12, N = 15 SE +/- 12.91, N = 15 SE +/- 4.73, N = 3 SE +/- 1.90, N = 3 960.86 889.17 864.56 567.02 567.07 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: fcn-resnet101-11 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.3969 0.7938 1.1907 1.5876 1.9845 SE +/- 0.013081, N = 14 SE +/- 0.010436, N = 15 SE +/- 0.018640, N = 15 SE +/- 0.014643, N = 3 SE +/- 0.005924, N = 3 1.042812 1.125970 1.160540 1.763840 1.763490 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: fcn-resnet101-11 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 11.74, N = 14 SE +/- 8.12, N = 15 SE +/- 12.91, N = 15 SE +/- 4.73, N = 3 SE +/- 1.90, N = 3 960.86 889.17 864.56 567.02 567.07 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.0039 2.0078 3.0117 4.0156 5.0195 SE +/- 0.03583, N = 3 SE +/- 0.02531, N = 15 SE +/- 0.00310, N = 3 SE +/- 0.03044, N = 15 SE +/- 0.05300, N = 15 4.46170 3.05542 3.09150 2.49599 2.38299 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 1.81, N = 3 SE +/- 2.89, N = 15 SE +/- 0.32, N = 3 SE +/- 4.78, N = 15 SE +/- 9.29, N = 15 224.11 327.53 323.38 401.35 422.42 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.0039 2.0078 3.0117 4.0156 5.0195 SE +/- 0.03583, N = 3 SE +/- 0.02531, N = 15 SE +/- 0.00310, N = 3 SE +/- 0.03044, N = 15 SE +/- 0.05300, N = 15 4.46170 3.05542 3.09150 2.49599 2.38299 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
miniBUDE MiniBUDE is a mini application for the the core computation of the Bristol University Docking Engine (BUDE). This test profile currently makes use of the OpenMP implementation of miniBUDE for CPU benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Billion Interactions/s, More Is Better miniBUDE 20210901 Implementation: OpenMP - Input Deck: BM2 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.06, N = 3 12.55 26.49 26.36 49.92 49.99 1. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm
OpenBenchmarking.org GFInst/s, More Is Better miniBUDE 20210901 Implementation: OpenMP - Input Deck: BM2 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 0.15, N = 3 SE +/- 0.43, N = 3 SE +/- 0.13, N = 3 SE +/- 0.90, N = 3 SE +/- 1.54, N = 3 313.74 662.34 658.93 1248.08 1249.62 1. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm
BRL-CAD BRL-CAD is a cross-platform, open-source solid modeling system with built-in benchmark mode. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org VGR Performance Metric, More Is Better BRL-CAD 7.38.2 VGR Performance Metric Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90K 180K 270K 360K 450K 194127 229365 226198 394575 397138 1. (CXX) g++ options: -std=c++17 -pipe -fvisibility=hidden -fno-strict-aliasing -fno-common -fexceptions -ftemplate-depth-128 -m64 -ggdb3 -O3 -fipa-pta -fstrength-reduce -finline-functions -flto -ltcl8.6 -lnetpbm -lregex_brl -lz_brl -lassimp -ldl -lm -ltk8.6
Apache IoTDB Apache IotDB is a time series database and this benchmark is facilitated using the IoT Benchmaark [https://github.com/thulab/iot-benchmark/]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 60 120 180 240 300 SE +/- 4.53, N = 3 SE +/- 1.83, N = 3 SE +/- 1.82, N = 3 SE +/- 2.23, N = 3 SE +/- 2.35, N = 3 290.14 267.99 268.42 238.34 235.16 MAX: 26535.45 MAX: 27513.85 MAX: 27440.82 MAX: 26718.64 MAX: 26882.16
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 17M 34M 51M 68M 85M SE +/- 787042.09, N = 3 SE +/- 239601.11, N = 3 SE +/- 261270.67, N = 3 SE +/- 362879.93, N = 3 SE +/- 360981.97, N = 3 62655549 68897602 69144164 75893047 77658914
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.20, N = 3 SE +/- 0.06, N = 3 SE +/- 0.29, N = 3 SE +/- 0.11, N = 3 SE +/- 0.29, N = 3 77.24 69.05 68.16 62.09 60.66 MAX: 23891.04 MAX: 23877.52 MAX: 23889.9 MAX: 23874.44 MAX: 23861.25
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 800 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 17M 34M 51M 68M 85M SE +/- 232749.13, N = 3 SE +/- 75302.81, N = 3 SE +/- 296062.71, N = 3 SE +/- 113056.05, N = 3 SE +/- 390613.00, N = 3 61842388 68672756 69561449 75706552 77605627
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80 160 240 320 400 SE +/- 10.20, N = 3 SE +/- 1.80, N = 3 SE +/- 6.12, N = 3 SE +/- 8.45, N = 3 SE +/- 5.40, N = 3 382.48 359.26 347.97 322.24 321.29 MAX: 27376.22 MAX: 27346.14 MAX: 27085.8 MAX: 27055.68 MAX: 27126.33
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20M 40M 60M 80M 100M SE +/- 922667.78, N = 3 SE +/- 357039.39, N = 3 SE +/- 255843.12, N = 3 SE +/- 919852.08, N = 3 SE +/- 818556.17, N = 3 64886364 69856949 71338740 77609975 78454316
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.54, N = 3 SE +/- 0.47, N = 3 SE +/- 0.24, N = 3 SE +/- 0.76, N = 3 SE +/- 0.82, N = 3 119.49 106.44 105.58 95.72 92.08 MAX: 13845.12 MAX: 13870.13 MAX: 15103.61 MAX: 11335.36 MAX: 11356.99
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20M 40M 60M 80M 100M SE +/- 262085.32, N = 3 SE +/- 351738.51, N = 3 SE +/- 282897.65, N = 3 SE +/- 224737.25, N = 3 SE +/- 368630.91, N = 3 62558098 69450850 70305285 76780782 79315319
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: ChaCha20 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30000M 60000M 90000M 120000M 150000M SE +/- 7700475.42, N = 3 SE +/- 85172787.78, N = 3 SE +/- 32786296.71, N = 3 SE +/- 23437434.86, N = 3 SE +/- 13732325.68, N = 3 34651855800 68362354417 68366108810 129320991313 129853526473 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: AES-128-GCM Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50000M 100000M 150000M 200000M 250000M SE +/- 45687236.99, N = 3 SE +/- 53677368.01, N = 3 SE +/- 48206635.82, N = 3 SE +/- 105284888.52, N = 3 SE +/- 197196146.33, N = 3 64812725680 132890512590 133134074673 250497225070 251224402513 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: ChaCha20-Poly1305 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20000M 40000M 60000M 80000M 100000M SE +/- 5622113.45, N = 3 SE +/- 17581304.10, N = 3 SE +/- 11858939.41, N = 3 SE +/- 14328313.13, N = 3 SE +/- 63328398.60, N = 3 23800263020 48632082067 48542739663 91743101110 92129622150 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: AES-256-GCM Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50000M 100000M 150000M 200000M 250000M SE +/- 5611535.27, N = 3 SE +/- 93170961.59, N = 3 SE +/- 101608115.02, N = 3 SE +/- 47970633.04, N = 3 SE +/- 27189836.88, N = 3 56193249190 113843625530 113941258877 214525919597 215336847677 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: SHA512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2000M 4000M 6000M 8000M 10000M SE +/- 2787035.99, N = 3 SE +/- 1105217.67, N = 3 SE +/- 749962.39, N = 3 SE +/- 1453158.01, N = 3 SE +/- 2951460.39, N = 3 4429397160 5813555230 5794405307 11004843967 11041579923 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: SHA256 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7000M 14000M 21000M 28000M 35000M SE +/- 27097230.48, N = 3 SE +/- 18056187.30, N = 3 SE +/- 8102567.82, N = 3 SE +/- 8523911.43, N = 3 SE +/- 7892274.19, N = 3 12336557263 17965660883 17876358337 33924881420 33989652273 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
Xcompact3d Incompact3d Xcompact3d Incompact3d is a Fortran-MPI based, finite difference high-performance code for solving the incompressible Navier-Stokes equation and as many as you need scalar transport equations. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Xcompact3d Incompact3d 2021-03-11 Input: input.i3d 193 Cells Per Direction Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.87, N = 15 SE +/- 0.76, N = 3 SE +/- 0.27, N = 3 SE +/- 1.01, N = 3 SE +/- 0.58, N = 10 73.72 90.69 87.77 79.60 73.88 1. (F9X) gfortran options: -cpp -O2 -funroll-loops -floop-optimize -fcray-pointer -fbacktrace -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 1 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.305 2.61 3.915 5.22 6.525 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 2.26 3.04 3.04 5.74 5.80 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
MariaDB mariadb-slap This is a MariaDB MySQL database server benchmark making use of mysqlslap (mariadb-slap). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB mariadb-slap 11.5 Clients: 128 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 130 260 390 520 650 SE +/- 1.05, N = 3 SE +/- 0.57, N = 3 SE +/- 0.74, N = 3 SE +/- 0.55, N = 3 SE +/- 0.54, N = 3 574 587 580 614 612 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
Pennant Pennant is an application focused on hydrodynamics on general unstructured meshes in 2D. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Hydro Cycle Time - Seconds, Fewer Is Better Pennant 1.0.1 Test: sedovbig Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14 28 42 56 70 SE +/- 0.67, N = 4 SE +/- 0.86, N = 15 SE +/- 0.59, N = 6 SE +/- 0.31, N = 15 SE +/- 0.35, N = 8 55.90 61.28 61.97 44.43 42.23 1. (CXX) g++ options: -fopenmp -lmpi_cxx -lmpi
CloverLeaf CloverLeaf is a Lagrangian-Eulerian hydrodynamics benchmark. This test profile currently makes use of CloverLeaf's OpenMP version. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better CloverLeaf 1.3 Input: clover_bm64_short Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 138.32 177.83 172.36 178.81 171.30 1. (F9X) gfortran options: -O3 -march=native -funroll-loops -fopenmp
Java JMH This very basic test profile runs the stock benchmark of the Java JMH benchmark via Maven. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/s, More Is Better Java JMH Throughput Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16000M 32000M 48000M 64000M 80000M 36765943472.27 38717919571.11 38776376503.47 75186159786.82 75134161573.04
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: particle_volume/scivis/real_time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.00749, N = 3 SE +/- 0.00354, N = 3 SE +/- 0.00347, N = 3 SE +/- 0.00801, N = 3 SE +/- 0.00296, N = 3 5.88640 4.64920 4.61871 8.81722 8.98195
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: yolov4 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 1.78, N = 3 SE +/- 1.34, N = 15 SE +/- 1.14, N = 15 SE +/- 0.47, N = 3 SE +/- 1.04, N = 4 123.59 107.51 107.79 85.63 86.84 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: yolov4 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.11600, N = 3 SE +/- 0.12476, N = 15 SE +/- 0.10193, N = 15 SE +/- 0.06318, N = 3 SE +/- 0.13812, N = 4 8.09421 9.32307 9.29179 11.67770 11.51990 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: yolov4 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 1.78, N = 3 SE +/- 1.34, N = 15 SE +/- 1.14, N = 15 SE +/- 0.47, N = 3 SE +/- 1.04, N = 4 123.59 107.51 107.79 85.63 86.84 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: GPT-2 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.02440, N = 3 SE +/- 0.07551, N = 15 SE +/- 0.01376, N = 3 SE +/- 0.09699, N = 3 SE +/- 0.08068, N = 15 7.89080 7.81820 8.06389 7.60699 7.51744 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: GPT-2 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.39, N = 3 SE +/- 1.28, N = 15 SE +/- 0.21, N = 3 SE +/- 1.70, N = 3 SE +/- 1.39, N = 15 126.67 128.01 123.95 131.44 133.17 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: GPT-2 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.02440, N = 3 SE +/- 0.07551, N = 15 SE +/- 0.01376, N = 3 SE +/- 0.09699, N = 3 SE +/- 0.08068, N = 15 7.89080 7.81820 8.06389 7.60699 7.51744 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
ScyllaDB This is a benchmark of ScyllaDB and is making use of Apache Cassandra's cassandra-stress for conducting the benchmark. ScyllaDB is an open-source distributed NoSQL data store that is compatible with Apache Cassandra while focusing on higher throughput and lower latency. ScyllaDB uses a sharded design on each node. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better ScyllaDB 5.2.9 Test: Writes Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 395.91, N = 3 SE +/- 446.24, N = 3 SE +/- 60.93, N = 3 SE +/- 78.75, N = 3 SE +/- 255.51, N = 3 104832 94899 94217 151849 151729
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: yolov4 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 1.27, N = 3 SE +/- 5.32, N = 15 SE +/- 5.00, N = 15 SE +/- 0.60, N = 3 SE +/- 0.70, N = 3 127.67 100.03 97.12 96.99 97.94 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: yolov4 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.07886, N = 3 SE +/- 0.63014, N = 15 SE +/- 0.57269, N = 15 SE +/- 0.06449, N = 3 SE +/- 0.07376, N = 3 7.83432 10.46448 10.70547 10.31040 10.21150 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: yolov4 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 1.27, N = 3 SE +/- 5.32, N = 15 SE +/- 5.00, N = 15 SE +/- 0.60, N = 3 SE +/- 0.70, N = 3 127.67 100.03 97.12 96.99 97.94 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Appleseed Appleseed is an open-source production renderer focused on physically-based global illumination rendering engine primarily designed for animation and visual effects. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Appleseed 2.0 Beta Scene: Emily Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 323.27 264.55 263.51 153.37 152.30
PostgreSQL This is a benchmark of PostgreSQL using the integrated pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 500 - Mode: Read Write - Average Latency Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.101, N = 3 SE +/- 0.141, N = 4 SE +/- 0.078, N = 3 SE +/- 0.094, N = 4 SE +/- 0.109, N = 3 11.791 12.454 12.355 8.696 8.850 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 500 - Mode: Read Write Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12K 24K 36K 48K 60K SE +/- 362.37, N = 3 SE +/- 453.73, N = 4 SE +/- 256.18, N = 3 SE +/- 617.95, N = 4 SE +/- 699.92, N = 3 42412 40163 40473 57520 56517 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
Apache IoTDB Apache IotDB is a time series database and this benchmark is facilitated using the IoT Benchmaark [https://github.com/thulab/iot-benchmark/]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 60 120 180 240 300 SE +/- 1.91, N = 3 SE +/- 3.81, N = 3 SE +/- 2.74, N = 3 SE +/- 3.26, N = 3 SE +/- 3.67, N = 3 275.14 248.62 243.61 235.20 230.41 MAX: 27185.57 MAX: 26892.13 MAX: 26975.13 MAX: 26852.2 MAX: 26735.87
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 400 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 15M 30M 45M 60M 75M SE +/- 245819.57, N = 3 SE +/- 382550.66, N = 3 SE +/- 418677.10, N = 3 SE +/- 438509.29, N = 3 SE +/- 489754.87, N = 3 57760197 63296864 64655502 67794681 69731166
LuxCoreRender LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.6 Scene: Danish Mood - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.9405 1.881 2.8215 3.762 4.7025 SE +/- 0.01, N = 15 SE +/- 0.02, N = 3 SE +/- 0.02, N = 11 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 1.45 2.06 2.10 4.14 4.18 MIN: 0.39 / MAX: 1.82 MIN: 0.66 / MAX: 2.48 MIN: 0.64 / MAX: 2.54 MIN: 1.71 / MAX: 4.75 MIN: 1.81 / MAX: 4.75
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: particle_volume/pathtracer/real_time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 0.42, N = 3 SE +/- 0.86, N = 3 SE +/- 0.61, N = 3 SE +/- 0.83, N = 3 SE +/- 0.68, N = 3 157.97 168.64 170.39 219.74 220.91
easyWave The easyWave software allows simulating tsunami generation and propagation in the context of early warning systems. EasyWave supports making use of OpenMP for CPU multi-threading and there are also GPU ports available but not currently incorporated as part of this test profile. The easyWave tsunami generation software is run with one of the example/reference input files for measuring the CPU execution time. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better easyWave r34 Input: e2Asean Grid + BengkuluSept2007 Source - Time: 1200 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.20, N = 3 SE +/- 0.14, N = 3 SE +/- 0.21, N = 3 SE +/- 0.20, N = 3 SE +/- 0.25, N = 3 145.99 159.04 152.79 132.13 127.02 1. (CXX) g++ options: -O3 -fopenmp
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: SP.C Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3K 6K 9K 12K 15K SE +/- 4.20, N = 3 SE +/- 2.24, N = 3 SE +/- 1.44, N = 3 SE +/- 175.17, N = 3 SE +/- 146.76, N = 5 11387.05 9707.08 10027.11 12630.55 13374.35 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
ACES DGEMM This is a multi-threaded DGEMM benchmark. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org GFLOP/s, More Is Better ACES DGEMM 1.0 Sustained Floating-Point Rate Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.054700, N = 3 SE +/- 0.133403, N = 15 SE +/- 0.144416, N = 15 SE +/- 0.166167, N = 15 SE +/- 0.258971, N = 12 4.204110 5.160364 5.330500 9.536711 9.577753 1. (CC) gcc options: -O3 -march=native -fopenmp
Primesieve Primesieve generates prime numbers using a highly optimized sieve of Eratosthenes implementation. Primesieve primarily benchmarks the CPU's L1/L2 cache performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Primesieve 12.1 Length: 1e13 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 0.18, N = 3 SE +/- 0.09, N = 3 SE +/- 0.15, N = 3 SE +/- 0.09, N = 3 SE +/- 0.20, N = 3 245.67 152.68 153.60 82.12 82.47 1. (CXX) g++ options: -O3
Apache IoTDB Apache IotDB is a time series database and this benchmark is facilitated using the IoT Benchmaark [https://github.com/thulab/iot-benchmark/]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Average Latency, Fewer Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.51, N = 3 SE +/- 0.41, N = 3 SE +/- 0.31, N = 3 SE +/- 0.42, N = 3 SE +/- 0.10, N = 3 80.56 72.28 70.70 67.07 65.63 MAX: 15084.51 MAX: 13851.96 MAX: 13822.83 MAX: 12570.5 MAX: 12576.01
OpenBenchmarking.org point/sec, More Is Better Apache IoTDB 1.2 Device Count: 500 - Batch Size Per Write: 100 - Sensor Count: 500 - Client Number: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 15M 30M 45M 60M 75M SE +/- 205982.51, N = 3 SE +/- 161254.73, N = 3 SE +/- 42980.57, N = 3 SE +/- 313014.56, N = 3 SE +/- 41672.18, N = 3 57028398 63627945 64781285 68166248 69857519
ASTC Encoder ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MT/s, More Is Better ASTC Encoder 4.7 Preset: Very Thorough Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.6039 1.2078 1.8117 2.4156 3.0195 SE +/- 0.0116, N = 3 SE +/- 0.0023, N = 3 SE +/- 0.0010, N = 3 SE +/- 0.0070, N = 3 SE +/- 0.0093, N = 3 1.4639 1.4109 1.3979 2.6608 2.6841 1. (CXX) g++ options: -O3 -flto -pthread
OpenBenchmarking.org MT/s, More Is Better ASTC Encoder 4.7 Preset: Exhaustive Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.3688 0.7376 1.1064 1.4752 1.844 SE +/- 0.0051, N = 3 SE +/- 0.0007, N = 3 SE +/- 0.0000, N = 3 SE +/- 0.0018, N = 3 SE +/- 0.0020, N = 3 0.8741 0.8633 0.8557 1.6262 1.6391 1. (CXX) g++ options: -O3 -flto -pthread
