Benchmarks by Michael Larabel for a future article. 2 x Intel Xeon Max 9480 testing with a Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS) and ASPEED on Ubuntu 24.04 via the Phoronix Test Suite.
WRF WRF, the Weather Research and Forecasting Model, is a "next-generation mesoscale numerical weather prediction system designed for both atmospheric research and operational forecasting applications. It features two dynamical cores, a data assimilation system, and a software architecture supporting parallel computation and system extensibility." Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better WRF 4.2.2 Input: conus 2.5km Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3K 6K 9K 12K 15K 5601.10 10602.41 6674.12 11980.81 6840.06 12410.29 7157.96 13273.69 1. (F9X) gfortran options: -O2 -ftree-vectorize -funroll-loops -ffree-form -fconvert=big-endian -frecord-marker=4 -fallow-invalid-boz -lesmf_time -lwrfio_nf -lnetcdff -lnetcdf -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 700 1400 2100 2800 3500 SE +/- 13.62, N = 3 SE +/- 2.08, N = 3 SE +/- 6.74, N = 3 SE +/- 1.53, N = 3 SE +/- 4.36, N = 3 SE +/- 13.96, N = 3 SE +/- 8.62, N = 3 SE +/- 6.66, N = 3 3222 2004 2716 1734 1852 1356 1776 1243 MIN: 248 / MAX: 36519 MIN: 134 / MAX: 29750 MIN: 188 / MAX: 30888 MIN: 104 / MAX: 26188 MIN: 191 / MAX: 28947 MIN: 107 / MAX: 20909 MIN: 177 / MAX: 27371 MIN: 95 / MAX: 19286
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 Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3K 6K 9K 12K 15K SE +/- 295.26, N = 8 SE +/- 228.53, N = 12 SE +/- 577.93, N = 3 SE +/- 150.41, N = 12 14463.40 9204.87 14665.00 8868.20 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
NWChem NWChem is an open-source high performance computational chemistry package. Per NWChem's documentation, "NWChem aims to provide its users with computational chemistry tools that are scalable both in their ability to treat large scientific computational chemistry problems efficiently, and in their use of available parallel computing resources from high-performance parallel supercomputers to conventional workstation clusters." Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better NWChem 7.0.2 Input: C240 Buckyball Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 500 1000 1500 2000 2500 1809.7 1878.3 1960.9 2070.3 1873.1 2245.2 1785.0 2530.5 1. (F9X) gfortran options: -lnwctask -lccsd -lmcscf -lselci -lmp2 -lmoints -lstepper -ldriver -loptim -lnwdft -lgradients -lcphf -lesp -lddscf -ldangchang -lguess -lhessian -lvib -lnwcutil -lrimp2 -lproperty -lsolvation -lnwints -lprepar -lnwmd -lnwpw -lofpw -lpaw -lpspw -lband -lnwpwlib -lcafe -lspace -lanalyze -lqhop -lpfft -ldplot -ldrdy -lvscf -lqmmm -lqmd -letrans -ltce -lbq -lmm -lcons -lperfm -ldntmc -lccca -ldimqm -lga -larmci -lpeigs -l64to32 -lopenblas -lpthread -lrt -llapack -lnwcblas -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz -lcomex -m64 -ffast-math -std=legacy -fdefault-integer-8 -finline-functions -O2
RELION RELION - REgularised LIkelihood OptimisatioN - is a stand-alone computer program for Maximum A Posteriori refinement of (multiple) 3D reconstructions or 2D class averages in cryo-electron microscopy (cryo-EM). It is developed in the research group of Sjors Scheres at the MRC Laboratory of Molecular Biology. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better RELION 4.0.1 Test: Basic - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 90 180 270 360 450 SE +/- 1.41, N = 3 SE +/- 2.68, N = 7 SE +/- 2.65, N = 3 SE +/- 7.14, N = 9 SE +/- 2.66, N = 12 SE +/- 4.65, N = 9 SE +/- 3.24, N = 12 SE +/- 2.98, N = 3 155.74 290.45 203.41 373.18 201.38 384.38 214.42 412.37 1. (CXX) g++ options: -fopenmp -std=c++11 -O3 -rdynamic -ldl -ltiff -lfftw3f -lfftw3 -lpng -ljpeg -lmpi_cxx -lmpi
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: 2400 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 9.88, N = 12 SE +/- 0.96, N = 15 SE +/- 0.99, N = 12 SE +/- 0.67, N = 3 SE +/- 18.26, N = 12 SE +/- 5.17, N = 9 SE +/- 15.80, N = 12 SE +/- 12.12, N = 12 102.13 75.73 126.39 119.84 173.55 145.24 182.12 173.01 1. (CXX) g++ options: -O3 -fopenmp
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 - Gridding Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2K 4K 6K 8K 10K SE +/- 60.72, N = 3 SE +/- 54.27, N = 5 SE +/- 75.19, N = 5 SE +/- 4.15, N = 3 SE +/- 195.28, N = 8 SE +/- 32.87, N = 12 SE +/- 23.56, N = 3 SE +/- 87.89, N = 12 6038.92 5062.07 7630.06 8088.11 4382.97 3406.96 3347.93 3375.32 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
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, 95% Reads - Concurrency: 512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30K 60K 90K 120K 150K SE +/- 1875.64, N = 15 SE +/- 2853.06, N = 12 SE +/- 1606.68, N = 15 SE +/- 3058.00, N = 15 SE +/- 1792.91, N = 12 SE +/- 2047.28, N = 12 SE +/- 1830.36, N = 15 SE +/- 2758.29, N = 15 85613.1 135861.9 82341.2 131999.7 76524.7 120494.3 80804.7 128125.9
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1M 2.2M 3.3M 4.4M 5.5M 5186677 916904 4136969 833892 4075342 767812 3658790 713937 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
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: 512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30K 60K 90K 120K 150K SE +/- 1250.17, N = 13 SE +/- 2440.36, N = 12 SE +/- 886.06, N = 15 SE +/- 1885.96, N = 15 SE +/- 970.10, N = 15 SE +/- 1384.66, N = 15 SE +/- 889.12, N = 15 SE +/- 1362.07, N = 3 73640.9 121465.2 72110.1 116603.7 71696.5 109388.2 71680.5 110666.9
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: 1000 - Mode: Read Only - Average Latency Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.2909 0.5818 0.8727 1.1636 1.4545 SE +/- 0.040, N = 12 SE +/- 0.005, N = 12 SE +/- 0.030, N = 12 SE +/- 0.007, N = 3 SE +/- 0.038, N = 9 SE +/- 0.007, N = 3 SE +/- 0.017, N = 9 SE +/- 0.008, N = 12 1.149 0.589 1.293 0.572 1.098 0.622 1.027 0.559 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: 1000 - Mode: Read Only Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400K 800K 1200K 1600K 2000K SE +/- 30933.90, N = 12 SE +/- 13863.97, N = 12 SE +/- 18635.21, N = 12 SE +/- 20350.34, N = 3 SE +/- 36627.33, N = 9 SE +/- 17697.00, N = 3 SE +/- 16726.04, N = 9 SE +/- 24978.23, N = 12 881881 1699046 778141 1748146 920725 1607424 976035 1792636 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -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: 512 - Model: ResNet-50 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.61, N = 3 SE +/- 0.48, N = 3 SE +/- 1.39, N = 3 SE +/- 0.01, N = 3 SE +/- 1.80, N = 3 SE +/- 0.61, N = 3 SE +/- 0.79, N = 3 SE +/- 1.17, N = 3 169.75 140.53 174.95 148.45 138.39 125.14 145.29 117.53
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_bm16 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 90 180 270 360 450 SE +/- 2.02, N = 3 SE +/- 0.54, N = 3 SE +/- 0.74, N = 3 SE +/- 0.13, N = 3 SE +/- 2.11, N = 9 SE +/- 2.03, N = 3 SE +/- 1.56, N = 3 SE +/- 1.43, N = 3 378.05 410.59 214.80 328.47 269.20 357.57 263.33 351.29 1. (F9X) gfortran options: -O3 -march=native -funroll-loops -fopenmp
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.5072 1.0144 1.5216 2.0288 2.536 SE +/- 0.08185, N = 13 SE +/- 0.02648, N = 15 SE +/- 0.04403, N = 15 SE +/- 0.01396, N = 15 SE +/- 0.05463, N = 15 SE +/- 0.05862, N = 15 SE +/- 0.05644, N = 12 SE +/- 0.05087, N = 15 2.25422 1.43787 1.10900 1.05274 1.76778 1.50214 1.88679 1.45223
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: 144 144 144 - RT: 60 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.94, N = 3 SE +/- 0.08, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.37, N = 3 SE +/- 0.29, N = 3 SE +/- 0.08, N = 3 69.30 35.14 60.85 31.07 62.64 30.59 54.41 25.24 1. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -lmpi_cxx -lmpi
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: 800 - Client Number: 400 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 70 140 210 280 350 SE +/- 3.40, N = 12 SE +/- 3.41, N = 3 SE +/- 3.79, N = 8 SE +/- 2.20, N = 3 SE +/- 3.74, N = 12 SE +/- 5.07, N = 3 SE +/- 2.59, N = 12 SE +/- 5.15, N = 3 286.78 280.26 317.36 294.33 325.97 309.03 328.41 314.44 MAX: 28496.67 MAX: 27450.82 MAX: 32341.07 MAX: 27740.38 MAX: 30042.47 MAX: 26720.06 MAX: 29309.84 MAX: 27244.84
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20M 40M 60M 80M 100M SE +/- 761276.42, N = 12 SE +/- 746822.47, N = 3 SE +/- 692701.79, N = 8 SE +/- 308756.34, N = 3 SE +/- 574699.97, N = 12 SE +/- 1037167.36, N = 3 SE +/- 565297.90, N = 12 SE +/- 662950.65, N = 3 89231850 92709751 80169620 87797206 78057728 83475645 77501248 82455664
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 110 220 330 440 550 SE +/- 5.39, N = 5 SE +/- 3.62, N = 6 SE +/- 2.55, N = 3 SE +/- 3.45, N = 3 SE +/- 3.54, N = 12 SE +/- 3.60, N = 3 SE +/- 3.08, N = 3 SE +/- 6.03, N = 3 438.60 458.23 501.63 481.66 471.79 476.91 501.07 470.02 MIN: 58.2 / MAX: 6000 MIN: 42.16 / MAX: 6000 MIN: 64.45 / MAX: 6666.67 MIN: 40.16 / MAX: 5454.55 MIN: 53.43 / MAX: 6000 MIN: 36.01 / MAX: 5000 MIN: 54.25 / MAX: 5000 MIN: 32.59 / MAX: 5454.55