PostgreSQL This is a benchmark of PostgreSQL using the integrated pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 800 - Mode: Read Only - Average Latency Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.4673 0.9346 1.4019 1.8692 2.3365 SE +/- 0.005, N = 3 SE +/- 0.030, N = 3 SE +/- 0.023, N = 3 SE +/- 0.009, N = 3 SE +/- 0.006, N = 3 1.852 2.077 2.051 0.709 0.722 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 800 - Mode: Read Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200K 400K 600K 800K 1000K SE +/- 1154.98, N = 3 SE +/- 5557.21, N = 3 SE +/- 4300.20, N = 3 SE +/- 14172.11, N = 3 SE +/- 8757.92, N = 3 431962 385411 390018 1128893 1108243 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 500 - Mode: Read Only - Average Latency Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.2162 0.4324 0.6486 0.8648 1.081 SE +/- 0.003, N = 3 SE +/- 0.010, N = 3 SE +/- 0.012, N = 3 SE +/- 0.001, N = 3 SE +/- 0.001, N = 3 0.899 0.940 0.961 0.360 0.361 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 500 - Mode: Read Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300K 600K 900K 1200K 1500K SE +/- 1617.01, N = 3 SE +/- 5499.26, N = 3 SE +/- 6448.13, N = 3 SE +/- 3056.15, N = 3 SE +/- 3948.07, N = 3 555873 532015 520502 1388157 1384166 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 800 - Mode: Read Write - Average Latency Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.06, N = 3 SE +/- 0.26, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.16, N = 3 23.20 24.76 24.50 16.10 16.16 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 800 - Mode: Read Write Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11K 22K 33K 44K 55K SE +/- 85.58, N = 3 SE +/- 337.25, N = 3 SE +/- 123.75, N = 3 SE +/- 155.76, N = 3 SE +/- 481.81, N = 3 34486 32316 32657 49683 49522 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
Pennant Pennant is an application focused on hydrodynamics on general unstructured meshes in 2D. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Hydro Cycle Time - Seconds, Fewer Is Better Pennant 1.0.1 Test: leblancbig Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 SE +/- 1.09, N = 12 SE +/- 0.82, N = 15 SE +/- 0.88, N = 15 SE +/- 0.33, N = 5 SE +/- 0.27, N = 3 32.87 43.80 42.94 30.24 29.55 1. (CXX) g++ options: -fopenmp -lmpi_cxx -lmpi
DuckDB DuckDB is an in-progress SQL OLAP database management system optimized for analytics and features a vectorized and parallel engine. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better DuckDB 0.9.1 Benchmark: IMDB Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.14, N = 3 SE +/- 0.10, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 99.58 101.94 100.63 83.01 82.38 1. (CXX) g++ options: -O3 -rdynamic -lssl -lcrypto -ldl
FFmpeg This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better FFmpeg 6.1 Encoder: libx265 - Scenario: Platform Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16 32 48 64 80 SE +/- 0.22, N = 3 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 SE +/- 0.13, N = 3 SE +/- 0.10, N = 3 58.75 65.45 65.26 69.24 69.80 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.org FPS, More Is Better FFmpeg 6.1 Encoder: libx265 - Scenario: Video On Demand Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16 32 48 64 80 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.09, N = 3 SE +/- 0.13, N = 3 SE +/- 0.13, N = 3 58.75 65.27 65.35 69.52 69.81 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
Zstd Compression This test measures the time needed to compress/decompress a sample file (silesia.tar) using Zstd (Zstandard) compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 19, Long Mode - Decompression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 400 800 1200 1600 2000 SE +/- 9.70, N = 3 SE +/- 97.47, N = 3 SE +/- 9.44, N = 3 SE +/- 6.90, N = 15 SE +/- 20.05, N = 3 1906.1 1836.4 1926.9 1890.5 1908.3 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 19, Long Mode - Compression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.00, N = 3 SE +/- 0.18, N = 15 SE +/- 0.09, N = 3 8.91 10.30 10.20 11.50 10.70 1. (CC) gcc options: -O3 -pthread -lz -llzma
Llamafile Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Tokens Per Second, More Is Better Llamafile 0.6 Test: mistral-7b-instruct-v0.2.Q8_0 - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 7.91 6.34 6.28 6.53 6.79
PyTorch This is a benchmark of PyTorch making use of pytorch-benchmark [https://github.com/LukasHedegaard/pytorch-benchmark]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 64 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 SE +/- 1.58, N = 15 SE +/- 0.35, N = 3 SE +/- 0.16, N = 3 SE +/- 0.27, N = 3 SE +/- 0.94, N = 15 32.25 45.20 45.11 45.76 43.38 MIN: 23.14 / MAX: 40.93 MIN: 29.05 / MAX: 46.24 MIN: 43.61 / MAX: 45.86 MIN: 34.75 / MAX: 46.5 MIN: 34.95 / MAX: 46.71
FFmpeg This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better FFmpeg 6.1 Encoder: libx265 - Scenario: Upload Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 8 16 24 32 40 SE +/- 0.08, N = 3 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.13, N = 3 SE +/- 0.07, N = 3 28.89 32.03 31.95 34.06 34.21 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: particle_volume/ao/real_time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.00839, N = 3 SE +/- 0.05544, N = 4 SE +/- 0.00129, N = 3 SE +/- 0.00268, N = 3 SE +/- 0.00600, N = 3 5.93161 4.58353 4.63828 8.81078 8.98399
Blender Blender is an open-source 3D creation and modeling software project. This test is of Blender's Cycles performance with various sample files. GPU computing via NVIDIA OptiX and NVIDIA CUDA is currently supported as well as HIP for AMD Radeon GPUs and Intel oneAPI for Intel Graphics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.1 Blend File: Junkshop - Compute: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.02, N = 3 SE +/- 0.19, N = 3 SE +/- 0.43, N = 3 SE +/- 0.01, N = 3 SE +/- 0.17, N = 3 173.17 139.94 139.48 73.73 72.75
Zstd Compression This test measures the time needed to compress/decompress a sample file (silesia.tar) using Zstd (Zstandard) compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 19 - Decompression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 400 800 1200 1600 2000 SE +/- 7.69, N = 3 SE +/- 20.68, N = 3 SE +/- 18.77, N = 3 SE +/- 7.95, N = 15 SE +/- 9.65, N = 3 1972.1 2026.0 2066.7 1998.3 2010.8 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 19 - Compression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.03, N = 3 SE +/- 0.15, N = 3 SE +/- 0.03, N = 3 SE +/- 0.23, N = 15 SE +/- 0.03, N = 3 16.3 17.9 18.0 19.5 18.8 1. (CC) gcc options: -O3 -pthread -lz -llzma
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.1.1 Model: Layered Halfspace Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 1.26, N = 3 SE +/- 0.89, N = 3 SE +/- 0.49, N = 3 SE +/- 0.17, N = 3 SE +/- 0.14, N = 3 144.54 130.53 133.31 86.00 83.97 1. (F9X) gfortran options: -O2 -fopenmp -std=f2008 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: IS.D Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 16.64, N = 3 SE +/- 18.32, N = 12 SE +/- 16.77, N = 15 SE +/- 10.32, N = 15 SE +/- 4.80, N = 3 1604.85 1228.58 1269.37 1223.22 1278.44 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.1.1 Model: Water-layered Halfspace Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.21, N = 3 SE +/- 0.40, N = 3 SE +/- 0.42, N = 3 SE +/- 0.26, N = 3 SE +/- 0.06, N = 3 135.42 129.55 129.33 87.86 85.25 1. (F9X) gfortran options: -O2 -fopenmp -std=f2008 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 32 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.04, N = 3 23.19 30.52 30.79 38.96 39.99
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.1.1 Model: Mount St. Helens Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12 24 36 48 60 SE +/- 0.34, N = 3 SE +/- 0.33, N = 15 SE +/- 0.25, N = 3 SE +/- 0.33, N = 3 SE +/- 0.29, N = 15 53.29 49.54 51.29 34.80 33.13 1. (F9X) gfortran options: -O2 -fopenmp -std=f2008 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
Llamafile Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Tokens Per Second, More Is Better Llamafile 0.6 Test: wizardcoder-python-34b-v1.0.Q6_K - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.495 0.99 1.485 1.98 2.475 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 2.20 1.84 1.82 1.94 2.01
Zstd Compression This test measures the time needed to compress/decompress a sample file (silesia.tar) using Zstd (Zstandard) compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 8 - Decompression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 2.75, N = 3 SE +/- 3.55, N = 3 SE +/- 4.31, N = 3 SE +/- 5.41, N = 9 SE +/- 24.36, N = 9 2375.6 2313.4 2322.3 2279.9 2245.7 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 8 - Compression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 3.15, N = 3 SE +/- 1.37, N = 3 SE +/- 1.88, N = 3 SE +/- 6.44, N = 9 SE +/- 6.27, N = 9 485.6 513.9 520.3 778.5 793.8 1. (CC) gcc options: -O3 -pthread -lz -llzma
Intel Open Image Denoise Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Images / Sec, More Is Better Intel Open Image Denoise 2.2 Run: RTLightmap.hdr.4096x4096 - Device: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.0923 0.1846 0.2769 0.3692 0.4615 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.21 0.23 0.23 0.40 0.41
Blender Blender is an open-source 3D creation and modeling software project. This test is of Blender's Cycles performance with various sample files. GPU computing via NVIDIA OptiX and NVIDIA CUDA is currently supported as well as HIP for AMD Radeon GPUs and Intel oneAPI for Intel Graphics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.1 Blend File: Fishy Cat - Compute: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.24, N = 3 SE +/- 0.38, N = 3 SE +/- 0.39, N = 3 SE +/- 0.07, N = 3 SE +/- 0.06, N = 3 165.00 127.12 127.01 67.34 66.78
NAMD NAMD is a parallel molecular dynamics code designed for high-performance simulation of large biomolecular systems. NAMD was developed by the Theoretical and Computational Biophysics Group in the Beckman Institute for Advanced Science and Technology at the University of Illinois at Urbana-Champaign. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ns/day, More Is Better NAMD 3.0b6 Input: STMV with 1,066,628 Atoms Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.1702 0.3404 0.5106 0.6808 0.851 SE +/- 0.00016, N = 3 SE +/- 0.00121, N = 3 SE +/- 0.00059, N = 3 SE +/- 0.00230, N = 3 SE +/- 0.00230, N = 3 0.20565 0.49007 0.49345 0.74072 0.75646
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.41, N = 15 SE +/- 0.03, N = 3 23.61 19.10 19.26 16.78 18.66 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 13 26 39 52 65 SE +/- 0.04, N = 3 SE +/- 0.20, N = 3 SE +/- 0.12, N = 3 SE +/- 1.44, N = 15 SE +/- 0.09, N = 3 42.35 52.34 51.91 60.06 53.58 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.41, N = 15 SE +/- 0.03, N = 3 23.61 19.10 19.26 16.78 18.66 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: bertsquad-12 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.14, N = 3 SE +/- 0.63, N = 3 SE +/- 4.77, N = 15 SE +/- 0.32, N = 3 SE +/- 0.03, N = 3 91.52 85.12 70.83 65.48 66.03 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: bertsquad-12 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.02, N = 3 SE +/- 0.09, N = 3 SE +/- 1.17, N = 15 SE +/- 0.08, N = 3 SE +/- 0.01, N = 3 10.93 11.75 15.22 15.27 15.14 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: bertsquad-12 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.14, N = 3 SE +/- 0.63, N = 3 SE +/- 4.77, N = 15 SE +/- 0.32, N = 3 SE +/- 0.03, N = 3 91.52 85.12 70.83 65.48 66.03 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: T5 Encoder - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.07186, N = 15 SE +/- 0.04051, N = 3 SE +/- 0.03666, N = 3 SE +/- 0.08219, N = 3 SE +/- 0.08220, N = 3 5.90786 7.03702 7.16751 6.36984 6.16145 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: T5 Encoder - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 2.25, N = 15 SE +/- 0.82, N = 3 SE +/- 0.71, N = 3 SE +/- 2.01, N = 3 SE +/- 2.20, N = 3 169.62 142.06 139.48 157.01 162.33 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: T5 Encoder - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.07186, N = 15 SE +/- 0.04051, N = 3 SE +/- 0.03666, N = 3 SE +/- 0.08219, N = 3 SE +/- 0.08220, N = 3 5.90786 7.03702 7.16751 6.36984 6.16145 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
CockroachDB CockroachDB is a cloud-native, distributed SQL database for data intensive applications. This test profile uses a server-less CockroachDB configuration to test various Coackroach workloads on the local host with a single node. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ops/s, More Is Better CockroachDB 22.2 Workload: KV, 60% Reads - Concurrency: 128 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 198.29, N = 3 SE +/- 204.39, N = 3 SE +/- 47.09, N = 3 SE +/- 137.27, N = 3 SE +/- 156.58, N = 3 63021.4 67444.0 67506.7 93864.1 95084.8
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 256 - Model: AlexNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 0.06, N = 3 SE +/- 0.15, N = 3 SE +/- 0.09, N = 3 SE +/- 0.13, N = 3 SE +/- 0.09, N = 3 189.14 237.76 237.57 405.64 409.96
CockroachDB CockroachDB is a cloud-native, distributed SQL database for data intensive applications. This test profile uses a server-less CockroachDB configuration to test various Coackroach workloads on the local host with a single node. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ops/s, More Is Better CockroachDB 22.2 Workload: KV, 50% Reads - Concurrency: 128 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 189.35, N = 3 SE +/- 103.24, N = 3 SE +/- 264.18, N = 3 SE +/- 764.38, N = 3 SE +/- 831.25, N = 3 59866.6 64458.4 64052.7 90528.3 92195.2
OpenBenchmarking.org ops/s, More Is Better CockroachDB 22.2 Workload: KV, 95% Reads - Concurrency: 128 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 237.27, N = 3 SE +/- 258.97, N = 3 SE +/- 302.28, N = 3 SE +/- 716.09, N = 3 SE +/- 765.52, N = 3 83987.5 84052.3 83155.8 119084.9 121061.7
Dragonflydb Dragonfly is an open-source database server that is a "modern Redis replacement" that aims to be the fastest memory store while being compliant with the Redis and Memcached protocols. For benchmarking Dragonfly, Memtier_benchmark is used as a NoSQL Redis/Memcache traffic generation plus benchmarking tool developed by Redis Labs. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 10 - Set To Get Ratio: 1:10 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.4M 2.8M 4.2M 5.6M 7M SE +/- 44122.15, N = 3 SE +/- 43642.69, N = 3 SE +/- 38431.66, N = 3 SE +/- 45878.41, N = 3 SE +/- 90168.82, N = 12 4522560.58 4759359.71 4729191.29 6406922.86 6266297.87 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
ASKAP ASKAP is a set of benchmarks from the Australian SKA Pathfinder. The principal ASKAP benchmarks are the Hogbom Clean Benchmark (tHogbomClean) and Convolutional Resamping Benchmark (tConvolve) as well as some previous ASKAP benchmarks being included as well for OpenCL and CUDA execution of tConvolve. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mpix/sec, More Is Better ASKAP 1.0 Test: tConvolve MPI - Gridding Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1600 3200 4800 6400 8000 SE +/- 48.43, N = 3 SE +/- 0.00, N = 3 SE +/- 180.65, N = 12 SE +/- 298.76, N = 15 SE +/- 298.02, N = 15 6199.09 3975.69 4256.28 7317.33 7477.19 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
OpenBenchmarking.org Mpix/sec, More Is Better ASKAP 1.0 Test: tConvolve MPI - Degridding Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1400 2800 4200 5600 7000 SE +/- 58.27, N = 3 SE +/- 14.20, N = 3 SE +/- 192.19, N = 12 SE +/- 328.86, N = 15 SE +/- 321.53, N = 15 5146.33 3349.85 3635.69 6613.59 6642.94 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
GROMACS The GROMACS (GROningen MAchine for Chemical Simulations) molecular dynamics package testing with the water_GMX50 data. This test profile allows selecting between CPU and GPU-based GROMACS builds. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2024 Implementation: MPI CPU - Input: water_GMX50_bare Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.5722 1.1444 1.7166 2.2888 2.861 SE +/- 0.000, N = 3 SE +/- 0.001, N = 3 SE +/- 0.002, N = 3 SE +/- 0.004, N = 3 SE +/- 0.001, N = 3 1.434 1.526 1.546 2.468 2.543 1. (CXX) g++ options: -O3 -lm
Zstd Compression This test measures the time needed to compress/decompress a sample file (silesia.tar) using Zstd (Zstandard) compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 12 - Decompression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 2.01, N = 3 SE +/- 7.37, N = 3 SE +/- 10.81, N = 3 SE +/- 11.41, N = 12 SE +/- 25.03, N = 3 2377.0 2394.8 2404.2 2308.6 2314.7 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 12 - Compression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 2.27, N = 3 SE +/- 1.44, N = 3 SE +/- 0.92, N = 3 SE +/- 4.91, N = 12 SE +/- 2.97, N = 3 170.6 168.4 171.0 207.5 220.9 1. (CC) gcc options: -O3 -pthread -lz -llzma
libavif avifenc This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 1.0 Encoder Speed: 0 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.49, N = 3 SE +/- 0.12, N = 3 SE +/- 0.18, N = 3 SE +/- 0.08, N = 3 SE +/- 0.33, N = 3 133.57 107.79 108.65 72.42 72.11 1. (CXX) g++ options: -O3 -fPIC -lm
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: BT.C Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10K 20K 30K 40K 50K SE +/- 96.46, N = 3 SE +/- 21.68, N = 3 SE +/- 14.25, N = 3 SE +/- 553.44, N = 3 SE +/- 110.88, N = 3 28326.36 22551.09 22935.27 43371.78 44844.13 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.11 Video Input: Bosphorus 4K - Video Preset: Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.018, N = 3 SE +/- 0.024, N = 3 SE +/- 0.021, N = 3 SE +/- 0.061, N = 3 SE +/- 0.058, N = 3 4.975 5.681 5.693 7.746 7.818 1. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 SE +/- 0.21, N = 3 SE +/- 0.69, N = 12 SE +/- 0.15, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 43.40 36.25 37.00 27.35 26.83 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.11, N = 3 SE +/- 0.60, N = 12 SE +/- 0.11, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 23.04 27.71 27.02 36.56 37.27 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 SE +/- 0.21, N = 3 SE +/- 0.69, N = 12 SE +/- 0.15, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 43.40 36.25 37.00 27.35 26.83 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Redis Redis is an open-source in-memory data structure store, used as a database, cache, and message broker. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better Redis 7.0.4 Test: SET - Parallel Connections: 500 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 800K 1600K 2400K 3200K 4000K SE +/- 128040.11, N = 15 SE +/- 181123.85, N = 15 SE +/- 201571.91, N = 12 SE +/- 72840.48, N = 12 SE +/- 164667.68, N = 15 3327348.65 3656582.35 3684418.98 3827317.96 3419833.50 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
PETSc PETSc, the Portable, Extensible Toolkit for Scientific Computation, is for the scalable (parallel) solution of scientific applications modeled by partial differential equations. This test profile runs the PETSc "make streams" benchmark and records the throughput rate when all available cores are utilized for the MPI Streams build. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better PETSc 3.19 Test: Streams Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 15K 30K 45K 60K 75K SE +/- 12.95, N = 3 SE +/- 40.69, N = 3 SE +/- 249.54, N = 3 SE +/- 31.25, N = 3 SE +/- 23.59, N = 3 69652.73 39918.97 26545.71 39471.07 41077.63 1. (CC) gcc options: -fPIC -O3 -O2 -lpthread -ludev -lpciaccess -lm
OpenFOAM OpenFOAM is the leading free, open-source software for computational fluid dynamics (CFD). This test profile currently uses the drivaerFastback test case for analyzing automotive aerodynamics or alternatively the older motorBike input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better OpenFOAM 10 Input: drivaerFastback, Small Mesh Size - Execution Time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 242.94 311.36 296.60 199.88 183.23 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfoamToVTK -ldynamicMesh -llagrangian -lgenericPatchFields -lfileFormats -lOpenFOAM -ldl -lm