OpenBenchmarking.org Queries Per Minute, Geo Mean, More Is Better ClickHouse 22.12.3.5 100M Rows Hits Dataset, Second Run Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 110 220 330 440 550 SE +/- 3.10, N = 5 SE +/- 2.86, N = 6 SE +/- 3.26, N = 3 SE +/- 2.93, N = 3 SE +/- 3.17, N = 12 SE +/- 4.95, N = 3 SE +/- 0.19, N = 3 SE +/- 1.94, N = 3 443.58 455.79 507.56 486.05 467.96 474.68 499.18 465.06 MIN: 62.43 / MAX: 6666.67 MIN: 42.19 / MAX: 6666.67 MIN: 66.45 / MAX: 5454.55 MIN: 40.43 / MAX: 6666.67 MIN: 52.36 / MAX: 6000 MIN: 36.43 / MAX: 6000 MIN: 54.2 / MAX: 4615.38 MIN: 32.59 / MAX: 5000
OpenBenchmarking.org Queries Per Minute, Geo Mean, More Is Better ClickHouse 22.12.3.5 100M Rows Hits Dataset, First Run / Cold Cache Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 110 220 330 440 550 SE +/- 4.12, N = 5 SE +/- 4.18, N = 6 SE +/- 4.37, N = 3 SE +/- 3.41, N = 3 SE +/- 3.85, N = 12 SE +/- 5.88, N = 3 SE +/- 1.35, N = 3 SE +/- 2.85, N = 3 423.30 437.20 485.63 465.50 460.66 461.62 491.89 446.40 MIN: 61.54 / MAX: 6000 MIN: 41.7 / MAX: 5454.55 MIN: 63.63 / MAX: 5000 MIN: 40.93 / MAX: 5454.55 MIN: 53.62 / MAX: 5454.55 MIN: 35.59 / MAX: 5454.55 MIN: 51.86 / MAX: 5454.55 MIN: 32.64 / MAX: 4615.38
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400K 800K 1200K 1600K 2000K SE +/- 6745.42, N = 3 SE +/- 92.62, N = 3 SE +/- 862.96, N = 3 SE +/- 92.46, N = 3 SE +/- 277.84, N = 3 SE +/- 200.66, N = 3 SE +/- 355.69, N = 3 SE +/- 165.80, N = 3 2057021.6 1036735.6 1916162.9 974189.9 1627941.4 813667.6 1395446.4 697707.2 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org sign/s, More Is Better OpenSSL 3.3 Algorithm: RSA4096 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13K 26K 39K 52K 65K SE +/- 56.68, N = 3 SE +/- 15.15, N = 3 SE +/- 14.34, N = 3 SE +/- 9.84, N = 3 SE +/- 83.09, N = 3 SE +/- 20.13, N = 3 SE +/- 6.92, N = 3 SE +/- 19.45, N = 3 61499.6 31594.4 54601.1 27431.1 48534.8 25441.5 45531.0 22861.0 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -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: 256 - Model: ResNet-50 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.21, N = 3 SE +/- 0.58, N = 3 SE +/- 1.57, N = 3 SE +/- 0.03, N = 3 SE +/- 1.04, N = 12 SE +/- 0.63, N = 3 SE +/- 1.40, N = 3 SE +/- 0.14, N = 3 150.33 129.02 161.06 143.67 135.28 123.69 141.95 119.73
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 90 180 270 360 450 SE +/- 0.56, N = 3 SE +/- 0.72, N = 3 SE +/- 0.86, N = 3 SE +/- 0.80, N = 3 SE +/- 0.73, N = 3 SE +/- 0.67, N = 3 SE +/- 0.56, N = 3 SE +/- 0.61, N = 3 189.74 316.90 218.57 362.45 231.03 384.06 246.94 428.39
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: X3D-benchmarking input.i3d Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 90 180 270 360 450 SE +/- 0.37, N = 3 SE +/- 2.51, N = 3 SE +/- 0.05, N = 3 SE +/- 3.88, N = 4 SE +/- 0.01, N = 3 SE +/- 0.14, N = 3 SE +/- 0.04, N = 3 SE +/- 1.25, N = 3 171.38 323.25 196.81 357.01 187.22 351.64 189.69 407.59 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50M 100M 150M 200M 250M SE +/- 2482737.87, N = 3 SE +/- 2966712.90, N = 9 SE +/- 2148606.38, N = 4 SE +/- 1304590.97, N = 3 SE +/- 2925715.37, N = 11 SE +/- 629927.17, N = 3 SE +/- 4344961.90, N = 10 SE +/- 1751369.24, N = 12 222610378 109766345 187761769 96887479 149409082 73592230 137977203 68019088 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 -mavx512f -mavx512bw -mavx512vnni -mavx512dq -mavx512vl -msse4.1 -mssse3 -msse2 -mbmi2 -flto -flto-partition=one -flto=jobserver
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 70K 140K 210K 280K 350K SE +/- 411.11, N = 3 SE +/- 209.97, N = 3 SE +/- 330.95, N = 3 SE +/- 625.56, N = 3 SE +/- 196.24, N = 3 SE +/- 670.43, N = 3 SE +/- 384.08, N = 3 SE +/- 256.85, N = 3 308434 309896 275472 277205 277298 277260 277589 277647 1. (CC) gcc options: -pedantic -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: 800 - Batch Size Per Write: 100 - Sensor Count: 800 - Client Number: 100 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.80, N = 7 SE +/- 0.68, N = 3 SE +/- 1.18, N = 3 SE +/- 0.15, N = 3 SE +/- 1.11, N = 4 SE +/- 0.87, N = 3 SE +/- 1.01, N = 3 SE +/- 0.48, N = 3 82.73 81.73 92.23 85.16 95.39 89.53 94.84 90.02 MAX: 24148.34 MAX: 24081.78 MAX: 23842.9 MAX: 23810.86 MAX: 23846.38 MAX: 23819.37 MAX: 23842.98 MAX: 23857.34
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20M 40M 60M 80M 100M SE +/- 862582.24, N = 7 SE +/- 739343.51, N = 3 SE +/- 928557.10, N = 3 SE +/- 276414.35, N = 3 SE +/- 872659.60, N = 4 SE +/- 782176.98, N = 3 SE +/- 594234.55, N = 3 SE +/- 435900.66, N = 3 92771639 94616807 83651785 90871770 80139566 86354518 80770465 85744542