OpenBenchmarking.org Seconds, Fewer Is Better OpenFOAM 10 Input: drivaerFastback, Small Mesh Size - Mesh Time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 35.19 41.68 38.17 29.74 30.77 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfoamToVTK -ldynamicMesh -llagrangian -lgenericPatchFields -lfileFormats -lOpenFOAM -ldl -lm
Dragonflydb Dragonfly is an open-source database server that is a "modern Redis replacement" that aims to be the fastest memory store while being compliant with the Redis and Memcached protocols. For benchmarking Dragonfly, Memtier_benchmark is used as a NoSQL Redis/Memcache traffic generation plus benchmarking tool developed by Redis Labs. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Dragonflydb 1.6.2 Clients Per Thread: 10 - Set To Get Ratio: 1:100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.5M 3M 4.5M 6M 7.5M SE +/- 43112.16, N = 3 SE +/- 16892.77, N = 3 SE +/- 6458.55, N = 3 SE +/- 57648.71, N = 9 SE +/- 34452.25, N = 3 4604951.36 4818585.25 4762998.42 7046043.06 6289782.08 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
nginx This is a benchmark of the lightweight Nginx HTTP(S) web-server. This Nginx web server benchmark test profile makes use of the wrk program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients/connections. HTTPS with a self-signed OpenSSL certificate is used by this test for local benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better nginx 1.23.2 Connections: 1000 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 46.60, N = 3 SE +/- 52.64, N = 3 SE +/- 98.21, N = 3 SE +/- 536.64, N = 3 SE +/- 301.50, N = 3 87132.22 84772.31 85990.12 144355.05 136360.72 1. (CC) gcc options: -lluajit-5.1 -lm -lssl -lcrypto -lpthread -ldl -std=c99 -O2
Apache HTTP Server This is a test of the Apache HTTPD web server. This Apache HTTPD web server benchmark test profile makes use of the wrk program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better Apache HTTP Server 2.4.56 Concurrent Requests: 1000 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 873.03, N = 3 SE +/- 59.58, N = 3 SE +/- 62.36, N = 3 SE +/- 802.85, N = 3 SE +/- 797.39, N = 3 113097.85 87686.63 87708.39 130637.70 131437.32 1. (CC) gcc options: -lluajit-5.1 -lm -lssl -lcrypto -lpthread -ldl -std=c99 -O2
nginx This is a benchmark of the lightweight Nginx HTTP(S) web-server. This Nginx web server benchmark test profile makes use of the wrk program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients/connections. HTTPS with a self-signed OpenSSL certificate is used by this test for local benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better nginx 1.23.2 Connections: 500 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 173.47, N = 3 SE +/- 239.98, N = 3 SE +/- 191.32, N = 3 SE +/- 126.64, N = 3 SE +/- 288.60, N = 3 92582.29 91505.24 94947.75 156935.63 154835.06 1. (CC) gcc options: -lluajit-5.1 -lm -lssl -lcrypto -lpthread -ldl -std=c99 -O2
Apache HTTP Server This is a test of the Apache HTTPD web server. This Apache HTTPD web server benchmark test profile makes use of the wrk program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better Apache HTTP Server 2.4.56 Concurrent Requests: 500 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 774.43, N = 3 SE +/- 141.36, N = 3 SE +/- 152.51, N = 3 SE +/- 667.30, N = 3 SE +/- 521.55, N = 3 125745.61 96630.23 96508.30 136714.05 133534.11 1. (CC) gcc options: -lluajit-5.1 -lm -lssl -lcrypto -lpthread -ldl -std=c99 -O2
nginx This is a benchmark of the lightweight Nginx HTTP(S) web-server. This Nginx web server benchmark test profile makes use of the wrk program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients/connections. HTTPS with a self-signed OpenSSL certificate is used by this test for local benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better nginx 1.23.2 Connections: 200 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 231.66, N = 3 SE +/- 204.87, N = 3 SE +/- 127.61, N = 3 SE +/- 223.05, N = 3 SE +/- 150.85, N = 3 98692.95 100527.89 101214.88 162256.48 162348.33 1. (CC) gcc options: -lluajit-5.1 -lm -lssl -lcrypto -lpthread -ldl -std=c99 -O2
Appleseed Appleseed is an open-source production renderer focused on physically-based global illumination rendering engine primarily designed for animation and visual effects. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Appleseed 2.0 Beta Scene: Material Tester Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 173.41 143.38 142.94 102.74 102.15
MariaDB mariadb-slap This is a MariaDB MySQL database server benchmark making use of mysqlslap (mariadb-slap). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB mariadb-slap 11.5 Clients: 64 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 4.64, N = 3 SE +/- 5.02, N = 3 SE +/- 5.66, N = 3 SE +/- 5.34, N = 3 SE +/- 4.85, N = 3 1128 1132 1124 1182 1183 -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -lnuma -lpcre2-8 -lcrypt -laio -lssl -lcrypto -shared -lrt -lstdc++ -shared -lrt -lstdc++ -shared -lrt -lstdc++ 1. (CXX) g++ options: -pie -fPIC -fstack-protector -O3 -lz -lm -lpthread -ldl
Appleseed Appleseed is an open-source production renderer focused on physically-based global illumination rendering engine primarily designed for animation and visual effects. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Appleseed 2.0 Beta Scene: Disney Material Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 174.95 156.37 156.07 85.93 85.32
Blender Blender is an open-source 3D creation and modeling software project. This test is of Blender's Cycles performance with various sample files. GPU computing via NVIDIA OptiX and NVIDIA CUDA is currently supported as well as HIP for AMD Radeon GPUs and Intel oneAPI for Intel Graphics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.1 Blend File: BMW27 - Compute: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.18, N = 3 SE +/- 0.24, N = 3 SE +/- 0.33, N = 3 SE +/- 0.16, N = 3 SE +/- 0.05, N = 3 123.51 101.52 101.72 53.56 53.25
Timed Linux Kernel Compilation This test times how long it takes to build the Linux kernel in a default configuration (defconfig) for the architecture being tested or alternatively an allmodconfig for building all possible kernel modules for the build. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Linux Kernel Compilation 6.8 Build: defconfig Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.45, N = 3 SE +/- 0.32, N = 3 SE +/- 0.32, N = 3 SE +/- 0.42, N = 3 SE +/- 0.36, N = 3 108.34 97.92 97.41 57.93 57.47
Renaissance Renaissance is a suite of benchmarks designed to test the Java JVM from Apache Spark to a Twitter-like service to Scala and other features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Renaissance 0.14 Test: Genetic Algorithm Using Jenetics + Futures Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 9.79, N = 3 SE +/- 9.08, N = 3 SE +/- 5.89, N = 3 SE +/- 8.75, N = 3 SE +/- 6.12, N = 15 886.6 819.5 821.7 887.0 882.7 MIN: 844.84 / MAX: 942.25 MIN: 773.51 / MAX: 850.94 MIN: 782.09 / MAX: 843.31 MIN: 771.69 / MAX: 950.63 MIN: 756.16 / MAX: 981
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: Recurrent Neural Network Training - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 9.27, N = 3 SE +/- 3.90, N = 3 SE +/- 1.64, N = 3 SE +/- 3.18, N = 3 SE +/- 1.43, N = 3 2931.95 2496.65 2491.14 1354.12 1348.37 MIN: 2882.15 MIN: 2449.87 MIN: 2459.9 MIN: 1317.27 MIN: 1314.05 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
CloverLeaf CloverLeaf is a Lagrangian-Eulerian hydrodynamics benchmark. This test profile currently makes use of CloverLeaf's OpenMP version. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better CloverLeaf 1.3 Input: clover_bm Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.27, N = 3 SE +/- 0.09, N = 3 SE +/- 0.21, N = 3 SE +/- 0.09, N = 3 SE +/- 0.45, N = 3 64.58 89.36 86.82 75.55 72.88 1. (F9X) gfortran options: -O3 -march=native -funroll-loops -fopenmp
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 3 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 76.92, N = 3 SE +/- 59.04, N = 3 SE +/- 273.73, N = 3 SE +/- 146.12, N = 3 SE +/- 101.69, N = 3 95447 84611 84209 44561 44247
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: Recurrent Neural Network Inference - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 6.38, N = 3 SE +/- 2.00, N = 3 SE +/- 7.18, N = 3 SE +/- 2.19, N = 3 SE +/- 1.91, N = 3 1548.71 1329.68 1337.47 703.75 706.65 MIN: 1512.56 MIN: 1297.27 MIN: 1299.81 MIN: 675.7 MIN: 679.48 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
PyTorch This is a benchmark of PyTorch making use of pytorch-benchmark [https://github.com/LukasHedegaard/pytorch-benchmark]. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 1 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 15 30 45 60 75 SE +/- 1.37, N = 12 SE +/- 1.71, N = 12 SE +/- 1.84, N = 15 SE +/- 0.29, N = 3 SE +/- 0.90, N = 15 61.49 68.30 67.48 66.63 65.69 MIN: 40.06 / MAX: 72.22 MIN: 48 / MAX: 73.52 MIN: 46.76 / MAX: 74.15 MIN: 57.04 / MAX: 69.34 MIN: 53.12 / MAX: 69.76
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 64 - Model: GoogLeNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 SE +/- 0.17, N = 3 SE +/- 0.07, N = 3 78.17 91.42 92.37 129.59 133.21
Chaos Group V-RAY This is a test of Chaos Group's V-RAY benchmark. V-RAY is a commercial renderer that can integrate with various creator software products like SketchUp and 3ds Max. The V-RAY benchmark is standalone and supports CPU and NVIDIA CUDA/RTX based rendering. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org vsamples, More Is Better Chaos Group V-RAY 6.0 Mode: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9K 18K 27K 36K 45K SE +/- 16.84, N = 3 SE +/- 57.13, N = 3 SE +/- 11.00, N = 3 SE +/- 38.48, N = 3 SE +/- 204.38, N = 3 20060 23056 23078 43567 43879
ASTC Encoder ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MT/s, More Is Better ASTC Encoder 4.7 Preset: Thorough Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.32, N = 12 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 12.23 10.32 10.21 19.68 19.90 1. (CXX) g++ options: -O3 -flto -pthread
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 3 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 0.88, N = 3 SE +/- 1.20, N = 3 SE +/- 4.04, N = 3 SE +/- 1.53, N = 3 SE +/- 0.33, N = 3 2884 2539 2526 1275 1265
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: CG.C Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2K 4K 6K 8K 10K SE +/- 210.57, N = 15 SE +/- 81.68, N = 15 SE +/- 92.49, N = 15 SE +/- 68.61, N = 9 SE +/- 89.79, N = 6 9750.34 7857.53 7898.41 8784.83 9036.65 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
QMCPACK QMCPACK is a modern high-performance open-source Quantum Monte Carlo (QMC) simulation code making use of MPI for this benchmark of the H20 example code. QMCPACK is an open-source production level many-body ab initio Quantum Monte Carlo code for computing the electronic structure of atoms, molecules, and solids. QMCPACK is supported by the U.S. Department of Energy. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Execution Time - Seconds, Fewer Is Better QMCPACK 3.17.1 Input: simple-H2O Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 1.38, N = 12 SE +/- 0.82, N = 15 SE +/- 0.73, N = 12 SE +/- 0.19, N = 3 SE +/- 0.05, N = 3 24.06 21.45 23.12 25.62 25.25 1. (CXX) g++ options: -fopenmp -foffload=disable -finline-limit=1000 -fstrict-aliasing -funroll-all-loops -ffast-math -march=native -O3 -lm -ldl
ASKAP ASKAP is a set of benchmarks from the Australian SKA Pathfinder. The principal ASKAP benchmarks are the Hogbom Clean Benchmark (tHogbomClean) and Convolutional Resamping Benchmark (tConvolve) as well as some previous ASKAP benchmarks being included as well for OpenCL and CUDA execution of tConvolve. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 1.0 Test: tConvolve MT - Degridding Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 700 1400 2100 2800 3500 SE +/- 6.39, N = 3 SE +/- 4.57, N = 3 SE +/- 6.80, N = 3 SE +/- 6.45, N = 3 SE +/- 14.04, N = 3 3033.56 1791.99 1806.71 2291.85 2387.89 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 1.0 Test: tConvolve MT - Gridding Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 4.91, N = 3 SE +/- 2.12, N = 3 SE +/- 3.42, N = 3 SE +/- 1.09, N = 3 SE +/- 1.46, N = 3 2260.82 1550.64 1619.82 1917.67 1977.00 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.18, N = 3 SE +/- 0.16, N = 3 SE +/- 0.17, N = 3 SE +/- 0.19, N = 5 SE +/- 0.07, N = 3 24.39 19.33 19.79 19.32 18.97 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12 24 36 48 60 SE +/- 0.30, N = 3 SE +/- 0.43, N = 3 SE +/- 0.43, N = 3 SE +/- 0.52, N = 5 SE +/- 0.21, N = 3 41.01 51.72 50.53 51.78 52.72 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.18, N = 3 SE +/- 0.16, N = 3 SE +/- 0.17, N = 3 SE +/- 0.19, N = 5 SE +/- 0.07, N = 3 24.39 19.33 19.79 19.32 18.97 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.1.1 Model: Tomographic Model Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12 24 36 48 60 SE +/- 0.06, N = 3 SE +/- 0.38, N = 3 SE +/- 0.48, N = 3 SE +/- 0.37, N = 3 SE +/- 0.20, N = 15 52.89 46.58 49.58 30.49 30.13 1. (F9X) gfortran options: -O2 -fopenmp -std=f2008 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 2 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 2.33, N = 3 SE +/- 2.73, N = 3 SE +/- 0.33, N = 3 SE +/- 1.00, N = 3 SE +/- 2.08, N = 3 2474 2171 2156 1090 1080
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 1 - Resolution: 1080p - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 0.88, N = 3 SE +/- 2.19, N = 3 SE +/- 1.76, N = 3 SE +/- 1.15, N = 3 SE +/- 1.00, N = 3 2441 2143 2133 1078 1069
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Person Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.73, N = 3 SE +/- 0.11, N = 3 SE +/- 0.15, N = 3 SE +/- 1.00, N = 5 SE +/- 1.01, N = 3 133.99 73.87 72.52 92.38 89.98 MIN: 103.21 / MAX: 233.72 MIN: 36.14 / MAX: 99.53 MIN: 28.47 / MAX: 87.6 MIN: 31.8 / MAX: 128.67 MIN: 63.39 / MAX: 124.53 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Person Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.16, N = 3 SE +/- 0.08, N = 3 SE +/- 0.11, N = 3 SE +/- 0.92, N = 5 SE +/- 0.98, N = 3 29.83 54.11 55.11 86.56 88.84 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: CPU Cache Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.3M 2.6M 3.9M 5.2M 6.5M SE +/- 59300.10, N = 15 SE +/- 44904.06, N = 10 SE +/- 27236.78, N = 3 SE +/- 29190.39, N = 3 SE +/- 18366.37, N = 3 4023464.53 5854131.56 5997921.20 2537425.46 2446936.66 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 2 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 34.91, N = 3 SE +/- 128.55, N = 3 SE +/- 152.10, N = 3 SE +/- 126.32, N = 3 SE +/- 61.02, N = 3 82403 72906 72250 38454 38192
Llama.cpp Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Tokens Per Second, More Is Better Llama.cpp b1808 Model: llama-2-13b.Q4_0.gguf Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 8.38 6.70 6.57 6.98 7.32 1. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 1 - Resolution: 1080p - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 135.20, N = 3 SE +/- 135.51, N = 3 SE +/- 145.56, N = 3 SE +/- 192.07, N = 3 SE +/- 168.15, N = 3 81236 72067 71530 37984 37923
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.00822, N = 3 SE +/- 0.00132, N = 3 SE +/- 0.00236, N = 3 SE +/- 0.00685, N = 3 SE +/- 0.00353, N = 3 3.79539 4.65305 4.66757 8.92149 8.97813
Memcached Memcached is a high performance, distributed memory object caching system. This Memcached test profiles makes use of memtier_benchmark for excuting this CPU/memory-focused server benchmark. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Memcached 1.6.19 Set To Get Ratio: 1:10 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 800K 1600K 2400K 3200K 4000K SE +/- 7388.22, N = 3 SE +/- 1626.82, N = 3 SE +/- 2761.61, N = 3 SE +/- 5916.60, N = 3 SE +/- 5303.54, N = 3 3072373.87 2217988.24 2195688.49 3775389.09 3761626.61 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
OpenBenchmarking.org Ops/sec, More Is Better Memcached 1.6.19 Set To Get Ratio: 1:100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 800K 1600K 2400K 3200K 4000K SE +/- 11523.53, N = 3 SE +/- 1526.11, N = 3 SE +/- 1471.26, N = 3 SE +/- 3096.61, N = 3 SE +/- 5665.65, N = 3 3048853.56 2124192.99 2104490.55 3850014.66 3844031.28 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: gravity_spheres_volume/dim_512/ao/real_time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.00479, N = 3 SE +/- 0.01340, N = 3 SE +/- 0.00682, N = 3 SE +/- 0.01707, N = 3 SE +/- 0.00259, N = 3 2.73496 3.87827 3.91070 7.45782 7.46649
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: gravity_spheres_volume/dim_512/scivis/real_time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.00622, N = 3 SE +/- 0.01059, N = 3 SE +/- 0.00260, N = 3 SE +/- 0.00203, N = 3 SE +/- 0.02548, N = 3 2.63851 3.82321 3.82138 7.31921 7.36909
LuxCoreRender LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.6 Scene: Orange Juice - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.04, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 3.58 4.42 4.42 7.70 7.72 MIN: 3.44 / MAX: 4.42 MIN: 4.35 / MAX: 4.49 MIN: 4.32 / MAX: 4.49 MIN: 6.73 / MAX: 8.14 MIN: 6.87 / MAX: 8.15
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Face Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 8.08, N = 3 SE +/- 0.34, N = 3 SE +/- 0.52, N = 3 SE +/- 1.15, N = 3 SE +/- 1.73, N = 3 1506.63 556.14 560.14 592.59 590.80 MIN: 1180.01 / MAX: 1642.12 MIN: 444.24 / MAX: 578.41 MIN: 538 / MAX: 581.72 MIN: 531.26 / MAX: 631.27 MIN: 555.5 / MAX: 623.22 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Face Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 2.65 7.19 7.14 13.46 13.52 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
Zstd Compression This test measures the time needed to compress/decompress a sample file (silesia.tar) using Zstd (Zstandard) compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 8, Long Mode - Decompression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 5.21, N = 3 SE +/- 7.92, N = 3 SE +/- 7.19, N = 3 SE +/- 3.89, N = 3 SE +/- 6.24, N = 3 2375.4 2324.5 2352.5 2278.0 2299.7 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.4 Compression Level: 8, Long Mode - Compression Speed Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 1.12, N = 3 SE +/- 0.67, N = 3 SE +/- 0.95, N = 3 SE +/- 3.67, N = 3 SE +/- 3.49, N = 3 504.1 518.5 527.9 778.8 797.0 1. (CC) gcc options: -O3 -pthread -lz -llzma
LuxCoreRender LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.6 Scene: LuxCore Benchmark - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.0463 2.0926 3.1389 4.1852 5.2315 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 1.69 2.41 2.41 4.62 4.65 MIN: 0.51 / MAX: 2.04 MIN: 0.83 / MAX: 2.84 MIN: 0.82 / MAX: 2.85 MIN: 2 / MAX: 5.24 MIN: 2.03 / MAX: 5.27
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Face Detection FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80 160 240 320 400 SE +/- 0.59, N = 3 SE +/- 0.44, N = 3 SE +/- 0.15, N = 3 SE +/- 0.18, N = 3 SE +/- 0.24, N = 3 384.09 283.71 285.97 301.83 301.79 MIN: 285.45 / MAX: 543.49 MIN: 240.12 / MAX: 303.33 MIN: 150.08 / MAX: 303.22 MIN: 264.12 / MAX: 322.86 MIN: 278.87 / MAX: 327.8 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Face Detection FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 10.41 14.07 13.97 26.44 26.46 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 4K - Video Preset: Medium Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 6.62 8.62 8.61 15.70 15.92
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: GPT-2 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.03060, N = 3 SE +/- 0.00562, N = 3 SE +/- 0.02822, N = 3 SE +/- 0.01596, N = 3 SE +/- 0.00809, N = 3 8.10864 8.26788 8.31107 8.40671 8.15155 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: GPT-2 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.46, N = 3 SE +/- 0.08, N = 3 SE +/- 0.41, N = 3 SE +/- 0.22, N = 3 SE +/- 0.12, N = 3 123.28 120.91 120.28 118.91 122.63 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: GPT-2 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.03060, N = 3 SE +/- 0.00562, N = 3 SE +/- 0.02822, N = 3 SE +/- 0.01596, N = 3 SE +/- 0.00809, N = 3 8.10864 8.26788 8.31107 8.40671 8.15155 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Xcompact3d Incompact3d Xcompact3d Incompact3d is a Fortran-MPI based, finite difference high-performance code for solving the incompressible Navier-Stokes equation and as many as you need scalar transport equations. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Xcompact3d Incompact3d 2021-03-11 Input: input.i3d 129 Cells Per Direction Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.22, N = 3 SE +/- 0.17, N = 15 SE +/- 0.18, N = 9 SE +/- 0.21, N = 3 SE +/- 0.12, N = 15 19.75 22.53 21.87 17.11 16.59 1. (F9X) gfortran options: -cpp -O2 -funroll-loops -floop-optimize -fcray-pointer -fbacktrace -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: bertsquad-12 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.78, N = 3 SE +/- 0.35, N = 3 SE +/- 0.76, N = 3 SE +/- 0.63, N = 3 SE +/- 0.22, N = 3 86.69 81.57 79.70 63.11 64.28 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: bertsquad-12 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.10, N = 3 SE +/- 0.05, N = 3 SE +/- 0.12, N = 3 SE +/- 0.16, N = 3 SE +/- 0.05, N = 3 11.54 12.26 12.55 15.85 15.56 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: bertsquad-12 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.78, N = 3 SE +/- 0.35, N = 3 SE +/- 0.76, N = 3 SE +/- 0.63, N = 3 SE +/- 0.22, N = 3 86.69 81.57 79.70 63.11 64.28 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
LuxCoreRender LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.6 Scene: DLSC - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.125 2.25 3.375 4.5 5.625 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 2.13 2.74 2.75 4.96 5.00 MIN: 2.03 / MAX: 2.65 MIN: 2.66 / MAX: 2.85 MIN: 2.66 / MAX: 2.86 MIN: 4.84 / MAX: 5.29 MIN: 4.85 / MAX: 5.35
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Sharpen Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 13 26 39 52 65 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 36 31 30 57 58 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Machine Translation EN To DE FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.08, N = 3 SE +/- 0.15, N = 3 SE +/- 0.10, N = 3 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 118.45 48.55 48.26 59.77 58.43 MIN: 68.57 / MAX: 244.71 MIN: 29.96 / MAX: 68.41 MIN: 29.16 / MAX: 69.56 MIN: 40 / MAX: 88.32 MIN: 31.85 / MAX: 94.55 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Machine Translation EN To DE FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.02, N = 3 SE +/- 0.25, N = 3 SE +/- 0.17, N = 3 SE +/- 0.07, N = 3 SE +/- 0.06, N = 3 33.74 82.32 82.80 133.65 136.73 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: T5 Encoder - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.03755, N = 3 SE +/- 0.09318, N = 3 SE +/- 0.06732, N = 3 SE +/- 0.01065, N = 3 SE +/- 0.02831, N = 3 6.19876 7.00331 7.06097 6.57601 6.37806 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: T5 Encoder - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.97, N = 3 SE +/- 1.92, N = 3 SE +/- 1.35, N = 3 SE +/- 0.25, N = 3 SE +/- 0.69, N = 3 161.31 142.82 141.63 152.04 156.77 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: T5 Encoder - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.03755, N = 3 SE +/- 0.09318, N = 3 SE +/- 0.06732, N = 3 SE +/- 0.01065, N = 3 SE +/- 0.02831, N = 3 6.19876 7.00331 7.06097 6.57601 6.37806 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Person Detection FP32 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.77, N = 3 SE +/- 0.12, N = 3 SE +/- 0.13, N = 3 SE +/- 0.63, N = 3 SE +/- 0.48, N = 3 134.54 73.71 72.70 91.34 90.49 MIN: 98.92 / MAX: 233.45 MIN: 31.23 / MAX: 89.7 MIN: 40.38 / MAX: 84.07 MIN: 68.51 / MAX: 134.9 MIN: 55.81 / MAX: 127.95 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Person Detection FP32 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.17, N = 3 SE +/- 0.09, N = 3 SE +/- 0.09, N = 3 SE +/- 0.61, N = 3 SE +/- 0.47, N = 3 29.71 54.23 54.98 87.50 88.32 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Noise Suppression Poconet-Like FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 15.21 8.03 7.95 10.18 9.90 MIN: 10.08 / MAX: 64.69 MIN: 5.57 / MAX: 23.51 MIN: 5 / MAX: 23.85 MIN: 5.82 / MAX: 38.39 MIN: 6.53 / MAX: 37.69 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Noise Suppression Poconet-Like FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 1.17, N = 3 SE +/- 0.72, N = 3 SE +/- 0.86, N = 3 SE +/- 1.94, N = 3 SE +/- 1.23, N = 3 523.02 980.84 991.32 1544.55 1587.74 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Road Segmentation ADAS FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 8 16 24 32 40 SE +/- 0.05, N = 3 SE +/- 0.06, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 36.35 14.59 14.48 17.22 16.73 MIN: 24.38 / MAX: 96.4 MIN: 9.42 / MAX: 26.08 MIN: 8.91 / MAX: 26.44 MIN: 10.01 / MAX: 28.19 MIN: 9.66 / MAX: 30.96 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Road Segmentation ADAS FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 100 200 300 400 500 SE +/- 0.28, N = 3 SE +/- 1.13, N = 3 SE +/- 0.50, N = 3 SE +/- 1.58, N = 3 SE +/- 1.94, N = 3 219.86 273.50 275.72 462.96 476.63 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Enhanced Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 49 50 50 93 94 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Noise-Gaussian Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 70 76 76 116 119 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Person Vehicle Bike Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 9.94 4.78 4.78 5.13 5.12 MIN: 6.2 / MAX: 41.65 MIN: 3.25 / MAX: 16.12 MIN: 3.18 / MAX: 16.31 MIN: 3.4 / MAX: 14.3 MIN: 3.61 / MAX: 15.43 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Person Vehicle Bike Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 0.87, N = 3 SE +/- 2.43, N = 3 SE +/- 0.80, N = 3 SE +/- 4.59, N = 3 SE +/- 2.52, N = 3 401.42 834.18 833.73 1551.75 1556.07 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Road Segmentation ADAS FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 39.48 18.02 17.79 26.38 25.15 MIN: 22.02 / MAX: 90.7 MIN: 11.1 / MAX: 33.13 MIN: 9.38 / MAX: 30.36 MIN: 13.08 / MAX: 46.13 MIN: 11.63 / MAX: 50.5 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Road Segmentation ADAS FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 0.22, N = 3 SE +/- 1.23, N = 3 SE +/- 1.10, N = 3 SE +/- 0.60, N = 3 SE +/- 0.46, N = 3 101.21 221.52 224.43 302.45 317.25 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Person Re-Identification Retail FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.11, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 18.22 3.85 3.87 4.11 4.11 MIN: 10.36 / MAX: 177.58 MIN: 2.35 / MAX: 15.38 MIN: 2.38 / MAX: 13.89 MIN: 2.52 / MAX: 16.5 MIN: 2.45 / MAX: 41.27 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Person Re-Identification Retail FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 400 800 1200 1600 2000 SE +/- 2.58, N = 3 SE +/- 0.54, N = 3 SE +/- 0.99, N = 3 SE +/- 0.55, N = 3 SE +/- 0.22, N = 3 437.96 1036.33 1029.88 1938.68 1938.32 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 3 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10K 20K 30K 40K 50K SE +/- 115.84, N = 3 SE +/- 61.68, N = 3 SE +/- 62.68, N = 3 SE +/- 13.61, N = 3 SE +/- 29.87, N = 3 48929 44013 43830 20461 20283
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Handwritten English Recognition FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 SE +/- 0.12, N = 3 SE +/- 0.00, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 45.70 18.90 19.13 20.11 20.09 MIN: 32.92 / MAX: 117.92 MIN: 13.12 / MAX: 33.09 MIN: 12.57 / MAX: 34 MIN: 14.53 / MAX: 37.6 MIN: 13.66 / MAX: 38.02 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Handwritten English Recognition FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 0.45, N = 3 SE +/- 0.13, N = 3 SE +/- 0.59, N = 3 SE +/- 0.92, N = 3 SE +/- 1.22, N = 3 174.92 422.62 417.64 794.33 795.05 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Random Fill Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200K 400K 600K 800K 1000K SE +/- 5918.01, N = 3 SE +/- 738.85, N = 3 SE +/- 1877.58, N = 3 SE +/- 826.71, N = 3 SE +/- 1278.43, N = 3 876130 1011788 1005973 869816 866759 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Vehicle Detection FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.05, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 15.95 4.52 4.53 4.83 4.82 MIN: 9.85 / MAX: 70.28 MIN: 2.64 / MAX: 15.69 MIN: 2.65 / MAX: 17.47 MIN: 2.79 / MAX: 19.86 MIN: 2.78 / MAX: 17.83 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Vehicle Detection FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 400 800 1200 1600 2000 SE +/- 1.52, N = 3 SE +/- 0.30, N = 3 SE +/- 0.43, N = 3 SE +/- 0.97, N = 3 SE +/- 1.19, N = 3 500.86 882.75 879.50 1648.32 1653.69 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Handwritten English Recognition FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12 24 36 48 60 SE +/- 0.12, N = 3 SE +/- 0.06, N = 3 SE +/- 0.05, N = 3 SE +/- 0.06, N = 3 SE +/- 0.09, N = 3 53.07 20.34 20.37 21.99 21.89 MIN: 35.4 / MAX: 130.32 MIN: 13.65 / MAX: 36.45 MIN: 14.74 / MAX: 35.54 MIN: 14.99 / MAX: 41.87 MIN: 14.64 / MAX: 40.37 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Handwritten English Recognition FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 160 320 480 640 800 SE +/- 0.33, N = 3 SE +/- 1.21, N = 3 SE +/- 1.00, N = 3 SE +/- 1.88, N = 3 SE +/- 3.18, N = 3 150.64 392.93 392.23 726.61 729.84 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Face Detection Retail FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.1408 2.2816 3.4224 4.5632 5.704 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 5.07 3.13 3.16 3.32 3.33 MIN: 2.36 / MAX: 27.99 MIN: 1.79 / MAX: 14.78 MIN: 1.77 / MAX: 13.91 MIN: 1.82 / MAX: 16.13 MIN: 1.88 / MAX: 15.67 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Face Detection Retail FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1000 2000 3000 4000 5000 SE +/- 3.53, N = 3 SE +/- 5.39, N = 3 SE +/- 1.63, N = 3 SE +/- 5.85, N = 3 SE +/- 1.99, N = 3 1559.06 2522.56 2495.55 4743.73 4734.27 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Random Fill Sync Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50K 100K 150K 200K 250K SE +/- 573.89, N = 3 SE +/- 257.69, N = 3 SE +/- 468.76, N = 3 SE +/- 490.37, N = 3 SE +/- 238.71, N = 3 225607 128855 127955 170365 171623 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.09 0.18 0.27 0.36 0.45 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.40 0.28 0.28 0.29 0.29 MIN: 0.18 / MAX: 8.35 MIN: 0.15 / MAX: 7.24 MIN: 0.15 / MAX: 7.71 MIN: 0.16 / MAX: 9.68 MIN: 0.16 / MAX: 8.63 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11K 22K 33K 44K 55K SE +/- 38.88, N = 3 SE +/- 67.36, N = 3 SE +/- 19.90, N = 3 SE +/- 49.45, N = 3 SE +/- 47.80, N = 3 18507.31 26465.33 26230.85 49580.07 49485.35 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Vehicle Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 16.04 8.57 8.44 11.72 11.36 MIN: 9.49 / MAX: 56.98 MIN: 3.77 / MAX: 23.35 MIN: 4.55 / MAX: 19.23 MIN: 4.13 / MAX: 31.49 MIN: 4.12 / MAX: 36.81 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Vehicle Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 150 300 450 600 750 SE +/- 0.14, N = 3 SE +/- 0.22, N = 3 SE +/- 1.08, N = 3 SE +/- 3.49, N = 3 SE +/- 2.53, N = 3 249.02 465.15 472.57 680.56 701.84 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Weld Porosity Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7 14 21 28 35 SE +/- 0.06, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 30.05 10.97 11.05 11.75 11.72 MIN: 17.38 / MAX: 87.99 MIN: 5.79 / MAX: 25.09 MIN: 5.83 / MAX: 23.56 MIN: 5.98 / MAX: 26.57 MIN: 6 / MAX: 26.83 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Weld Porosity Detection FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 0.54, N = 3 SE +/- 0.11, N = 3 SE +/- 0.13, N = 3 SE +/- 1.41, N = 3 SE +/- 1.37, N = 3 265.83 728.18 723.13 1359.29 1362.00 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Update Random Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200K 400K 600K 800K 1000K SE +/- 1460.51, N = 3 SE +/- 819.19, N = 3 SE +/- 1312.04, N = 3 SE +/- 726.71, N = 3 SE +/- 1173.38, N = 3 828184 556040 553791 624743 623366 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.225 0.45 0.675 0.9 1.125 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 1.00 0.40 0.40 0.41 0.41 MIN: 0.4 / MAX: 10.73 MIN: 0.2 / MAX: 8.52 MIN: 0.2 / MAX: 7.69 MIN: 0.21 / MAX: 10.47 MIN: 0.21 / MAX: 10.02 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7K 14K 21K 28K 35K SE +/- 31.71, N = 3 SE +/- 3.66, N = 3 SE +/- 8.32, N = 3 SE +/- 30.10, N = 3 SE +/- 16.51, N = 3 7499.47 18754.59 18615.12 34850.47 34815.39 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Weld Porosity Detection FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 7.78 5.59 5.63 5.96 5.96 MIN: 3.64 / MAX: 44.07 MIN: 2.93 / MAX: 16.72 MIN: 2.98 / MAX: 16.09 MIN: 3.09 / MAX: 18.13 MIN: 3.08 / MAX: 18.52 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Weld Porosity Detection FP16-INT8 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 2.31, N = 3 SE +/- 0.44, N = 3 SE +/- 0.60, N = 3 SE +/- 3.06, N = 3 SE +/- 4.11, N = 3 1022.10 1427.32 1417.33 2669.64 2672.51 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: CaffeNet 12-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.4035 0.807 1.2105 1.614 2.0175 SE +/- 0.01234, N = 3 SE +/- 0.00719, N = 3 SE +/- 0.01555, N = 3 SE +/- 0.00633, N = 3 SE +/- 0.01197, N = 3 1.79347 1.46770 1.47531 1.19855 1.20504 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: CaffeNet 12-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 3.86, N = 3 SE +/- 3.38, N = 3 SE +/- 7.12, N = 3 SE +/- 4.40, N = 3 SE +/- 8.14, N = 3 557.41 680.95 677.58 833.52 829.19 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: CaffeNet 12-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.4035 0.807 1.2105 1.614 2.0175 SE +/- 0.01234, N = 3 SE +/- 0.00719, N = 3 SE +/- 0.01555, N = 3 SE +/- 0.00633, N = 3 SE +/- 0.01197, N = 3 1.79347 1.46770 1.47531 1.19855 1.20504 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: CaffeNet 12-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.3827 0.7654 1.1481 1.5308 1.9135 SE +/- 0.01304, N = 3 SE +/- 0.00355, N = 3 SE +/- 0.00807, N = 3 SE +/- 0.00557, N = 3 SE +/- 0.00549, N = 3 1.70107 1.48191 1.47738 1.11712 1.11577 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: CaffeNet 12-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 4.53, N = 3 SE +/- 1.59, N = 3 SE +/- 3.70, N = 3 SE +/- 4.40, N = 3 SE +/- 4.41, N = 3 587.73 674.51 676.65 894.66 895.77 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: CaffeNet 12-int8 - Device: CPU - Executor: Standard Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.3827 0.7654 1.1481 1.5308 1.9135 SE +/- 0.01304, N = 3 SE +/- 0.00355, N = 3 SE +/- 0.00807, N = 3 SE +/- 0.00557, N = 3 SE +/- 0.00549, N = 3 1.70107 1.48191 1.47738 1.11712 1.11577 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2024.0 Model: Face Detection Retail FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 9.76 2.20 2.21 2.37 2.37 MIN: 4.34 / MAX: 46.89 MIN: 1.23 / MAX: 12.5 MIN: 1.23 / MAX: 13.12 MIN: 1.26 / MAX: 11.94 MIN: 1.27 / MAX: 12.75 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
OpenBenchmarking.org FPS, More Is Better OpenVINO 2024.0 Model: Face Detection Retail FP16 - Device: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 700 1400 2100 2800 3500 SE +/- 1.96, N = 3 SE +/- 1.03, N = 3 SE +/- 0.66, N = 3 SE +/- 6.99, N = 3 SE +/- 6.25, N = 3 816.17 1791.06 1781.62 3283.98 3269.12 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
John The Ripper This is a benchmark of John The Ripper, which is a password cracker. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: MD5 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 900K 1800K 2700K 3600K 4500K SE +/- 2504.83, N = 3 SE +/- 10837.18, N = 3 SE +/- 7423.69, N = 3 SE +/- 19341.95, N = 3 SE +/- 11590.23, N = 3 919791 2230333 2234333 4066667 4118000 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Swirl Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 130 138 139 246 250 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Rotate Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.67, N = 3 SE +/- 1.53, N = 3 317 169 168 161 167 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Read Random Write Random Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600K 1200K 1800K 2400K 3000K SE +/- 4914.43, N = 3 SE +/- 1481.36, N = 3 SE +/- 1406.43, N = 3 SE +/- 5448.80, N = 3 SE +/- 2327.92, N = 3 2120915 2144575 2138776 2770899 2770195 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: HWB Color Space Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 0.67, N = 3 SE +/- 0.88, N = 3 SE +/- 0.33, N = 3 SE +/- 0.88, N = 3 SE +/- 0.58, N = 3 315 244 255 269 278 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Read While Writing Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.1M 2.2M 3.3M 4.4M 5.5M SE +/- 20209.24, N = 3 SE +/- 13925.00, N = 3 SE +/- 14207.26, N = 3 SE +/- 22040.69, N = 3 SE +/- 26941.48, N = 3 2548837 3212276 3194971 5146187 5119891 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Resizing Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 1.86, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.88, N = 3 SE +/- 1.00, N = 3 244 251 255 332 338 1. (CC) gcc options: -fopenmp -O2 -ltiff -ljbig -lsharpyuv -lwebp -lwebpmux -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lzstd -llzma -lbz2 -lz -lm -lpthread -lgomp