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: 512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20K 40K 60K 80K 100K SE +/- 325.76, N = 3 SE +/- 1659.04, N = 15 SE +/- 639.27, N = 15 SE +/- 1037.18, N = 15 SE +/- 556.62, N = 3 SE +/- 1352.20, N = 3 SE +/- 833.19, N = 3 SE +/- 297.86, N = 3 77327.1 115715.8 68524.7 108953.6 66816.9 112265.3 69221.4 103089.8
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-152 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.15, N = 12 SE +/- 0.34, N = 3 SE +/- 0.17, N = 3 SE +/- 0.18, N = 3 SE +/- 0.05, N = 3 SE +/- 0.17, N = 12 SE +/- 0.21, N = 3 SE +/- 0.19, N = 5 16.25 23.49 14.03 20.13 14.96 19.24 15.84 19.01 MIN: 6.19 / MAX: 17.5 MIN: 19.73 / MAX: 24.56 MIN: 8.98 / MAX: 15.41 MIN: 17.06 / MAX: 21.01 MIN: 10.89 / MAX: 15.42 MIN: 10.95 / MAX: 20.39 MIN: 6.82 / MAX: 16.51 MIN: 11.15 / MAX: 20.08
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.11, N = 15 SE +/- 0.10, N = 12 SE +/- 0.18, N = 12 SE +/- 0.08, N = 3 SE +/- 0.20, N = 15 SE +/- 0.07, N = 15 SE +/- 0.22, N = 15 SE +/- 0.02, N = 3 7.70 7.63 7.00 6.85 5.91 6.29 6.23 6.03 MIN: 3.44 / MAX: 9.79 MIN: 3.26 / MAX: 9.01 MIN: 2.74 / MAX: 9.4 MIN: 3.05 / MAX: 8.02 MIN: 2.31 / MAX: 9.24 MIN: 2.44 / MAX: 7.56 MIN: 2.27 / MAX: 9.18 MIN: 2.52 / MAX: 7
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 SE +/- 0.89, N = 3 SE +/- 2.17, N = 3 SE +/- 0.58, N = 3 SE +/- 0.92, N = 3 SE +/- 1.01, N = 3 SE +/- 2.11, N = 6 SE +/- 1.72, N = 3 SE +/- 2.59, N = 4 157.36 187.51 178.28 210.93 181.63 215.50 184.30 220.57
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3M 6M 9M 12M 15M SE +/- 215352.04, N = 15 SE +/- 28201.44, N = 3 SE +/- 343737.14, N = 15 SE +/- 61518.27, N = 10 SE +/- 254193.37, N = 15 SE +/- 26369.56, N = 3 SE +/- 218479.19, N = 15 SE +/- 83772.27, N = 13 12425519 8926893 12321300 7841738 11223329 7366500 9624417 7439663 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13 26 39 52 65 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 SE +/- 0.07, N = 3 SE +/- 0.01, N = 3 56.40 37.46 50.52 34.89 51.82 32.16 46.71 28.28 1. (CXX) g++ options: -O3 -lm -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 70 140 210 280 350 SE +/- 0.21, N = 3 SE +/- 0.69, N = 3 SE +/- 0.42, N = 3 SE +/- 0.18, N = 3 SE +/- 0.06, N = 3 SE +/- 1.09, N = 3 SE +/- 0.23, N = 3 SE +/- 0.12, N = 3 125.87 236.73 141.54 261.97 156.70 290.56 169.44 320.11
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 0.7 Test: wizardcoder-python-34b-v1.0.Q6_K - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.3163 2.6326 3.9489 5.2652 6.5815 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.12, N = 6 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 3.05 4.08 3.82 5.85 1.54 1.54 1.77 1.47
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 60 120 180 240 300 SE +/- 0.12, N = 3 SE +/- 0.52, N = 3 SE +/- 0.48, N = 3 SE +/- 0.71, N = 3 SE +/- 0.52, N = 3 SE +/- 0.74, N = 3 SE +/- 0.84, N = 3 SE +/- 0.53, N = 3 145.41 214.43 172.99 254.00 171.17 252.65 175.63 270.38
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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.44, N = 3 SE +/- 0.71, N = 3 SE +/- 0.69, N = 15 SE +/- 0.09, N = 3 SE +/- 2.59, N = 15 SE +/- 6.41, N = 15 SE +/- 5.82, N = 12 SE +/- 3.36, N = 15 100.74 109.67 92.35 116.32 121.78 155.79 121.67 139.10 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.04271, N = 3 SE +/- 0.05824, N = 3 SE +/- 0.07619, N = 15 SE +/- 0.00664, N = 3 SE +/- 0.16487, N = 15 SE +/- 0.22869, N = 15 SE +/- 0.34632, N = 12 SE +/- 0.15651, N = 15 9.92684 9.11885 10.83623 8.59719 8.26016 6.54965 8.39962 7.24187 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.44, N = 3 SE +/- 0.71, N = 3 SE +/- 0.69, N = 15 SE +/- 0.09, N = 3 SE +/- 2.59, N = 15 SE +/- 6.41, N = 15 SE +/- 5.82, N = 12 SE +/- 3.36, N = 15 100.74 109.67 92.35 116.32 121.78 155.79 121.67 139.10 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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: TPC-H Parquet Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.44, N = 3 SE +/- 0.34, N = 3 SE +/- 0.55, N = 3 SE +/- 0.12, N = 3 SE +/- 0.95, N = 3 SE +/- 0.58, N = 3 SE +/- 0.57, N = 3 SE +/- 0.15, N = 3 144.26 132.87 155.73 145.57 158.31 145.24 156.14 146.87 1. (CXX) g++ options: -O3 -rdynamic -lssl -lcrypto -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.23, N = 3 SE +/- 0.32, N = 4 SE +/- 0.24, N = 3 SE +/- 0.10, N = 3 SE +/- 3.75, N = 15 SE +/- 3.75, N = 15 SE +/- 4.11, N = 15 SE +/- 3.12, N = 12 37.66 28.70 50.83 46.29 68.86 64.88 74.87 63.43 1. (CXX) g++ options: -O3 -fopenmp