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Random Fill Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 400K 800K 1200K 1600K 2000K SE +/- 10692.91, N = 3 SE +/- 1270.91, N = 3 SE +/- 4512.41, N = 3 SE +/- 2055.67, N = 3 SE +/- 3781.08, N = 3 1647988 1154716 1159865 1229256 1225921 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.0486 2.0972 3.1458 4.1944 5.243 SE +/- 0.03176, N = 3 SE +/- 0.01507, N = 3 SE +/- 0.00856, N = 3 SE +/- 0.00459, N = 3 SE +/- 0.02511, N = 3 4.66040 3.07441 3.09905 2.58704 2.62524 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80 160 240 320 400 SE +/- 1.45, N = 3 SE +/- 1.58, N = 3 SE +/- 0.89, N = 3 SE +/- 0.69, N = 3 SE +/- 3.68, N = 3 214.55 325.21 322.61 386.41 380.82 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: ResNet50 v1-12-int8 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.0486 2.0972 3.1458 4.1944 5.243 SE +/- 0.03176, N = 3 SE +/- 0.01507, N = 3 SE +/- 0.00856, N = 3 SE +/- 0.00459, N = 3 SE +/- 0.02511, N = 3 4.66040 3.07441 3.09905 2.58704 2.62524 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Random Fill Sync Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40K 80K 120K 160K 200K SE +/- 710.34, N = 3 SE +/- 196.51, N = 3 SE +/- 48.85, N = 3 SE +/- 414.30, N = 3 SE +/- 270.54, N = 3 201594 98054 99746 170735 171199 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Overwrite Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 400K 800K 1200K 1600K 2000K SE +/- 14969.35, N = 3 SE +/- 1933.24, N = 3 SE +/- 1465.82, N = 3 SE +/- 1384.95, N = 3 SE +/- 1555.52, N = 3 1653489 1155121 1152820 1229747 1225361 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Update Random Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200K 400K 600K 800K 1000K SE +/- 2978.74, N = 3 SE +/- 1914.75, N = 3 SE +/- 748.73, N = 3 SE +/- 818.39, N = 3 SE +/- 1578.78, N = 3 908300 609510 618187 703294 700416 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Random Read Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30M 60M 90M 120M 150M SE +/- 170759.02, N = 3 SE +/- 103627.09, N = 3 SE +/- 34435.41, N = 3 SE +/- 185772.57, N = 3 SE +/- 286737.26, N = 3 57824426 72087947 71648683 135603493 134971626 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Read While Writing Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 900K 1800K 2700K 3600K 4500K SE +/- 3646.67, N = 3 SE +/- 15255.97, N = 3 SE +/- 22460.23, N = 3 SE +/- 13356.72, N = 3 SE +/- 7045.87, N = 3 1942760 2476441 2519594 4368397 4294631 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Read Random Write Random Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600K 1200K 1800K 2400K 3000K SE +/- 17144.87, N = 3 SE +/- 2242.91, N = 3 SE +/- 5891.62, N = 3 SE +/- 24632.94, N = 3 SE +/- 24549.22, N = 3 1976533 2137892 2131122 2912747 2915758 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
ONNX Runtime ONNX Runtime is developed by Microsoft and partners as a open-source, cross-platform, high performance machine learning inferencing and training accelerator. This test profile runs the ONNX Runtime with various models available from the ONNX Model Zoo. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: super-resolution-10 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.09981, N = 3 SE +/- 0.14350, N = 3 SE +/- 0.02135, N = 3 SE +/- 0.02431, N = 3 SE +/- 0.06152, N = 3 13.17920 10.73790 10.91820 7.55840 7.50080 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Result
OpenBenchmarking.org Inferences Per Second, More Is Better ONNX Runtime 1.17 Model: super-resolution-10 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.57, N = 3 SE +/- 1.26, N = 3 SE +/- 0.18, N = 3 SE +/- 0.42, N = 3 SE +/- 1.10, N = 3 75.88 93.16 91.59 132.29 133.32 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Inference Time Cost (ms)
OpenBenchmarking.org Inference Time Cost (ms), Fewer Is Better ONNX Runtime 1.17 Model: super-resolution-10 - Device: CPU - Executor: Parallel Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.09981, N = 3 SE +/- 0.14350, N = 3 SE +/- 0.02135, N = 3 SE +/- 0.02431, N = 3 SE +/- 0.06152, N = 3 13.17920 10.73790 10.91820 7.55840 7.50080 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
John The Ripper This is a benchmark of John The Ripper, which is a password cracker. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: HMAC-SHA512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40M 80M 120M 160M 200M SE +/- 518797.01, N = 3 SE +/- 719862.56, N = 3 SE +/- 411753.30, N = 3 SE +/- 877445.98, N = 3 SE +/- 580393.26, N = 3 40721000 117065667 116770333 178283333 181331000 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Random Read Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30M 60M 90M 120M 150M SE +/- 236765.68, N = 3 SE +/- 112684.33, N = 3 SE +/- 144351.01, N = 3 SE +/- 342858.26, N = 3 SE +/- 780691.20, N = 3 47512260 69366631 69408254 131218639 130693063 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
x265 This is a simple test of the x265 encoder run on the CPU with 1080p and 4K options for H.265 video encode performance with x265. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better x265 3.4 Video Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 8 16 24 32 40 SE +/- 0.17, N = 3 SE +/- 0.21, N = 4 SE +/- 0.15, N = 15 SE +/- 0.10, N = 3 SE +/- 0.08, N = 3 19.48 17.73 17.61 34.13 34.65 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
SPECFEM3D simulates acoustic (fluid), elastic (solid), coupled acoustic/elastic, poroelastic or seismic wave propagation in any type of conforming mesh of hexahedra. This test profile currently relies on CPU-based execution for SPECFEM3D and using a variety of their built-in examples/models for benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SPECFEM3D 4.1.1 Model: Homogeneous Halfspace Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 15 30 45 60 75 SE +/- 0.41, N = 3 SE +/- 0.66, N = 3 SE +/- 0.35, N = 3 SE +/- 0.10, N = 3 SE +/- 0.28, N = 3 67.11 57.79 63.87 39.92 39.56 1. (F9X) gfortran options: -O2 -fopenmp -std=f2008 -fimplicit-none -fmax-errors=10 -pedantic -pedantic-errors -O3 -finline-functions -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 1 - Buffer Length: 256 - Filter Length: 512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4M 8M 12M 16M 20M SE +/- 13650.40, N = 3 SE +/- 17835.67, N = 3 SE +/- 250051.55, N = 3 SE +/- 154682.83, N = 3 SE +/- 410224.76, N = 15 20255000 20106667 19611667 19299333 18964400 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Intel Open Image Denoise Open Image Denoise is a denoising library for ray-tracing and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Images / Sec, More Is Better Intel Open Image Denoise 2.2 Run: RT.hdr_alb_nrm.3840x2160 - Device: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.1935 0.387 0.5805 0.774 0.9675 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.42 0.48 0.48 0.85 0.86
OpenBenchmarking.org Images / Sec, More Is Better Intel Open Image Denoise 2.2 Run: RT.ldr_alb_nrm.3840x2160 - Device: CPU-Only Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.1935 0.387 0.5805 0.774 0.9675 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.42 0.48 0.48 0.85 0.86
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 2 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9K 18K 27K 36K 45K SE +/- 172.31, N = 3 SE +/- 82.83, N = 3 SE +/- 127.68, N = 3 SE +/- 24.89, N = 3 SE +/- 19.75, N = 3 42275 38010 37900 17491 17285
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 1 - Resolution: 1080p - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9K 18K 27K 36K 45K SE +/- 112.38, N = 3 SE +/- 115.78, N = 3 SE +/- 43.80, N = 3 SE +/- 11.15, N = 3 SE +/- 33.74, N = 3 41514 37677 37458 17247 17116
Apache Hadoop This is a benchmark of the Apache Hadoop making use of its built-in name-node throughput benchmark (NNThroughputBenchmark). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops per sec, More Is Better Apache Hadoop 3.3.6 Operation: Create - Threads: 100 - Files: 100000 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 1027.34, N = 3 SE +/- 496.20, N = 3 SE +/- 670.20, N = 11 SE +/- 292.68, N = 3 SE +/- 376.90, N = 3 88992 90476 90938 74946 75381
Node.js V8 Web Tooling Benchmark Running the V8 project's Web-Tooling-Benchmark under Node.js. The Web-Tooling-Benchmark stresses JavaScript-related workloads common to web developers like Babel and TypeScript and Babylon. This test profile can test the system's JavaScript performance with Node.js. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org runs/s, More Is Better Node.js V8 Web Tooling Benchmark Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.15, N = 3 SE +/- 0.04, N = 3 SE +/- 0.12, N = 3 SE +/- 0.11, N = 3 SE +/- 0.12, N = 3 23.30 21.89 22.14 20.06 20.81
Z3 Theorem Prover The Z3 Theorem Prover / SMT solver is developed by Microsoft Research under the MIT license. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Z3 Theorem Prover 4.12.1 SMT File: 2.smt2 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14 28 42 56 70 SE +/- 0.51, N = 3 SE +/- 0.21, N = 3 SE +/- 0.27, N = 3 SE +/- 0.35, N = 3 SE +/- 0.53, N = 3 63.08 48.12 48.68 49.00 48.88 1. (CXX) g++ options: -lpthread -std=c++17 -fvisibility=hidden -mfpmath=sse -msse -msse2 -O3 -fPIC
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 2 - Buffer Length: 256 - Filter Length: 512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9M 18M 27M 36M 45M SE +/- 14666.67, N = 3 SE +/- 267690.74, N = 3 SE +/- 368485.05, N = 12 SE +/- 184763.21, N = 3 SE +/- 254700.04, N = 3 40821333 39693000 39094333 38605667 38712667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Myriad-Groestl Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2K 4K 6K 8K 10K SE +/- 30.23, N = 3 SE +/- 0.86, N = 3 SE +/- 6.11, N = 3 SE +/- 195.61, N = 12 SE +/- 13.33, N = 3 4843.72 6023.38 5972.09 11228.13 11337.00 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
libavif avifenc This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 1.0 Encoder Speed: 2 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14 28 42 56 70 SE +/- 0.18, N = 3 SE +/- 0.14, N = 3 SE +/- 0.13, N = 3 SE +/- 0.11, N = 3 SE +/- 0.03, N = 3 61.84 51.38 51.94 36.34 36.33 1. (CXX) g++ options: -O3 -fPIC -lm
m-queens A solver for the N-queens problem with multi-threading support via the OpenMP library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better m-queens 1.2 Time To Solve Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16 32 48 64 80 SE +/- 0.62, N = 3 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 72.85 53.08 53.50 27.82 27.91 1. (CXX) g++ options: -fopenmp -O2 -march=native
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.3 Binary: Pathtracer ISPC - Model: Asian Dragon Obj Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.10, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 14.55 15.57 15.69 26.88 27.39 MIN: 14.14 / MAX: 16.68 MIN: 15.43 / MAX: 15.99 MIN: 15.56 / MAX: 16.03 MIN: 26.58 / MAX: 27.83 MIN: 27.11 / MAX: 28.25
Coremark This is a test of EEMBC CoreMark processor benchmark. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations/Sec, More Is Better Coremark 1.0 CoreMark Size 666 - Iterations Per Second Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200K 400K 600K 800K 1000K SE +/- 3256.58, N = 15 SE +/- 1288.71, N = 3 SE +/- 1076.82, N = 3 SE +/- 2950.54, N = 3 SE +/- 9073.63, N = 3 430017.46 505633.42 499103.50 954988.73 951674.51 1. (CC) gcc options: -O2 -lrt" -lrt
RawTherapee RawTherapee is a cross-platform, open-source multi-threaded RAW image processing program. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better RawTherapee Total Benchmark Time Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.02, N = 3 SE +/- 0.06, N = 3 SE +/- 0.05, N = 3 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 42.66 46.73 46.66 42.79 42.55 1. RawTherapee, version 5.10, command line.
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 1 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 9.16 12.31 12.28 22.86 23.10 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 4K - Video Preset: Medium Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 9.28 12.48 12.49 21.80 22.02 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Renaissance Renaissance is a suite of benchmarks designed to test the Java JVM from Apache Spark to a Twitter-like service to Scala and other features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Renaissance 0.14 Test: Savina Reactors.IO Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 800 1600 2400 3200 4000 SE +/- 16.31, N = 3 SE +/- 27.86, N = 3 SE +/- 28.17, N = 3 SE +/- 33.81, N = 3 SE +/- 15.03, N = 3 3695.4 3036.5 3042.0 3270.6 3265.2 MIN: 3664.8 / MAX: 5363.87 MIN: 2985.79 / MAX: 4135.64 MIN: 3011.87 / MAX: 4321.7 MIN: 3216.47 / MAX: 4638.43 MIN: 3236.75 / MAX: 4741.82
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.11 Video Input: Bosphorus 4K - Video Preset: Faster Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 11.77 13.08 13.14 17.32 17.44 1. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
JPEG-XL libjxl The JPEG XL Image Coding System is designed to provide next-generation JPEG image capabilities with JPEG XL offering better image quality and compression over legacy JPEG. This test profile is currently focused on the multi-threaded JPEG XL image encode performance using the reference libjxl library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better JPEG-XL libjxl 0.10.1 Input: JPEG - Quality: 90 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.09, N = 3 SE +/- 0.20, N = 3 SE +/- 0.30, N = 3 SE +/- 0.27, N = 3 SE +/- 0.37, N = 10 32.51 36.32 37.18 46.77 48.59 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 8 - Buffer Length: 256 - Filter Length: 512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30M 60M 90M 120M 150M SE +/- 1116779.09, N = 8 SE +/- 368435.13, N = 3 SE +/- 612490.36, N = 3 SE +/- 1579485.68, N = 4 SE +/- 292137.71, N = 3 130002500 151906667 150576667 147935000 149903333 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
FFmpeg This is a benchmark of the FFmpeg multimedia framework. The FFmpeg test profile is making use of a modified version of vbench from Columbia University's Architecture and Design Lab (ARCADE) [http://arcade.cs.columbia.edu/vbench/] that is a benchmark for video-as-a-service workloads. The test profile offers the options of a range of vbench scenarios based on freely distributable video content and offers the options of using the x264 or x265 video encoders for transcoding. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better FFmpeg 6.1 Encoder: libx265 - Scenario: Live Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 1.39, N = 3 SE +/- 0.16, N = 3 SE +/- 0.17, N = 3 SE +/- 0.58, N = 3 SE +/- 0.71, N = 3 178.15 179.98 179.99 186.71 186.42 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
Llama.cpp Llama.cpp is a port of Facebook's LLaMA model in C/C++ developed by Georgi Gerganov. Llama.cpp allows the inference of LLaMA and other supported models in C/C++. For CPU inference Llama.cpp supports AVX2/AVX-512, ARM NEON, and other modern ISAs along with features like OpenBLAS usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Tokens Per Second, More Is Better Llama.cpp b1808 Model: llama-2-7b.Q4_0.gguf Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.18, N = 4 SE +/- 0.12, N = 6 SE +/- 0.11, N = 3 SE +/- 0.02, N = 3 SE +/- 0.12, N = 3 16.07 12.73 12.54 13.12 13.82 1. (CXX) g++ options: -std=c++11 -fPIC -O3 -pthread -march=native -mtune=native -lopenblas
Java SciMark This test runs the Java version of SciMark 2, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This benchmark is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.2 Computational Test: Composite Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1200 2400 3600 4800 6000 SE +/- 52.94, N = 3 SE +/- 56.78, N = 3 SE +/- 42.37, N = 3 SE +/- 46.58, N = 8 SE +/- 35.75, N = 3 4311.80 5519.55 5476.91 5321.50 5441.20
Renaissance Renaissance is a suite of benchmarks designed to test the Java JVM from Apache Spark to a Twitter-like service to Scala and other features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Renaissance 0.14 Test: Finagle HTTP Requests Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 25.68, N = 12 SE +/- 7.85, N = 3 SE +/- 22.75, N = 3 SE +/- 14.49, N = 3 SE +/- 15.93, N = 3 1668.7 1925.0 1963.8 2288.1 2300.7 MIN: 1155.42 / MAX: 1849.98 MIN: 1772.86 / MAX: 2146.59 MIN: 1776.88 / MAX: 2143.78 MIN: 2126.73 / MAX: 2373.63 MIN: 2117.57 / MAX: 2419.92
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.3 Binary: Pathtracer ISPC - Model: Crown Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7 14 21 28 35 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.09, N = 3 SE +/- 0.02, N = 3 13.39 16.33 16.46 30.79 31.37 MIN: 13.27 / MAX: 16.49 MIN: 16.12 / MAX: 16.74 MIN: 16.24 / MAX: 16.97 MIN: 30.28 / MAX: 31.99 MIN: 30.96 / MAX: 32.3
QuantLib QuantLib is an open-source library/framework around quantitative finance for modeling, trading and risk management scenarios. QuantLib is written in C++ with Boost and its built-in benchmark used reports the QuantLib Benchmark Index benchmark score. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MFLOPS, More Is Better QuantLib 1.32 Configuration: Multi-Threaded Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 294.52, N = 3 SE +/- 56.56, N = 3 SE +/- 49.82, N = 3 SE +/- 103.02, N = 3 SE +/- 169.22, N = 3 37214.4 44058.2 43490.7 83133.5 83365.1 1. (CXX) g++ options: -O3 -march=native -fPIE -pie
NAMD NAMD is a parallel molecular dynamics code designed for high-performance simulation of large biomolecular systems. NAMD was developed by the Theoretical and Computational Biophysics Group in the Beckman Institute for Advanced Science and Technology at the University of Illinois at Urbana-Champaign. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ns/day, More Is Better NAMD 3.0b6 Input: ATPase with 327,506 Atoms Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.6274 1.2548 1.8822 2.5096 3.137 SE +/- 0.00164, N = 3 SE +/- 0.00243, N = 3 SE +/- 0.00212, N = 3 SE +/- 0.01282, N = 3 SE +/- 0.03311, N = 4 0.65977 1.73454 1.73764 2.74999 2.78846
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 32 - Model: GoogLeNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.10, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.12, N = 3 SE +/- 0.14, N = 3 78.79 92.82 93.71 131.79 135.30
miniFE MiniFE Finite Element is an application for unstructured implicit finite element codes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org CG Mflops, More Is Better miniFE 2.2 Problem Size: Small Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2K 4K 6K 8K 10K SE +/- 8.52, N = 3 SE +/- 7.52, N = 3 SE +/- 3.31, N = 3 SE +/- 16.40, N = 3 SE +/- 8.13, N = 3 9869.74 7278.87 7325.10 7330.07 7673.66 1. (CXX) g++ options: -O3 -fopenmp -lmpi_cxx -lmpi
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs (and GPUs via SYCL) and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Embree 4.3 Binary: Pathtracer ISPC - Model: Asian Dragon Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7 14 21 28 35 SE +/- 0.12, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 16.58 18.35 18.45 31.58 32.22 MIN: 16.29 / MAX: 19.34 MIN: 18.21 / MAX: 18.8 MIN: 18.3 / MAX: 18.87 MIN: 31.27 / MAX: 32.45 MIN: 31.88 / MAX: 33.23
Apache Hadoop This is a benchmark of the Apache Hadoop making use of its built-in name-node throughput benchmark (NNThroughputBenchmark). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops per sec, More Is Better Apache Hadoop 3.3.6 Operation: Create - Threads: 500 - Files: 100000 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 17K 34K 51K 68K 85K SE +/- 717.72, N = 3 SE +/- 761.92, N = 3 SE +/- 465.56, N = 3 SE +/- 488.73, N = 3 SE +/- 682.82, N = 3 76956 76371 77585 71080 70436
Timed Wasmer Compilation This test times how long it takes to compile Wasmer. Wasmer is written in the Rust programming language and is a WebAssembly runtime implementation that supports WASI and EmScripten. This test profile builds Wasmer with the Cranelift and Singlepast compiler features enabled. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Wasmer Compilation 2.3 Time To Compile Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.23, N = 3 SE +/- 0.18, N = 3 SE +/- 0.23, N = 3 SE +/- 0.10, N = 3 SE +/- 0.13, N = 3 38.13 36.55 36.90 29.89 29.44 1. (CC) gcc options: -m64 -ldl -lgcc_s -lutil -lrt -lpthread -lm -lc -pie -nodefaultlibs
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: System V Message Passing Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4M 8M 12M 16M 20M SE +/- 185878.60, N = 5 SE +/- 6257.32, N = 3 SE +/- 7976.77, N = 3 SE +/- 20135.83, N = 3 SE +/- 18408.29, N = 3 17858353.80 12527831.36 12445240.76 16207347.12 16150369.04 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Ringcoin Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 800 1600 2400 3200 4000 SE +/- 3.73, N = 3 SE +/- 11.68, N = 3 SE +/- 4.79, N = 3 SE +/- 37.02, N = 5 SE +/- 20.59, N = 3 1437.47 1868.89 1851.25 3541.50 3500.13 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Eclipse Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1500 3000 4500 6000 7500 SE +/- 61.34, N = 8 SE +/- 12.50, N = 3 SE +/- 37.34, N = 3 SE +/- 28.45, N = 3 SE +/- 26.61, N = 3 6943 6483 6410 6627 6587
miniBUDE MiniBUDE is a mini application for the the core computation of the Bristol University Docking Engine (BUDE). This test profile currently makes use of the OpenMP implementation of miniBUDE for CPU benchmarking. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Billion Interactions/s, More Is Better miniBUDE 20210901 Implementation: OpenMP - Input Deck: BM1 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.06, N = 3 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 SE +/- 0.17, N = 4 SE +/- 0.03, N = 4 12.45 26.37 26.29 49.81 49.95 1. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm
OpenBenchmarking.org GFInst/s, More Is Better miniBUDE 20210901 Implementation: OpenMP - Input Deck: BM1 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 1.56, N = 3 SE +/- 2.35, N = 3 SE +/- 2.56, N = 3 SE +/- 4.15, N = 4 SE +/- 0.84, N = 4 311.31 659.13 657.35 1245.30 1248.81 1. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm
SVT-VP9 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: Visual Quality Optimized - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.47, N = 15 SE +/- 0.50, N = 5 SE +/- 0.56, N = 5 SE +/- 1.37, N = 15 SE +/- 1.42, N = 15 59.56 61.45 61.90 97.12 100.00 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 64 - Model: AlexNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 0.21, N = 3 SE +/- 0.21, N = 3 SE +/- 0.21, N = 3 SE +/- 0.14, N = 3 SE +/- 0.34, N = 3 173.78 205.32 206.40 316.11 322.53
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Magi Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 140 280 420 560 700 SE +/- 0.39, N = 3 SE +/- 0.10, N = 3 SE +/- 0.30, N = 3 SE +/- 7.55, N = 4 SE +/- 0.26, N = 3 278.09 332.40 330.62 644.74 633.44 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
JPEG-XL libjxl The JPEG XL Image Coding System is designed to provide next-generation JPEG image capabilities with JPEG XL offering better image quality and compression over legacy JPEG. This test profile is currently focused on the multi-threaded JPEG XL image encode performance using the reference libjxl library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better JPEG-XL libjxl 0.10.1 Input: PNG - Quality: 90 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.19, N = 3 SE +/- 0.21, N = 3 SE +/- 0.36, N = 3 SE +/- 0.68, N = 3 SE +/- 0.28, N = 3 31.81 36.04 36.02 47.23 48.53 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.11 Video Input: Bosphorus 1080p - Video Preset: Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.06, N = 3 SE +/- 0.06, N = 3 SE +/- 0.08, N = 3 SE +/- 0.15, N = 3 SE +/- 0.07, N = 3 15.46 18.90 18.77 23.00 23.19 1. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PUSCH Processor Benchmark, Throughput Thread Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 7.12, N = 15 SE +/- 0.00, N = 4 SE +/- 0.00, N = 4 SE +/- 0.00, N = 4 SE +/- 0.00, N = 4 153.6 183.5 183.6 183.6 183.5 MIN: 105.1 / MAX: 182.1 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 105.9 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 106 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 106 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 105.9 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -fno-trapping-math -fno-math-errno -ldl
Aircrack-ng Aircrack-ng is a tool for assessing WiFi/WLAN network security. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org k/s, More Is Better Aircrack-ng 1.7 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 17.95, N = 3 SE +/- 65.97, N = 3 SE +/- 45.75, N = 3 SE +/- 194.99, N = 3 SE +/- 153.97, N = 3 35441.71 48803.00 48527.82 92030.35 92259.95 1. (CXX) g++ options: -std=gnu++17 -O3 -fvisibility=hidden -fcommon -rdynamic -lnl-3 -lnl-genl-3 -lpcre -lsqlite3 -lpthread -lz -lssl -lcrypto -lhwloc -ldl -lm -pthread
John The Ripper This is a benchmark of John The Ripper, which is a password cracker. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: WPA PSK Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 135.70, N = 3 SE +/- 245.32, N = 3 SE +/- 221.78, N = 3 SE +/- 573.94, N = 3 SE +/- 425.10, N = 3 29560 85712 85532 160805 160664 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: bcrypt Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9K 18K 27K 36K 45K SE +/- 127.68, N = 3 SE +/- 45.41, N = 3 SE +/- 43.32, N = 3 SE +/- 61.43, N = 3 SE +/- 104.20, N = 3 18787 23271 23235 43512 43695 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: Blowfish Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9K 18K 27K 36K 45K SE +/- 68.54, N = 3 SE +/- 121.86, N = 3 SE +/- 44.47, N = 3 SE +/- 344.23, N = 3 SE +/- 39.87, N = 3 18841 23174 23243 42924 43795 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 16 - Buffer Length: 256 - Filter Length: 512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 60M 120M 180M 240M 300M SE +/- 1245404.89, N = 3 SE +/- 176162.80, N = 3 SE +/- 75055.53, N = 3 SE +/- 2047567.66, N = 3 SE +/- 520000.00, N = 3 159270000 212960000 211670000 286960000 292390000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Garlicoin Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 700 1400 2100 2800 3500 SE +/- 4.51, N = 3 SE +/- 4.36, N = 3 SE +/- 1.43, N = 3 SE +/- 1.95, N = 3 SE +/- 4.61, N = 3 1356.31 1691.29 1674.99 3196.42 3179.32 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Vector Floating Point Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 13K 26K 39K 52K 65K SE +/- 378.33, N = 3 SE +/- 404.29, N = 3 SE +/- 415.80, N = 3 SE +/- 131.10, N = 3 SE +/- 261.75, N = 3 28846.06 30715.14 30936.50 59053.01 59037.14 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Matrix 3D Math Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 0.40, N = 3 SE +/- 3.38, N = 3 SE +/- 2.55, N = 3 SE +/- 1.97, N = 3 SE +/- 0.58, N = 3 2778.89 1953.52 2029.45 1924.70 2011.17 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 4 - Buffer Length: 256 - Filter Length: 512 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20M 40M 60M 80M 100M SE +/- 493083.61, N = 3 SE +/- 444695.78, N = 3 SE +/- 342214.52, N = 3 SE +/- 302227.47, N = 3 SE +/- 190528.21, N = 3 79553667 77811000 77508667 76415667 76809000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Floating Point Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2K 4K 6K 8K 10K SE +/- 7.41, N = 3 SE +/- 2.36, N = 3 SE +/- 6.00, N = 3 SE +/- 8.60, N = 3 SE +/- 15.20, N = 3 4740.28 4283.48 4251.39 8242.96 8199.65 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Vector Shuffle Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4K 8K 12K 16K 20K SE +/- 104.90, N = 3 SE +/- 3.63, N = 3 SE +/- 2.63, N = 3 SE +/- 12.64, N = 3 SE +/- 12.57, N = 3 9138.29 8606.94 8556.72 16537.83 16452.47 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Fused Multiply-Add Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12M 24M 36M 48M 60M SE +/- 267691.10, N = 3 SE +/- 28572.99, N = 3 SE +/- 252666.10, N = 3 SE +/- 59842.82, N = 3 SE +/- 500794.58, N = 3 29762395.57 30389834.82 29927440.69 57429956.15 56410032.25 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 2 - Buffer Length: 256 - Filter Length: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30M 60M 90M 120M 150M SE +/- 724829.48, N = 3 SE +/- 15275.25, N = 3 SE +/- 13333.33, N = 3 SE +/- 14529.66, N = 3 SE +/- 282921.27, N = 3 127376667 107140000 107146667 103326667 103696667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: scrypt Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 0.08, N = 3 SE +/- 0.23, N = 3 SE +/- 0.10, N = 3 SE +/- 1.89, N = 3 SE +/- 1.34, N = 3 119.86 163.88 162.88 310.84 308.21 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: Wide Vector Math Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200K 400K 600K 800K 1000K SE +/- 373.64, N = 3 SE +/- 313.30, N = 3 SE +/- 205.90, N = 3 SE +/- 1012.30, N = 3 SE +/- 1811.04, N = 3 576480.15 416118.50 413926.82 781906.97 778238.33 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
Liquid-DSP LiquidSDR's Liquid-DSP is a software-defined radio (SDR) digital signal processing library. This test profile runs a multi-threaded benchmark of this SDR/DSP library focused on embedded platform usage. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 4 - Buffer Length: 256 - Filter Length: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50M 100M 150M 200M 250M SE +/- 17320.51, N = 3 SE +/- 1006495.57, N = 3 SE +/- 325695.84, N = 3 SE +/- 452413.28, N = 3 SE +/- 674487.46, N = 3 253450000 212860000 209743333 206306667 205890000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 16 - Buffer Length: 256 - Filter Length: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200M 400M 600M 800M 1000M SE +/- 5153390.47, N = 3 SE +/- 199192.82, N = 3 SE +/- 812656.82, N = 3 SE +/- 1556216.92, N = 3 SE +/- 67412.49, N = 3 711050000 740166667 735566667 787633333 795966667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 8 - Buffer Length: 256 - Filter Length: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90M 180M 270M 360M 450M SE +/- 2917967.33, N = 3 SE +/- 1727319.70, N = 3 SE +/- 388858.73, N = 3 SE +/- 1946723.74, N = 3 SE +/- 2535286.53, N = 3 439150000 411290000 409386667 404010000 404976667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 1 - Buffer Length: 256 - Filter Length: 32 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14M 28M 42M 56M 70M SE +/- 637003.49, N = 3 SE +/- 881.92, N = 3 SE +/- 2645.75, N = 3 SE +/- 469067.51, N = 3 SE +/- 318365.79, N = 3 63236667 53661333 53659000 52493000 53018667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Deepcoin Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2K 4K 6K 8K 10K SE +/- 36.27, N = 3 SE +/- 3.81, N = 3 SE +/- 3.02, N = 3 SE +/- 3.98, N = 3 SE +/- 5.77, N = 3 3294.49 4405.85 4377.74 8421.15 8356.74 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
Stress-NG Stress-NG is a Linux stress tool developed by Colin Ian King. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: AVX-512 VNNI Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600K 1200K 1800K 2400K 3000K SE +/- 5501.16, N = 3 SE +/- 1569.35, N = 3 SE +/- 1225.91, N = 3 SE +/- 3334.21, N = 3 SE +/- 2539.18, N = 3 1125583.01 1376354.00 1371242.64 2572055.58 2571394.82 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.16.04 Test: x86_64 RdRand Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2M 4M 6M 8M 10M SE +/- 7.65, N = 3 SE +/- 1009.93, N = 3 SE +/- 788.85, N = 3 SE +/- 980.70, N = 3 SE +/- 6736.08, N = 3 81538.34 4400145.01 4374980.16 8497821.55 8469095.71 1. (CXX) g++ options: -lm -laio -lapparmor -latomic -lc -lcrypt -ldl -lEGL -ljpeg -lmpfr -lpthread -lrt -lsctp -lz
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: LBC, LBRY Credits Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3K 6K 9K 12K 15K SE +/- 15.18, N = 3 SE +/- 4.08, N = 3 SE +/- 7.26, N = 3 SE +/- 47.02, N = 3 SE +/- 5.77, N = 3 6037.26 8437.93 8389.15 16317.00 16120.00 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Skeincoin Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7K 14K 21K 28K 35K SE +/- 139.08, N = 3 SE +/- 5.77, N = 3 SE +/- 39.30, N = 3 SE +/- 17.64, N = 3 SE +/- 11.55, N = 3 13933 18060 17983 34753 34640 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Triple SHA-256, Onecoin Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20K 40K 60K 80K 100K SE +/- 3.33, N = 3 SE +/- 5.77, N = 3 SE +/- 8.82, N = 3 SE +/- 73.33, N = 3 SE +/- 146.40, N = 3 38847 51370 51087 97753 97340 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Quad SHA-256, Pyrite Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 13K 26K 39K 52K 65K SE +/- 3.33, N = 3 SE +/- 5.77, N = 3 SE +/- 78.81, N = 3 SE +/- 297.56, N = 3 SE +/- 331.71, N = 3 26527 32010 31903 61347 61360 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 23.5 Algorithm: Blake-2 S Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30K 60K 90K 120K 150K SE +/- 460.90, N = 3 SE +/- 74.24, N = 3 SE +/- 29.06, N = 3 SE +/- 125.03, N = 3 SE +/- 99.39, N = 3 56160 74883 74513 143480 142753 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PUSCH Processor Benchmark, Throughput Total Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 9.26, N = 15 SE +/- 17.52, N = 15 SE +/- 0.10, N = 6 SE +/- 0.04, N = 4 SE +/- 0.05, N = 4 1068.9 1517.4 1470.5 1611.5 1611.5 MIN: 637.2 / MAX: 1111.5 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 849.7 / MAX: 1610.7 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 849.1 / MAX: 1470.7 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 1019.3 / MAX: 1611.6 -mavx512f -mavx512cd -mavx512bw -mavx512dq - MIN: 1019.4 / MAX: 1611.6 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -fno-trapping-math -fno-math-errno -ldl
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.2.4 Encode Settings: Quality 100, Lossless, Highest Compression Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.198 0.396 0.594 0.792 0.99 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 0.88 0.84 0.84 0.82 0.82 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 4 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.024, N = 3 SE +/- 0.014, N = 3 SE +/- 0.016, N = 3 SE +/- 0.016, N = 3 SE +/- 0.009, N = 3 4.348 5.711 5.742 7.831 7.838 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
Llamafile Mozilla's Llamafile allows distributing and running large language models (LLMs) as a single file. Llamafile aims to make open-source LLMs more accessible to developers and users. Llamafile supports a variety of models, CPUs and GPUs, and other options. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Tokens Per Second, More Is Better Llamafile 0.6 Test: llava-v1.5-7b-q4 - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 13.32 10.57 10.48 10.90 11.37
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: SP.B Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4K 8K 12K 16K 20K SE +/- 44.80, N = 3 SE +/- 48.58, N = 3 SE +/- 12.84, N = 3 SE +/- 211.58, N = 3 SE +/- 253.63, N = 3 12750.97 10661.66 11152.31 18360.82 18660.90 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 13 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.99, N = 15 SE +/- 1.96, N = 15 SE +/- 2.01, N = 15 SE +/- 4.13, N = 15 SE +/- 4.27, N = 15 104.61 121.44 122.91 160.65 165.29 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
Y-Cruncher Y-Cruncher is a multi-threaded Pi benchmark capable of computing Pi to trillions of digits. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Y-Cruncher 0.8.3 Pi Digits To Calculate: 1B Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 8 16 24 32 40 SE +/- 0.21, N = 3 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 33.89 25.24 25.46 18.80 18.21
JPEG-XL libjxl The JPEG XL Image Coding System is designed to provide next-generation JPEG image capabilities with JPEG XL offering better image quality and compression over legacy JPEG. This test profile is currently focused on the multi-threaded JPEG XL image encode performance using the reference libjxl library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better JPEG-XL libjxl 0.10.1 Input: JPEG - Quality: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 13.48 16.58 16.52 26.96 27.22 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
OpenBenchmarking.org MP/s, More Is Better JPEG-XL libjxl 0.10.1 Input: PNG - Quality: 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 7 14 21 28 35 SE +/- 0.07, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 13.65 16.69 16.66 27.56 27.75 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Tradesoap Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 16.67, N = 4 SE +/- 7.41, N = 4 SE +/- 20.46, N = 5 SE +/- 16.84, N = 13 SE +/- 17.19, N = 4 1914 1964 1979 2439 2474
Timed Mesa Compilation This test profile times how long it takes to compile Mesa with Meson/Ninja. For minimizing build dependencies and avoid versioning conflicts, test this is just the core Mesa build without LLVM or the extra Gallium3D/Mesa drivers enabled. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed Mesa Compilation 24.0 Time To Compile Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 8 16 24 32 40 SE +/- 0.07, N = 3 SE +/- 0.02, N = 3 SE +/- 0.11, N = 3 SE +/- 0.06, N = 3 SE +/- 0.21, N = 3 32.52 28.20 28.26 17.73 17.46
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 4K - Video Preset: Very Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 10 20 30 40 50 SE +/- 0.06, N = 3 SE +/- 0.03, N = 3 SE +/- 0.05, N = 3 SE +/- 0.07, N = 4 SE +/- 0.08, N = 4 18.97 25.32 25.25 42.20 42.60
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: H2 Database Engine Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 23.84, N = 3 SE +/- 12.68, N = 3 SE +/- 7.51, N = 3 SE +/- 9.17, N = 3 SE +/- 13.96, N = 3 2128 2000 1977 1872 1826
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Apache Lucene Search Engine Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 700 1400 2100 2800 3500 SE +/- 50.09, N = 15 SE +/- 17.06, N = 5 SE +/- 18.64, N = 5 SE +/- 9.20, N = 5 SE +/- 9.53, N = 5 3113 2404 2409 2083 2080
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 4K - Video Preset: Ultra Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.30, N = 10 SE +/- 0.05, N = 4 SE +/- 0.04, N = 4 SE +/- 0.08, N = 6 SE +/- 0.07, N = 6 39.59 52.25 52.30 80.50 81.23 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: EP.C Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 55.41, N = 15 SE +/- 4.34, N = 6 SE +/- 41.08, N = 15 SE +/- 18.71, N = 8 SE +/- 13.01, N = 8 1470.81 1101.23 1169.68 2117.29 2123.37 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
ASTC Encoder ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MT/s, More Is Better ASTC Encoder 4.7 Preset: Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 0.04, N = 3 SE +/- 0.35, N = 3 SE +/- 0.29, N = 3 SE +/- 0.65, N = 4 SE +/- 0.54, N = 4 205.93 204.84 203.10 388.63 392.24 1. (CXX) g++ options: -O3 -flto -pthread
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 32 - Model: AlexNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 1.37, N = 3 SE +/- 0.05, N = 3 SE +/- 0.23, N = 3 SE +/- 0.14, N = 4 SE +/- 0.14, N = 4 160.52 171.11 172.89 236.71 242.49
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 4K - Video Preset: Super Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.08, N = 3 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 SE +/- 0.05, N = 4 SE +/- 0.07, N = 4 20.17 26.73 26.65 45.87 46.41