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 SE +/- 0.14, N = 3 SE +/- 0.20, N = 3 SE +/- 0.11, N = 3 SE +/- 0.15, N = 3 SE +/- 0.14, N = 3 SE +/- 0.18, N = 3 SE +/- 0.13, N = 3 SE +/- 0.08, N = 3 86.93 157.56 115.89 207.20 117.91 212.99 125.15 228.74
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.24, N = 15 SE +/- 0.02, N = 3 SE +/- 0.23, N = 15 SE +/- 0.06, N = 15 SE +/- 0.20, N = 15 SE +/- 0.10, N = 3 SE +/- 0.12, N = 3 SE +/- 0.08, N = 3 12.74 9.38 11.12 8.72 10.44 8.00 10.17 7.46 MIN: 10.69 / MAX: 19.19 MIN: 8.94 / MAX: 11.49 MIN: 9.18 / MAX: 17.51 MIN: 8.12 / MAX: 10.6 MIN: 8.86 / MAX: 16.08 MIN: 7.65 / MAX: 9.12 MIN: 9.34 / MAX: 14.75 MIN: 7.17 / MAX: 8.21
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 7.0 Encoder: libx265 - Scenario: Upload Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.10, N = 3 SE +/- 0.04, N = 3 SE +/- 0.07, N = 3 SE +/- 0.07, N = 3 SE +/- 0.12, N = 3 SE +/- 0.00, N = 3 26.80 26.93 24.56 24.97 23.30 23.77 22.34 23.33 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.35, N = 3 SE +/- 0.10, N = 3 SE +/- 0.28, N = 3 SE +/- 0.12, N = 3 SE +/- 0.47, N = 14 SE +/- 0.55, N = 15 SE +/- 1.43, N = 15 SE +/- 0.55, N = 15 58.94 43.54 64.29 50.35 69.20 59.18 69.45 54.96 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.10, N = 3 SE +/- 0.05, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.10, N = 14 SE +/- 0.15, N = 15 SE +/- 0.29, N = 15 SE +/- 0.17, N = 15 16.97 22.97 15.56 19.86 14.46 16.92 14.48 18.22 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.35, N = 3 SE +/- 0.10, N = 3 SE +/- 0.28, N = 3 SE +/- 0.12, N = 3 SE +/- 0.47, N = 14 SE +/- 0.55, N = 15 SE +/- 1.43, N = 15 SE +/- 0.55, N = 15 58.94 43.54 64.29 50.35 69.20 59.18 69.45 54.96 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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: AES-128-GCM Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400000M 800000M 1200000M 1600000M 2000000M SE +/- 1622921637.03, N = 3 SE +/- 1253503549.61, N = 3 SE +/- 159598753.20, N = 3 SE +/- 28627701.54, N = 3 SE +/- 697208281.49, N = 3 SE +/- 56355870.99, N = 3 SE +/- 268465891.11, N = 3 SE +/- 43545680.66, N = 3 1769313146013 889567506540 1533026300727 768804693697 1484669627050 747710702660 1277583382757 641200848170 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 300000M 600000M 900000M 1200000M 1500000M SE +/- 1320809584.03, N = 3 SE +/- 3966858436.56, N = 3 SE +/- 1397764228.85, N = 3 SE +/- 2848140790.67, N = 3 SE +/- 112278705.55, N = 3 SE +/- 63026881.96, N = 3 SE +/- 108288170.03, N = 3 SE +/- 119857064.56, N = 3 1497529939377 744387683670 1369534328470 700239352717 1181148957353 592014115453 1011898993323 507456728087 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: ChaCha20-Poly1305 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100000M 200000M 300000M 400000M 500000M SE +/- 82161991.11, N = 3 SE +/- 120899798.30, N = 3 SE +/- 132227134.30, N = 3 SE +/- 766713911.50, N = 3 SE +/- 36941390.18, N = 3 SE +/- 37795783.50, N = 3 SE +/- 4148679.28, N = 3 SE +/- 6370607.14, N = 3 444099144270 223625738013 442202532977 225038850730 372735215297 187796174440 320396720447 161321058213 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: SHA512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10000M 20000M 30000M 40000M 50000M SE +/- 14734422.48, N = 3 SE +/- 37654528.63, N = 3 SE +/- 481040738.21, N = 3 SE +/- 203213936.42, N = 3 SE +/- 14339684.98, N = 3 SE +/- 1250834.62, N = 3 SE +/- 3500318.33, N = 3 SE +/- 500176.69, N = 3 46520760707 23886054770 42846254477 22219927723 37564522053 18893980783 32304222993 16199261277 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: ChaCha20 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 150000M 300000M 450000M 600000M 750000M SE +/- 303075210.03, N = 3 SE +/- 3335467.26, N = 3 SE +/- 271318554.29, N = 3 SE +/- 232288636.76, N = 3 SE +/- 863978337.15, N = 3 SE +/- 312824840.64, N = 3 SE +/- 329510485.23, N = 3 SE +/- 415487944.57, N = 3 684284382760 344150916937 624402972947 316126421173 522912403347 263740690067 449908494927 226597354153 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.3 Algorithm: SHA256 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30000M 60000M 90000M 120000M 150000M SE +/- 29142783.86, N = 3 SE +/- 23323764.84, N = 3 SE +/- 66583540.43, N = 3 SE +/- 159515825.53, N = 3 SE +/- 70647143.49, N = 3 SE +/- 33334799.97, N = 3 SE +/- 126029257.64, N = 3 SE +/- 27466135.54, N = 3 131976050120 66573957920 119106802107 61806994523 103382860113 51840436763 88635753413 44439185187 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -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 Read Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 130M 260M 390M 520M 650M SE +/- 5399265.88, N = 15 SE +/- 242146.75, N = 3 SE +/- 5393159.45, N = 3 SE +/- 168680.90, N = 3 SE +/- 2955608.21, N = 3 SE +/- 2763560.64, N = 13 SE +/- 8491098.40, N = 15 SE +/- 3698537.65, N = 15 586147871 315178381 551286958 276179617 497907490 244070674 419943570 211457048 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
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 7.0 Encoder: libx265 - Scenario: Video On Demand Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.14, N = 3 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.13, N = 3 SE +/- 0.07, N = 3 SE +/- 0.43, N = 3 SE +/- 0.10, N = 3 53.29 53.99 48.77 49.84 46.79 48.14 45.29 47.34 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
OpenBenchmarking.org FPS, More Is Better FFmpeg 7.0 Encoder: libx265 - Scenario: Platform Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.15, N = 3 SE +/- 0.14, N = 3 SE +/- 0.12, N = 3 SE +/- 0.16, N = 3 SE +/- 0.27, N = 3 SE +/- 0.08, N = 3 SE +/- 0.09, N = 3 SE +/- 0.03, N = 3 53.22 53.92 48.76 49.79 46.59 48.35 45.95 47.32 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
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: 512 - Model: ResNet-152 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.18, N = 3 SE +/- 0.16, N = 3 SE +/- 0.09, N = 3 SE +/- 0.04, N = 3 SE +/- 0.16, N = 3 SE +/- 0.18, N = 7 SE +/- 0.09, N = 3 SE +/- 0.18, N = 7 16.17 23.41 14.59 20.46 14.74 19.29 16.68 19.47 MIN: 11.54 / MAX: 16.69 MIN: 22.41 / MAX: 24.3 MIN: 10.11 / MAX: 16.06 MIN: 17.25 / MAX: 21.16 MIN: 10.63 / MAX: 15.36 MIN: 10.66 / MAX: 20.07 MIN: 7.02 / MAX: 17.07 MIN: 11.79 / MAX: 20.53
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 300 600 900 1200 1500 SE +/- 1.88, N = 3 SE +/- 3.01, N = 4 SE +/- 4.96, N = 3 SE +/- 4.45, N = 3 SE +/- 0.65, N = 15 SE +/- 1.08, N = 15 SE +/- 0.66, N = 15 SE +/- 2.12, N = 3 1165.8 1221.8 963.6 1032.0 1039.2 1127.5 1040.2 1126.4 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.03, N = 3 SE +/- 0.21, N = 4 SE +/- 0.12, N = 3 SE +/- 0.06, N = 3 SE +/- 0.13, N = 15 SE +/- 0.11, N = 15 SE +/- 0.09, N = 15 SE +/- 0.15, N = 3 17.8 18.0 15.2 14.6 13.3 14.5 13.8 14.4 1. (CC) gcc options: -O3 -pthread -lz -llzma
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 300 600 900 1200 1500 SE +/- 3.17, N = 3 SE +/- 2.24, N = 3 SE +/- 2.30, N = 3 SE +/- 0.29, N = 3 SE +/- 7.22, N = 3 SE +/- 20.21, N = 15 SE +/- 113.47, N = 12 SE +/- 21.06, N = 12 538.23 488.92 633.29 409.59 719.23 632.15 1498.58 734.80 MIN: 529.58 MIN: 457.65 MIN: 621.42 MIN: 390.22 MIN: 696.95 MIN: 459.4 MIN: 641.2 MIN: 546.95 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4M 8M 12M 16M 20M SE +/- 317498.76, N = 15 SE +/- 1201.85, N = 3 SE +/- 22338.31, N = 3 SE +/- 3511.88, N = 3 SE +/- 695042.48, N = 12 SE +/- 5000.00, N = 3 SE +/- 601801.63, N = 12 SE +/- 54649.77, N = 15 18438600 10922667 18166000 9583000 14609417 8957000 12958000 7626333 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
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: 256 - Model: ResNet-152 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.16, N = 3 SE +/- 0.29, N = 3 SE +/- 0.13, N = 7 SE +/- 0.08, N = 3 SE +/- 0.18, N = 4 SE +/- 0.15, N = 3 SE +/- 0.09, N = 3 SE +/- 0.15, N = 3 16.00 23.33 14.08 20.25 14.92 19.58 16.30 19.20 MIN: 11.43 / MAX: 16.66 MIN: 19.78 / MAX: 24.54 MIN: 6.98 / MAX: 15.46 MIN: 17.03 / MAX: 21.21 MIN: 8.28 / MAX: 15.85 MIN: 13.13 / MAX: 20.27 MIN: 11.86 / MAX: 16.95 MIN: 10.93 / MAX: 20
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 SE +/- 0.06, N = 3 SE +/- 0.01, N = 3 53.93 30.68 38.23 21.89 33.13 18.49 29.35 16.09
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.30, N = 3 SE +/- 0.59, N = 3 SE +/- 0.67, N = 3 SE +/- 0.10, N = 3 SE +/- 0.41, N = 3 SE +/- 0.07, N = 3 SE +/- 0.07, N = 3 SE +/- 0.23, N = 3 127.57 99.18 125.88 99.49 131.80 99.62 127.70 100.09 1. (CXX) g++ options: -O3 -rdynamic -lssl -lcrypto -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.89, N = 3 SE +/- 0.25, N = 3 SE +/- 0.49, N = 3 SE +/- 0.17, N = 3 SE +/- 0.81, N = 3 SE +/- 0.29, N = 3 SE +/- 0.42, N = 3 SE +/- 0.09, N = 3 176.40 193.14 149.09 160.24 135.44 143.48 137.50 144.48