LULESH LULESH is the Livermore Unstructured Lagrangian Explicit Shock Hydrodynamics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org z/s, More Is Better LULESH 2.0.3 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2K 4K 6K 8K 10K SE +/- 210.25, N = 15 SE +/- 39.33, N = 5 SE +/- 55.23, N = 6 SE +/- 15.60, N = 6 SE +/- 57.24, N = 12 8126.32 6454.67 6744.86 8006.35 8106.15 1. (CXX) g++ options: -O3 -fopenmp -lm -lmpi_cxx -lmpi
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: jMonkeyEngine Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1500 3000 4500 6000 7500 SE +/- 4.67, N = 3 SE +/- 0.88, N = 3 SE +/- 0.33, N = 3 SE +/- 1.20, N = 3 SE +/- 0.67, N = 3 6827 6804 6802 6808 6806
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: Deconvolution Batch shapes_1d - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.02381, N = 3 SE +/- 0.01491, N = 3 SE +/- 0.00421, N = 3 SE +/- 0.00488, N = 3 SE +/- 0.01295, N = 3 6.85318 3.87665 3.91089 2.98654 3.00636 MIN: 3.93 MIN: 3.48 MIN: 3.5 MIN: 2.41 MIN: 2.4 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
7-Zip Compression This is a test of 7-Zip compression/decompression with its integrated benchmark feature. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 22.01 Test: Decompression Rating Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40K 80K 120K 160K 200K SE +/- 900.25, N = 3 SE +/- 835.95, N = 3 SE +/- 299.70, N = 3 SE +/- 110.28, N = 3 SE +/- 369.09, N = 3 61527 87804 88945 168957 170373 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 22.01 Test: Compression Rating Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40K 80K 120K 160K 200K SE +/- 436.66, N = 3 SE +/- 118.34, N = 3 SE +/- 195.02, N = 3 SE +/- 227.16, N = 3 SE +/- 313.05, N = 3 94990 101849 103571 167186 170888 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
SVT-VP9 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: VMAF Optimized - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.46, N = 15 SE +/- 0.09, N = 5 SE +/- 0.09, N = 5 SE +/- 0.05, N = 6 SE +/- 0.15, N = 6 79.85 70.29 71.48 99.84 103.52 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 4K - Video Preset: Super Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14 28 42 56 70 SE +/- 0.31, N = 4 SE +/- 0.02, N = 4 SE +/- 0.05, N = 4 SE +/- 0.07, N = 5 SE +/- 0.07, N = 5 28.64 38.85 38.82 62.23 62.56 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Spring Boot Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 300 600 900 1200 1500 SE +/- 6.95, N = 5 SE +/- 7.11, N = 5 SE +/- 6.93, N = 5 SE +/- 10.62, N = 13 SE +/- 8.13, N = 5 1220 1159 1162 1554 1554
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: IP Shapes 1D - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.6557 1.3114 1.9671 2.6228 3.2785 SE +/- 0.01544, N = 4 SE +/- 0.00306, N = 4 SE +/- 0.00536, N = 4 SE +/- 0.00498, N = 4 SE +/- 0.00269, N = 4 2.91408 1.98773 1.99986 1.13203 1.13000 MIN: 2.51 MIN: 1.84 MIN: 1.85 MIN: 1.03 MIN: 1.03 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
Z3 Theorem Prover The Z3 Theorem Prover / SMT solver is developed by Microsoft Research under the MIT license. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Z3 Theorem Prover 4.12.1 SMT File: 1.smt2 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.08, N = 3 SE +/- 0.05, N = 3 SE +/- 0.11, N = 3 SE +/- 0.16, N = 3 SE +/- 0.15, N = 3 18.66 20.50 20.43 20.78 20.33 1. (CXX) g++ options: -lpthread -std=c++17 -fvisibility=hidden -mfpmath=sse -msse -msse2 -O3 -fPIC
QuantLib QuantLib is an open-source library/framework around quantitative finance for modeling, trading and risk management scenarios. QuantLib is written in C++ with Boost and its built-in benchmark used reports the QuantLib Benchmark Index benchmark score. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MFLOPS, More Is Better QuantLib 1.32 Configuration: Single-Threaded Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1100 2200 3300 4400 5500 SE +/- 34.53, N = 4 SE +/- 43.08, N = 3 SE +/- 17.50, N = 3 SE +/- 56.37, N = 3 SE +/- 46.95, N = 3 5019.2 4518.4 4457.8 4425.0 4425.2 1. (CXX) g++ options: -O3 -march=native -fPIE -pie
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 4K - Video Preset: Ultra Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 12 24 36 48 60 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 SE +/- 0.03, N = 5 SE +/- 0.08, N = 5 24.69 32.09 31.86 54.07 54.54
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Apache Tomcat Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1400 2800 4200 5600 7000 SE +/- 14.71, N = 3 SE +/- 2.73, N = 3 SE +/- 9.21, N = 3 SE +/- 4.40, N = 4 SE +/- 2.95, N = 4 6687 6467 6482 3367 3370
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 4K - Video Preset: Very Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.19, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 4 SE +/- 0.09, N = 4 23.05 31.05 31.04 47.67 48.11 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Algebraic Multi-Grid Benchmark AMG is a parallel algebraic multigrid solver for linear systems arising from problems on unstructured grids. The driver provided with AMG builds linear systems for various 3-dimensional problems. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Figure Of Merit, More Is Better Algebraic Multi-Grid Benchmark 1.2 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 100M 200M 300M 400M 500M SE +/- 269181.94, N = 5 SE +/- 333950.91, N = 4 SE +/- 445215.93, N = 4 SE +/- 201300.38, N = 3 SE +/- 143363.91, N = 3 453703880 350656750 352361775 362574433 378697000 1. (CC) gcc options: -lparcsr_ls -lparcsr_mv -lseq_mv -lIJ_mv -lkrylov -lHYPRE_utilities -lm -fopenmp -lmpi
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Tradebeans Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 700 1400 2100 2800 3500 SE +/- 14.57, N = 4 SE +/- 31.95, N = 4 SE +/- 22.67, N = 4 SE +/- 38.01, N = 4 SE +/- 21.46, N = 4 3410 2775 2815 3359 3329
Parallel BZIP2 Compression This test measures the time needed to compress a file (FreeBSD-13.0-RELEASE-amd64-memstick.img) using Parallel BZIP2 compression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Parallel BZIP2 Compression 1.1.13 FreeBSD-13.0-RELEASE-amd64-memstick.img Compression Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.044479, N = 6 SE +/- 0.045334, N = 15 SE +/- 0.045809, N = 15 SE +/- 0.023270, N = 9 SE +/- 0.025811, N = 10 6.612486 6.176853 6.154460 3.325543 3.299879 1. (CXX) g++ options: -O2 -pthread -lbz2 -lpthread
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: PMD Source Code Analyzer Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 8.96, N = 15 SE +/- 9.22, N = 15 SE +/- 9.55, N = 7 SE +/- 10.09, N = 7 SE +/- 5.68, N = 7 1122 1142 1113 1074 1053
LuxCoreRender LuxCoreRender is an open-source 3D physically based renderer formerly known as LuxRender. LuxCoreRender supports CPU-based rendering as well as GPU acceleration via OpenCL, NVIDIA CUDA, and NVIDIA OptiX interfaces. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.6 Scene: Rainbow Colors and Prism - Acceleration: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.03, N = 3 SE +/- 0.01, N = 4 SE +/- 0.04, N = 4 SE +/- 0.11, N = 5 SE +/- 0.06, N = 5 9.35 11.89 11.93 16.94 16.98 MIN: 8.52 / MAX: 10.72 MIN: 11.05 / MAX: 12 MIN: 11.1 / MAX: 12.12 MIN: 15.4 / MAX: 17.31 MIN: 15.46 / MAX: 17.3
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: BioJava Biological Data Framework Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 900 1800 2700 3600 4500 SE +/- 25.22, N = 4 SE +/- 19.50, N = 4 SE +/- 20.56, N = 4 SE +/- 15.83, N = 4 SE +/- 33.62, N = 4 3708 4347 4327 4413 4425
Y-Cruncher Y-Cruncher is a multi-threaded Pi benchmark capable of computing Pi to trillions of digits. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Y-Cruncher 0.8.3 Pi Digits To Calculate: 500M Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.124, N = 4 SE +/- 0.040, N = 4 SE +/- 0.036, N = 4 SE +/- 0.013, N = 5 SE +/- 0.010, N = 5 14.926 11.555 11.589 8.920 8.652
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: MG.C Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5K 10K 15K 20K 25K SE +/- 532.70, N = 12 SE +/- 158.41, N = 5 SE +/- 213.35, N = 5 SE +/- 94.12, N = 5 SE +/- 161.84, N = 5 24688.58 19566.67 20265.52 20223.60 20961.35 1. (F9X) gfortran options: -O3 -march=native -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz 2. Open MPI 4.1.6
VVenC VVenC is the Fraunhofer Versatile Video Encoder as a fast/efficient H.266/VVC encoder. The vvenc encoder makes use of SIMD Everywhere (SIMDe). The vvenc software is published under the Clear BSD License. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better VVenC 1.11 Video Input: Bosphorus 1080p - Video Preset: Faster Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 11 22 33 44 55 SE +/- 0.25, N = 4 SE +/- 0.05, N = 4 SE +/- 0.04, N = 4 SE +/- 0.09, N = 4 SE +/- 0.11, N = 4 40.07 44.58 45.14 50.41 50.91 1. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
Dolfyn Dolfyn is a Computational Fluid Dynamics (CFD) code of modern numerical simulation techniques. The Dolfyn test profile measures the execution time of the bundled computational fluid dynamics demos that are bundled with Dolfyn. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Dolfyn 0.527 Computational Fluid Dynamics Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.03, N = 5 SE +/- 0.04, N = 5 SE +/- 0.02, N = 5 SE +/- 0.06, N = 5 SE +/- 0.11, N = 5 10.42 10.63 10.73 10.93 10.96
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Jython Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 800 1600 2400 3200 4000 SE +/- 14.73, N = 5 SE +/- 14.78, N = 4 SE +/- 16.55, N = 4 SE +/- 15.77, N = 4 SE +/- 34.16, N = 4 2433 3790 3762 3795 3871
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 1080p - Video Preset: Medium Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14 28 42 56 70 SE +/- 0.14, N = 4 SE +/- 0.07, N = 4 SE +/- 0.03, N = 4 SE +/- 0.05, N = 5 SE +/- 0.07, N = 5 36.59 45.75 45.63 63.89 64.16
ASTC Encoder ASTC Encoder (astcenc) is for the Adaptive Scalable Texture Compression (ASTC) format commonly used with OpenGL, OpenGL ES, and Vulkan graphics APIs. This test profile does a coding test of both compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MT/s, More Is Better ASTC Encoder 4.7 Preset: Medium Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.00, N = 3 SE +/- 0.05, N = 4 SE +/- 0.04, N = 4 SE +/- 0.09, N = 5 SE +/- 0.18, N = 5 89.14 80.53 79.60 153.70 154.72 1. (CXX) g++ options: -O3 -flto -pthread
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Apache Lucene Search Index Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 19.22, N = 3 SE +/- 17.23, N = 3 SE +/- 18.27, N = 3 SE +/- 11.06, N = 3 SE +/- 12.68, N = 3 2296 2407 2390 2517 2471
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.2.4 Encode Settings: Quality 100, Lossless Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.4883 0.9766 1.4649 1.9532 2.4415 SE +/- 0.01, N = 5 SE +/- 0.00, N = 5 SE +/- 0.01, N = 5 SE +/- 0.01, N = 4 SE +/- 0.00, N = 4 2.17 2.16 2.17 2.08 2.09 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
x265 This is a simple test of the x265 encoder run on the CPU with 1080p and 4K options for H.265 video encode performance with x265. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better x265 3.4 Video Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 0.77, N = 15 SE +/- 0.38, N = 6 SE +/- 0.53, N = 6 SE +/- 0.13, N = 7 SE +/- 0.21, N = 7 84.33 89.23 89.12 117.25 118.06 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Apache Kafka Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1100 2200 3300 4400 5500 SE +/- 2.00, N = 3 SE +/- 29.48, N = 3 SE +/- 29.67, N = 3 SE +/- 27.06, N = 3 SE +/- 32.17, N = 3 4984 5031 5027 5030 5055
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 8 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 16 32 48 64 80 SE +/- 0.30, N = 3 SE +/- 0.02, N = 4 SE +/- 0.10, N = 4 SE +/- 0.25, N = 5 SE +/- 0.15, N = 5 36.14 46.84 47.29 71.14 72.74 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
easyWave The easyWave software allows simulating tsunami generation and propagation in the context of early warning systems. EasyWave supports making use of OpenMP for CPU multi-threading and there are also GPU ports available but not currently incorporated as part of this test profile. The easyWave tsunami generation software is run with one of the example/reference input files for measuring the CPU execution time. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better easyWave r34 Input: e2Asean Grid + BengkuluSept2007 Source - Time: 240 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.089, N = 15 SE +/- 0.009, N = 6 SE +/- 0.030, N = 6 SE +/- 0.008, N = 8 SE +/- 0.010, N = 9 6.278 6.166 5.990 3.164 3.037 1. (CXX) g++ options: -O3 -fopenmp
PyBench This test profile reports the total time of the different average timed test results from PyBench. PyBench reports average test times for different functions such as BuiltinFunctionCalls and NestedForLoops, with this total result providing a rough estimate as to Python's average performance on a given system. This test profile runs PyBench each time for 20 rounds. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Milliseconds, Fewer Is Better PyBench 2018-02-16 Total For Average Test Times Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 120 240 360 480 600 SE +/- 2.82, N = 5 SE +/- 1.38, N = 4 SE +/- 2.53, N = 4 SE +/- 2.65, N = 4 SE +/- 6.41, N = 4 415 532 527 536 533
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 10 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 1.10, N = 15 SE +/- 0.10, N = 6 SE +/- 0.12, N = 6 SE +/- 0.15, N = 8 SE +/- 0.14, N = 8 101.83 107.65 109.03 153.78 158.94 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
SVT-VP9 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: PSNR/SSIM Optimized - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.64, N = 10 SE +/- 0.14, N = 5 SE +/- 0.08, N = 5 SE +/- 0.14, N = 6 SE +/- 0.13, N = 6 83.19 74.93 76.06 107.37 111.01 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 7 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.08, N = 4 SE +/- 0.07, N = 5 SE +/- 0.04, N = 5 SE +/- 0.08, N = 6 SE +/- 0.06, N = 6 47.98 56.95 57.23 97.70 99.61 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
ASKAP ASKAP is a set of benchmarks from the Australian SKA Pathfinder. The principal ASKAP benchmarks are the Hogbom Clean Benchmark (tHogbomClean) and Convolutional Resamping Benchmark (tConvolve) as well as some previous ASKAP benchmarks being included as well for OpenCL and CUDA execution of tConvolve. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 1.0 Test: tConvolve OpenMP - Degridding Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1100 2200 3300 4400 5500 SE +/- 41.03, N = 5 SE +/- 29.29, N = 6 SE +/- 27.18, N = 8 SE +/- 0.00, N = 6 SE +/- 49.85, N = 6 4282.18 3429.45 3426.03 5023.70 5172.44 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 1.0 Test: tConvolve OpenMP - Gridding Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1100 2200 3300 4400 5500 SE +/- 25.17, N = 5 SE +/- 11.75, N = 6 SE +/- 24.04, N = 8 SE +/- 38.41, N = 6 SE +/- 41.40, N = 6 2556.22 2703.36 2739.34 4886.94 5073.23 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 1 - Model: ResNet-50 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.01, N = 5 SE +/- 0.04, N = 5 SE +/- 0.01, N = 5 SE +/- 0.02, N = 5 SE +/- 0.03, N = 5 13.71 16.77 17.47 14.78 14.93
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: GraphChi Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 18.95, N = 4 SE +/- 18.46, N = 4 SE +/- 21.22, N = 4 SE +/- 10.30, N = 4 SE +/- 14.73, N = 4 2675 2535 2561 2521 2472
Primesieve Primesieve generates prime numbers using a highly optimized sieve of Eratosthenes implementation. Primesieve primarily benchmarks the CPU's L1/L2 cache performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Primesieve 12.1 Length: 1e12 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.302, N = 3 SE +/- 0.018, N = 4 SE +/- 0.007, N = 4 SE +/- 0.008, N = 6 SE +/- 0.007, N = 6 21.095 12.416 12.506 6.669 6.691 1. (CXX) g++ options: -O3
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 12 - Input: Bosphorus 4K Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 1.00, N = 15 SE +/- 0.15, N = 7 SE +/- 0.14, N = 7 SE +/- 0.69, N = 8 SE +/- 0.33, N = 8 105.28 123.64 125.28 164.85 170.21 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 1080p - Video Preset: Very Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.95, N = 15 SE +/- 0.05, N = 7 SE +/- 0.14, N = 7 SE +/- 0.14, N = 9 SE +/- 0.11, N = 9 89.31 112.05 111.78 165.55 166.81
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: IP Shapes 3D - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.01930, N = 5 SE +/- 0.02020, N = 5 SE +/- 0.00833, N = 5 SE +/- 0.00590, N = 5 SE +/- 0.00611, N = 5 7.59695 4.61211 4.67840 4.07056 3.88685 MIN: 7.29 MIN: 4.4 MIN: 4.46 MIN: 3.76 MIN: 3.57 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Avrora AVR Simulation Framework Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 500 1000 1500 2000 2500 SE +/- 2.49, N = 8 SE +/- 3.46, N = 6 SE +/- 3.26, N = 6 SE +/- 21.04, N = 8 SE +/- 20.96, N = 6 1451 2430 2415 2393 2359
libavif avifenc This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 1.0 Encoder Speed: 6 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.088, N = 15 SE +/- 0.007, N = 7 SE +/- 0.014, N = 7 SE +/- 0.018, N = 9 SE +/- 0.010, N = 9 6.260 5.301 5.338 3.447 3.421 1. (CXX) g++ options: -O3 -fPIC -lm
OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 1.0 Encoder Speed: 6, Lossless Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.014, N = 5 SE +/- 0.012, N = 6 SE +/- 0.021, N = 6 SE +/- 0.018, N = 7 SE +/- 0.044, N = 7 8.969 8.140 8.147 5.932 5.890 1. (CXX) g++ options: -O3 -fPIC -lm
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 1080p - Video Preset: Super Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.97, N = 15 SE +/- 0.07, N = 7 SE +/- 0.08, N = 7 SE +/- 0.19, N = 9 SE +/- 0.16, N = 9 96.33 120.30 119.83 180.54 181.83
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 4 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.15, N = 5 SE +/- 0.08, N = 5 SE +/- 0.07, N = 5 SE +/- 0.05, N = 6 SE +/- 0.13, N = 6 15.92 19.96 19.89 23.94 24.17 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 1080p - Video Preset: Medium Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 20 40 60 80 100 SE +/- 0.37, N = 4 SE +/- 0.04, N = 5 SE +/- 0.04, N = 5 SE +/- 0.15, N = 6 SE +/- 0.08, N = 6 50.47 64.80 64.68 87.51 88.30 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 1 - Model: AlexNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 5 10 15 20 25 SE +/- 0.01, N = 6 SE +/- 0.01, N = 5 SE +/- 0.00, N = 5 SE +/- 0.00, N = 5 SE +/- 0.01, N = 5 18.50 14.53 15.01 14.37 14.96
RNNoise RNNoise is a recurrent neural network for audio noise reduction developed by Mozilla and Xiph.Org. This test profile is a single-threaded test measuring the time to denoise a sample 26 minute long 16-bit RAW audio file using this recurrent neural network noise suppression library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better RNNoise 0.2 Input: 26 Minute Long Talking Sample Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.059, N = 7 SE +/- 0.016, N = 6 SE +/- 0.050, N = 6 SE +/- 0.063, N = 6 SE +/- 0.033, N = 6 6.400 7.028 7.184 7.291 7.414 1. (CC) gcc options: -O2 -pedantic -fvisibility=hidden