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: gravity_spheres_volume/dim_512/scivis/real_time Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 SE +/- 0.22, N = 15 SE +/- 0.18, N = 15 SE +/- 0.24, N = 12 SE +/- 0.22, N = 15 41.91 23.18 35.60 19.58 15.82 16.27 15.34 15.50
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.30, N = 15 SE +/- 0.05, N = 3 SE +/- 0.12, N = 3 SE +/- 0.07, N = 3 SE +/- 0.28, N = 15 SE +/- 0.04, N = 3 SE +/- 0.16, N = 15 SE +/- 0.10, N = 3 19.91 14.74 18.70 13.46 16.64 12.50 16.91 11.86 MIN: 15 / MAX: 30.59 MIN: 12.17 / MAX: 18.21 MIN: 16.22 / MAX: 25.05 MIN: 11.09 / MAX: 16.56 MIN: 12.39 / MAX: 24.93 MIN: 10.38 / MAX: 14.93 MIN: 12.9 / MAX: 23.07 MIN: 9.93 / MAX: 13.28
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: Ninja Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 SE +/- 0.27, N = 3 SE +/- 0.12, N = 3 SE +/- 0.15, N = 3 SE +/- 0.07, N = 3 SE +/- 0.79, N = 3 SE +/- 0.36, N = 3 SE +/- 0.17, N = 3 SE +/- 0.19, N = 3 106.82 165.05 122.31 187.02 128.07 197.84 134.42 212.48
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.34017, N = 12 SE +/- 0.54361, N = 15 SE +/- 0.17435, N = 15 SE +/- 0.21567, N = 12 SE +/- 0.44989, N = 15 SE +/- 0.14537, N = 15 SE +/- 0.26203, N = 15 SE +/- 0.16058, N = 15 5.85267 4.64814 2.21948 2.97021 2.56319 5.92501 7.30065 5.97130
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.038, N = 3 SE +/- 0.023, N = 3 SE +/- 0.047, N = 3 SE +/- 0.022, N = 3 SE +/- 0.043, N = 15 SE +/- 0.033, N = 3 SE +/- 0.081, N = 3 SE +/- 0.010, N = 3 6.960 7.392 6.480 6.719 6.077 6.501 6.023 6.324 1. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
Numpy Benchmark This is a test to obtain the general Numpy performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Score, More Is Better Numpy Benchmark Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 110 220 330 440 550 SE +/- 1.42, N = 3 SE +/- 1.84, N = 3 SE +/- 0.26, N = 3 SE +/- 0.33, N = 3 SE +/- 1.30, N = 3 SE +/- 0.79, N = 3 SE +/- 1.13, N = 3 SE +/- 0.60, N = 3 519.89 520.22 456.18 455.57 439.84 464.72 435.21 461.61
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: 512 - Model: ResNet-50 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.30, N = 15 SE +/- 0.19, N = 3 SE +/- 0.21, N = 3 SE +/- 0.12, N = 3 SE +/- 0.34, N = 15 SE +/- 0.43, N = 15 SE +/- 0.26, N = 3 SE +/- 0.72, N = 3 40.83 63.30 37.55 56.44 37.84 51.86 41.65 51.05 MIN: 15.4 / MAX: 43.41 MIN: 48.94 / MAX: 65.25 MIN: 24.5 / MAX: 40.69 MIN: 48.09 / MAX: 58.6 MIN: 19.63 / MAX: 40.77 MIN: 30.55 / MAX: 55.87 MIN: 31.6 / MAX: 43.65 MIN: 40.71 / MAX: 53.89
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.17, N = 12 SE +/- 0.11, N = 3 SE +/- 0.12, N = 3 SE +/- 0.04, N = 3 SE +/- 0.16, N = 15 SE +/- 0.05, N = 3 SE +/- 0.22, N = 12 SE +/- 0.03, N = 3 8.88 8.06 8.82 7.10 7.41 6.82 6.81 6.30 MIN: 3.73 / MAX: 11.48 MIN: 3.62 / MAX: 9.37 MIN: 3.72 / MAX: 10.41 MIN: 3.06 / MAX: 8.2 MIN: 2.67 / MAX: 9.42 MIN: 2.81 / MAX: 7.94 MIN: 2.55 / MAX: 9.19 MIN: 2.53 / MAX: 7.3
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100 200 300 400 500 137.04 302.06 196.50 420.65 186.33 410.29 201.11 440.83 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfiniteVolume -lmeshTools -lparallel -llagrangian -lregionModels -lgenericPatchFields -lOpenFOAM -ldl -lm
OpenBenchmarking.org Seconds, Fewer Is Better OpenFOAM 10 Input: drivaerFastback, Medium Mesh Size - Mesh Time Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 139.62 148.40 162.03 160.73 159.31 150.67 134.78 153.12 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfiniteVolume -lmeshTools -lparallel -llagrangian -lregionModels -lgenericPatchFields -lOpenFOAM -ldl -lm
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: 1000 - Mode: Read Write - Average Latency Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.10, N = 3 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.12, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 SE +/- 0.02, N = 3 30.20 13.26 14.86 14.04 16.66 13.85 16.96 14.29 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: 1000 - Mode: Read Write Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 16K 32K 48K 64K 80K SE +/- 106.16, N = 3 SE +/- 198.41, N = 3 SE +/- 137.82, N = 3 SE +/- 78.02, N = 3 SE +/- 440.63, N = 3 SE +/- 143.25, N = 3 SE +/- 195.77, N = 3 SE +/- 95.84, N = 3 33110 75396 67315 71227 60046 72225 58952 69969 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.72, N = 8 SE +/- 0.34, N = 3 SE +/- 0.06, N = 3 SE +/- 0.12, N = 3 SE +/- 0.51, N = 15 SE +/- 0.32, N = 3 SE +/- 0.02, N = 3 SE +/- 0.60, N = 3 86.85 54.14 88.39 59.91 81.65 59.32 70.73 59.58