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 8 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 1.11, N = 15 SE +/- 0.12, N = 8 SE +/- 0.13, N = 8 SE +/- 0.22, N = 9 SE +/- 0.29, N = 9 120.53 147.31 148.34 187.54 189.41 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PDSCH Processor Benchmark, Throughput Total Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3K 6K 9K 12K 15K SE +/- 5.11, N = 5 SE +/- 8.66, N = 7 SE +/- 38.56, N = 7 SE +/- 66.74, N = 6 SE +/- 51.53, N = 6 2908.5 7792.6 7530.0 12647.5 12852.6 -mavx512f -mavx512cd -mavx512bw -mavx512dq -mavx512f -mavx512cd -mavx512bw -mavx512dq -mavx512f -mavx512cd -mavx512bw -mavx512dq -mavx512f -mavx512cd -mavx512bw -mavx512dq 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -fno-trapping-math -fno-math-errno -ldl
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: Convolution Batch Shapes Auto - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.04485, N = 7 SE +/- 0.00879, N = 7 SE +/- 0.00717, N = 7 SE +/- 0.01136, N = 7 SE +/- 0.01415, N = 7 6.38053 8.23220 8.16800 6.72536 6.40719 MIN: 5.92 MIN: 7.77 MIN: 7.7 MIN: 6.2 MIN: 5.93 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
Google Draco Draco is a library developed by Google for compressing/decompressing 3D geometric meshes and point clouds. This test profile uses some Artec3D PLY models as the sample 3D model input formats for Draco compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Google Draco 1.5.6 Model: Church Facade Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1200 2400 3600 4800 6000 SE +/- 17.70, N = 6 SE +/- 12.60, N = 6 SE +/- 10.00, N = 7 SE +/- 31.00, N = 6 SE +/- 29.27, N = 6 5607 5163 5083 5237 5224 1. (CXX) g++ options: -O3
uvg266 uvg266 is an open-source VVC/H.266 (Versatile Video Coding) encoder based on Kvazaar as part of the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better uvg266 0.4.1 Video Input: Bosphorus 1080p - Video Preset: Ultra Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 1.21, N = 15 SE +/- 0.14, N = 8 SE +/- 0.06, N = 8 SE +/- 0.15, N = 9 SE +/- 0.20, N = 9 115.40 143.28 142.97 201.92 203.39
GNU Octave Benchmark This test profile measures how long it takes to complete several reference GNU Octave files via octave-benchmark. GNU Octave is used for numerical computations and is an open-source alternative to MATLAB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better GNU Octave Benchmark 8.4.0 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 2 4 6 8 10 SE +/- 0.011, N = 6 SE +/- 0.016, N = 7 SE +/- 0.017, N = 7 SE +/- 0.043, N = 7 SE +/- 0.021, N = 7 6.462 5.668 5.587 5.754 5.780
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: FOP Print Formatter Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 6.07, N = 15 SE +/- 3.32, N = 15 SE +/- 5.35, N = 15 SE +/- 3.38, N = 15 SE +/- 5.40, N = 15 426 423 414 403 399
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.2.4 Encode Settings: Quality 100, Highest Compression Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.1925 2.385 3.5775 4.77 5.9625 SE +/- 0.02, N = 7 SE +/- 0.03, N = 8 SE +/- 0.03, N = 8 SE +/- 0.03, N = 8 SE +/- 0.01, N = 8 4.73 5.30 5.28 5.15 5.10 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
libavif avifenc This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 1.0 Encoder Speed: 10, Lossless Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.9461 1.8922 2.8383 3.7844 4.7305 SE +/- 0.011, N = 8 SE +/- 0.009, N = 8 SE +/- 0.025, N = 8 SE +/- 0.020, N = 8 SE +/- 0.038, N = 15 4.162 4.202 4.205 3.804 3.841 1. (CXX) g++ options: -O3 -fPIC -lm
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Zxing 1D/2D Barcode Image Processing Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 30.56, N = 15 SE +/- 3.37, N = 8 SE +/- 2.37, N = 8 SE +/- 4.29, N = 9 SE +/- 4.09, N = 9 1053 1016 1019 597 589
Google Draco Draco is a library developed by Google for compressing/decompressing 3D geometric meshes and point clouds. This test profile uses some Artec3D PLY models as the sample 3D model input formats for Draco compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Google Draco 1.5.6 Model: Lion Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 900 1800 2700 3600 4500 SE +/- 9.84, N = 7 SE +/- 5.99, N = 7 SE +/- 12.07, N = 7 SE +/- 18.70, N = 7 SE +/- 17.41, N = 7 3830 3922 3891 4049 3878 1. (CXX) g++ options: -O3
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 1080p - Video Preset: Super Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 50 100 150 200 250 SE +/- 2.29, N = 15 SE +/- 0.11, N = 9 SE +/- 0.09, N = 9 SE +/- 0.29, N = 10 SE +/- 0.26, N = 10 139.72 168.47 168.08 245.56 248.39 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 1080p - Video Preset: Very Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 40 80 120 160 200 SE +/- 0.87, N = 7 SE +/- 0.08, N = 8 SE +/- 0.10, N = 8 SE +/- 0.24, N = 9 SE +/- 0.25, N = 9 103.04 129.55 129.02 184.98 186.56 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Batik SVG Toolkit Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 5.39, N = 8 SE +/- 6.28, N = 8 SE +/- 5.55, N = 8 SE +/- 5.97, N = 8 SE +/- 3.65, N = 8 897 927 930 969 936
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 7 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 2.05, N = 15 SE +/- 0.16, N = 8 SE +/- 0.21, N = 8 SE +/- 0.34, N = 10 SE +/- 0.26, N = 10 168.39 178.68 178.20 307.89 311.95 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
Helsing Helsing is an open-source POSIX vampire number generator. This test profile measures the time it takes to generate vampire numbers between varying numbers of digits. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Helsing 1.0-beta Digit Range: 12 digit Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 0.918 1.836 2.754 3.672 4.59 SE +/- 0.028, N = 15 SE +/- 0.004, N = 8 SE +/- 0.004, N = 8 SE +/- 0.002, N = 10 SE +/- 0.003, N = 11 4.080 3.753 3.828 2.011 2.016 1. (CC) gcc options: -O2 -pthread
TensorFlow This is a benchmark of the TensorFlow deep learning framework using the TensorFlow reference benchmarks (tensorflow/benchmarks with tf_cnn_benchmarks.py). Note with the Phoronix Test Suite there is also pts/tensorflow-lite for benchmarking the TensorFlow Lite binaries if desired for complementary metrics. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 1 - Model: GoogLeNet Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 14 28 42 56 70 SE +/- 0.05, N = 8 SE +/- 0.11, N = 9 SE +/- 0.10, N = 9 SE +/- 0.05, N = 8 SE +/- 0.18, N = 8 51.33 62.02 62.99 57.21 57.34
SVT-VP9 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: Visual Quality Optimized - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 80 160 240 320 400 SE +/- 3.49, N = 15 SE +/- 0.90, N = 9 SE +/- 0.76, N = 9 SE +/- 1.84, N = 15 SE +/- 1.71, N = 15 221.33 201.53 203.16 355.84 364.90 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the Intel oneAPI toolkit. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.4 Harness: Deconvolution Batch shapes_3d - Engine: CPU Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 1.1133 2.2266 3.3399 4.4532 5.5665 SE +/- 0.10597, N = 15 SE +/- 0.00353, N = 9 SE +/- 0.00661, N = 9 SE +/- 0.00342, N = 9 SE +/- 0.00405, N = 9 4.94792 3.79890 3.83412 2.46164 2.46159 MIN: 4.45 MIN: 3.63 MIN: 3.58 MIN: 2.35 MIN: 2.35 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl
LAMMPS Molecular Dynamics Simulator LAMMPS is a classical molecular dynamics code, and an acronym for Large-scale Atomic/Molecular Massively Parallel Simulator. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ns/day, More Is Better LAMMPS Molecular Dynamics Simulator 23Jun2022 Model: Rhodopsin Protein Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 3 6 9 12 15 SE +/- 0.250, N = 15 SE +/- 0.032, N = 9 SE +/- 0.055, N = 9 SE +/- 0.084, N = 15 SE +/- 0.059, N = 11 7.236 6.848 6.826 12.449 12.481 1. (CXX) g++ options: -O3 -lm -ldl
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PDSCH Processor Benchmark, Throughput Thread Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 5.58, N = 9 SE +/- 5.66, N = 10 SE +/- 3.24, N = 10 SE +/- 2.53, N = 9 SE +/- 5.92, N = 10 809.6 1082.7 1061.5 1027.8 1049.1 -mavx512f -mavx512cd -mavx512bw -mavx512dq -mavx512f -mavx512cd -mavx512bw -mavx512dq -mavx512f -mavx512cd -mavx512bw -mavx512dq -mavx512f -mavx512cd -mavx512bw -mavx512dq 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -fno-trapping-math -fno-math-errno -ldl
Kvazaar This is a test of Kvazaar as a CPU-based H.265/HEVC video encoder written in the C programming language and optimized in Assembly. Kvazaar is the winner of the 2016 ACM Open-Source Software Competition and developed at the Ultra Video Group, Tampere University, Finland. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.2 Video Input: Bosphorus 1080p - Video Preset: Ultra Fast Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 70 140 210 280 350 SE +/- 4.30, N = 15 SE +/- 0.17, N = 10 SE +/- 0.14, N = 10 SE +/- 0.21, N = 11 SE +/- 0.21, N = 11 197.51 226.69 226.13 299.12 301.57 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
SVT-VP9 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample YUV input video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: VMAF Optimized - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 0.18, N = 10 SE +/- 0.20, N = 10 SE +/- 0.20, N = 10 SE +/- 0.22, N = 11 SE +/- 0.58, N = 11 289.26 240.66 242.58 398.58 408.42 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.3 Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 90 180 270 360 450 SE +/- 0.53, N = 10 SE +/- 0.36, N = 10 SE +/- 0.28, N = 10 SE +/- 0.27, N = 11 SE +/- 0.42, N = 11 295.69 249.86 251.44 408.36 419.26 1. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm
DaCapo Benchmark This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance of various popular real-world Java workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org msec, Fewer Is Better DaCapo Benchmark 23.11 Java Test: Apache Xalan XSLT Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 110 220 330 440 550 SE +/- 2.76, N = 11 SE +/- 2.21, N = 10 SE +/- 2.71, N = 10 SE +/- 2.09, N = 10 SE +/- 3.47, N = 10 408 448 450 508 506
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.5.0 Tuning: 10 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 120 240 360 480 600 SE +/- 4.53, N = 15 SE +/- 0.26, N = 11 SE +/- 0.24, N = 11 SE +/- 1.05, N = 12 SE +/- 1.04, N = 12 383.64 358.54 361.05 545.31 560.81 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.2.4 Encode Settings: Quality 100 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4 8 12 16 20 SE +/- 0.05, N = 11 SE +/- 0.04, N = 12 SE +/- 0.06, N = 12 SE +/- 0.02, N = 11 SE +/- 0.10, N = 11 15.37 16.59 16.54 15.84 16.11 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
SVT-AV1 This is a benchmark of the SVT-AV1 open-source video encoder/decoder. SVT-AV1 was originally developed by Intel as part of their Open Visual Cloud / Scalable Video Technology (SVT). Development of SVT-AV1 has since moved to the Alliance for Open Media as part of upstream AV1 development. SVT-AV1 is a CPU-based multi-threaded video encoder for the AV1 video format with a sample YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 13 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 160 320 480 640 800 SE +/- 13.14, N = 15 SE +/- 14.78, N = 15 SE +/- 15.50, N = 15 SE +/- 20.45, N = 15 SE +/- 20.65, N = 15 649.43 621.63 633.29 711.53 728.58 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 2.0 Encoder Mode: Preset 12 - Input: Bosphorus 1080p Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 140 280 420 560 700 SE +/- 0.76, N = 12 SE +/- 0.68, N = 12 SE +/- 0.74, N = 12 SE +/- 0.81, N = 12 SE +/- 1.66, N = 15 548.04 539.87 546.71 649.20 660.34 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
WebP Image Encode This is a test of Google's libwebp with the cwebp image encode utility and using a sample 6000x4000 pixel JPEG image as the input. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.2.4 Encode Settings: Default Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 6 12 18 24 30 SE +/- 0.07, N = 13 SE +/- 0.10, N = 13 SE +/- 0.07, N = 13 SE +/- 0.08, N = 13 SE +/- 0.13, N = 13 23.93 26.78 26.81 25.78 26.25 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
HeFFTe - Highly Efficient FFT for Exascale HeFFTe is the Highly Efficient FFT for Exascale software developed as part of the Exascale Computing Project. This test profile uses HeFFTe's built-in speed benchmarks under a variety of configuration options and currently catering to CPU/processor testing. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org GFLOP/s, More Is Better HeFFTe - Highly Efficient FFT for Exascale 2.4 Test: r2c - Backend: FFTW - Precision: double - X Y Z: 128 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 9 18 27 36 45 SE +/- 0.39, N = 15 SE +/- 0.17, N = 15 SE +/- 0.15, N = 15 SE +/- 0.96, N = 15 SE +/- 0.58, N = 15 26.48 22.47 23.72 35.07 37.24 1. (CXX) g++ options: -O3
OpenBenchmarking.org GFLOP/s, More Is Better HeFFTe - Highly Efficient FFT for Exascale 2.4 Test: r2c - Backend: FFTW - Precision: float - X Y Z: 128 Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 30 60 90 120 150 SE +/- 3.73, N = 15 SE +/- 0.33, N = 15 SE +/- 1.52, N = 15 SE +/- 1.50, N = 15 SE +/- 1.14, N = 15 85.01 74.13 77.52 125.54 127.93 1. (CXX) g++ options: -O3
Java SciMark This test runs the Java version of SciMark 2, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This benchmark is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.2 Computational Test: Dense LU Matrix Factorization Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 4K 8K 12K 16K 20K SE +/- 8.09, N = 3 SE +/- 217.22, N = 3 SE +/- 146.88, N = 3 SE +/- 171.55, N = 3 SE +/- 115.98, N = 3 11976.06 18030.92 17934.59 17230.85 17774.36
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.2 Computational Test: Monte Carlo Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 0.88, N = 3 SE +/- 32.12, N = 3 SE +/- 19.82, N = 3 SE +/- 26.32, N = 3 SE +/- 17.64, N = 3 1426.11 2570.60 2541.20 2456.47 2529.18
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.2 Computational Test: Jacobi Successive Over-Relaxation Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 600 1200 1800 2400 3000 SE +/- 2.05, N = 3 SE +/- 22.92, N = 3 SE +/- 22.65, N = 3 SE +/- 18.83, N = 3 SE +/- 10.01, N = 3 2667.90 2487.93 2489.61 2464.14 2518.58
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.2 Computational Test: Fast Fourier Transform Xeon E-2488 @ DDR5-4800 EPYC 4364P @ DDR5-4800 EPYC 4364P @ DDR5-5200 EPYC 4564P @ DDR5-4800 EPYC 4564P @ DDR5-5200 200 400 600 800 1000 SE +/- 8.67, N = 3 SE +/- 0.60, N = 3 SE +/- 0.41, N = 3 SE +/- 3.44, N = 3 SE +/- 0.73, N = 3 1023.23 470.91 450.10 453.05 443.49
Xeon E-2488 @ DDR5-4800 Processor: Intel Xeon E E-2488 @ 3.20GHz (8 Cores / 16 Threads), Motherboard: Supermicro Super Server X13SCL-F v0123456789 (1.1 BIOS), Chipset: Intel Raptor Lake-S PCH, Memory: 2 x 32GB DRAM-4800MT/s Micron MTC20C2085S1EC48BA1, Disk: 3201GB Micron_7450_MTFDKCC3T2TFS + 960GB Micron_7450_MTFDKBA960TFR, Graphics: ASPEED, Network: 2 x Intel I210
OS: Ubuntu 24.04, Kernel: 6.8.0-11-generic (x86_64), Desktop: GNOME Shell 45.3, Display Server: X Server 1.21.1.11, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / errors=remount-ro,relatime,rw,stripe=64 / Block Size: 4096Processor Notes: Scaling Governor: acpi-cpufreq performance (Boost: Enabled) - CPU Microcode: 0x122 - Thermald 2.5.6Java Notes: OpenJDK Runtime Environment (build 21.0.2+13-Ubuntu-2)Python Notes: Python 3.12.2Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional RSB filling PBRSB-eIBRS: SW sequence + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 1 April 2024 17:44 by user phoronix.
EPYC 4364P @ DDR5-4800 Processor: AMD EPYC 4364P 8-Core @ 5.57GHz (8 Cores / 16 Threads), Motherboard: Supermicro AS-3015A-I H13SAE-MF v1.00 (1.1 BIOS), Chipset: AMD Device 14d8, Memory: 2 x 32GB DRAM-4800MT/s Micron MTC20C2085S1EC48BA1, Disk: 3201GB Micron_7450_MTFDKCC3T2TFS + 960GB Micron_7450_MTFDKBA960TFR, Graphics: ASPEED, Audio: AMD Rembrandt Radeon HD Audio, Network: 2 x Intel I210
OS: Ubuntu 24.04, Kernel: 6.8.0-11-generic (x86_64), Desktop: GNOME Shell 45.3, Display Server: X Server 1.21.1.11, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / errors=remount-ro,relatime,rw,stripe=64 / Block Size: 4096Processor Notes: Scaling Governor: amd-pstate-epp performance (EPP: performance) - CPU Microcode: 0xa601203Java Notes: OpenJDK Runtime Environment (build 21.0.2+13-Ubuntu-2)Python Notes: Python 3.12.2Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Vulnerable: Safe RET no microcode + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 26 March 2024 23:58 by user phoronix.
EPYC 4364P @ DDR5-5200 Processor: AMD EPYC 4364P 8-Core @ 5.57GHz (8 Cores / 16 Threads), Motherboard: Supermicro AS-3015A-I H13SAE-MF v1.00 (1.1 BIOS), Chipset: AMD Device 14d8, Memory: 2 x 32GB DRAM-5200MT/s, Disk: 3201GB Micron_7450_MTFDKCC3T2TFS + 960GB Micron_7450_MTFDKBA960TFR, Graphics: ASPEED, Audio: AMD Rembrandt Radeon HD Audio, Network: 2 x Intel I210
OS: Ubuntu 24.04, Kernel: 6.8.0-11-generic (x86_64), Desktop: GNOME Shell 45.3, Display Server: X Server 1.21.1.11, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate-epp performance (EPP: performance) - CPU Microcode: 0xa601203Java Notes: OpenJDK Runtime Environment (build 21.0.2+13-Ubuntu-2)Python Notes: Python 3.12.2Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Vulnerable: Safe RET no microcode + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 28 April 2024 17:44 by user phoronix.
EPYC 4564P @ DDR5-4800 Processor: AMD EPYC 4564P 16-Core @ 5.88GHz (16 Cores / 32 Threads), Motherboard: Supermicro AS-3015A-I H13SAE-MF v1.00 (1.1 BIOS), Chipset: AMD Device 14d8, Memory: 2 x 32GB DRAM-4800MT/s Micron MTC20C2085S1EC48BA1, Disk: 3201GB Micron_7450_MTFDKCC3T2TFS + 960GB Micron_7450_MTFDKBA960TFR, Graphics: ASPEED, Audio: AMD Rembrandt Radeon HD Audio, Network: 2 x Intel I210
OS: Ubuntu 24.04, Kernel: 6.8.0-11-generic (x86_64), Desktop: GNOME Shell 45.3, Display Server: X Server 1.21.1.11, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate-epp performance (EPP: performance) - CPU Microcode: 0xa601203Java Notes: OpenJDK Runtime Environment (build 21.0.2+13-Ubuntu-2)Python Notes: Python 3.12.2Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Vulnerable: Safe RET no microcode + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 21 April 2024 01:56 by user phoronix.
EPYC 4564P @ DDR5-5200 Processor: AMD EPYC 4564P 16-Core @ 5.88GHz (16 Cores / 32 Threads), Motherboard: Supermicro AS-3015A-I H13SAE-MF v1.00 (1.1 BIOS), Chipset: AMD Device 14d8, Memory: 2 x 32GB DRAM-5200MT/s, Disk: 3201GB Micron_7450_MTFDKCC3T2TFS + 960GB Micron_7450_MTFDKBA960TFR, Graphics: ASPEED, Audio: AMD Rembrandt Radeon HD Audio, Network: 2 x Intel I210
OS: Ubuntu 24.04, Kernel: 6.8.0-11-generic (x86_64), Desktop: GNOME Shell 45.3, Display Server: X Server 1.21.1.11, Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-fxIygj/gcc-13-13.2.0/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate-epp performance (EPP: performance) - CPU Microcode: 0xa601203Java Notes: OpenJDK Runtime Environment (build 21.0.2+13-Ubuntu-2)Python Notes: Python 3.12.2Security Notes: gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_rstack_overflow: Vulnerable: Safe RET no microcode + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced / Automatic IBRS IBPB: conditional STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 27 April 2024 00:18 by user phoronix.