OpenBenchmarking.org Seconds, Fewer Is Better OpenRadioss 2023.09.15 Model: INIVOL and Fluid Structure Interaction Drop Container Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.30, N = 3 SE +/- 0.12, N = 3 SE +/- 0.17, N = 3 SE +/- 0.19, N = 3 SE +/- 0.09, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 SE +/- 0.08, N = 3 100.00 128.74 107.36 140.43 110.56 146.03 114.29 160.56
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 300 600 900 1200 1500 SE +/- 1.60, N = 3 SE +/- 0.84, N = 3 SE +/- 3.69, N = 3 SE +/- 0.43, N = 3 SE +/- 1.78, N = 15 SE +/- 1.45, N = 3 SE +/- 1.00, N = 3 SE +/- 0.56, N = 3 1189.7 1242.6 992.1 1069.1 1046.0 1133.3 1041.8 1131.1 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 15 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 9.67 9.80 8.16 7.94 7.46 7.72 7.62 7.76 1. (CC) gcc options: -O3 -pthread -lz -llzma
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.29, N = 3 SE +/- 0.91, N = 3 SE +/- 0.56, N = 3 SE +/- 0.76, N = 3 SE +/- 1.69, N = 3 SE +/- 0.32, N = 3 SE +/- 0.56, N = 3 SE +/- 0.18, N = 3 103.38 122.72 118.57 140.23 122.70 143.81 123.88 149.38
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3K 6K 9K 12K 15K SE +/- 64.69, N = 3 SE +/- 44.68, N = 3 SE +/- 4.03, N = 3 SE +/- 23.16, N = 3 SE +/- 7.44, N = 3 SE +/- 14.78, N = 3 SE +/- 14.02, N = 3 SE +/- 12.93, N = 3 16324.7 8378.3 11263.2 5808.8 12115.0 6122.3 10975.8 5446.8 1. (CXX) g++ options: -fexceptions -fno-rtti -maes -O3 -Ofast -static-libgcc -static-libstdc++ -rdynamic -lssl -lcrypto -luv -lpthread -lrt -ldl -lhwloc
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 120 240 360 480 600 SE +/- 1.52, N = 3 SE +/- 3.68, N = 3 SE +/- 2.62, N = 3 SE +/- 0.34, N = 3 SE +/- 2.46, N = 3 SE +/- 0.37, N = 3 SE +/- 1.14, N = 3 SE +/- 0.41, N = 3 577.76 462.78 576.08 461.23 477.33 398.40 467.91 383.57
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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.3867 0.7734 1.1601 1.5468 1.9335 SE +/- 0.01541, N = 3 SE +/- 0.00291, N = 3 SE +/- 0.01405, N = 15 SE +/- 0.01594, N = 3 SE +/- 0.01853, N = 5 SE +/- 0.01456, N = 3 SE +/- 0.00498, N = 3 SE +/- 0.02719, N = 15 1.32454 1.10042 1.43573 1.17292 1.71875 1.41662 1.65270 1.37831 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200 400 600 800 1000 SE +/- 8.88, N = 3 SE +/- 2.43, N = 3 SE +/- 6.50, N = 15 SE +/- 11.46, N = 3 SE +/- 6.21, N = 5 SE +/- 7.34, N = 3 SE +/- 1.83, N = 3 SE +/- 12.91, N = 15 754.81 908.19 696.98 852.33 581.77 705.61 604.76 728.47 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.3867 0.7734 1.1601 1.5468 1.9335 SE +/- 0.01541, N = 3 SE +/- 0.00291, N = 3 SE +/- 0.01405, N = 15 SE +/- 0.01594, N = 3 SE +/- 0.01853, N = 5 SE +/- 0.01456, N = 3 SE +/- 0.00498, N = 3 SE +/- 0.02719, N = 15 1.32454 1.10042 1.43573 1.17292 1.71875 1.41662 1.65270 1.37831 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.15083, N = 15 SE +/- 0.01631, N = 3 SE +/- 0.03218, N = 3 SE +/- 0.03075, N = 3 SE +/- 0.06105, N = 3 SE +/- 0.06620, N = 3 SE +/- 0.02902, N = 3 SE +/- 0.06596, N = 15 4.29531 3.69649 5.55101 4.49976 7.53259 5.91961 7.04111 5.67700 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 60 120 180 240 300 SE +/- 8.14, N = 15 SE +/- 1.19, N = 3 SE +/- 1.05, N = 3 SE +/- 1.50, N = 3 SE +/- 1.08, N = 3 SE +/- 1.88, N = 3 SE +/- 0.58, N = 3 SE +/- 2.01, N = 15 236.65 270.33 180.03 222.03 132.71 168.86 141.95 176.33 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: GPT-2 - Device: CPU - Executor: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.15083, N = 15 SE +/- 0.01631, N = 3 SE +/- 0.03218, N = 3 SE +/- 0.03075, N = 3 SE +/- 0.06105, N = 3 SE +/- 0.06620, N = 3 SE +/- 0.02902, N = 3 SE +/- 0.06596, N = 15 4.29531 3.69649 5.55101 4.49976 7.53259 5.91961 7.04111 5.67700 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: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.9132 1.8264 2.7396 3.6528 4.566 SE +/- 0.01381, N = 3 SE +/- 0.01801, N = 3 SE +/- 0.04801, N = 3 SE +/- 0.00807, N = 3 SE +/- 0.02633, N = 15 SE +/- 0.01443, N = 3 SE +/- 0.04072, N = 3 SE +/- 0.03338, N = 15 2.81739 2.15980 4.05881 3.06500 3.95819 2.88734 4.05270 2.97094 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100 200 300 400 500 SE +/- 1.73, N = 3 SE +/- 3.84, N = 3 SE +/- 2.94, N = 3 SE +/- 0.86, N = 3 SE +/- 1.67, N = 15 SE +/- 1.74, N = 3 SE +/- 2.44, N = 3 SE +/- 3.48, N = 15 354.75 462.86 246.34 326.14 252.67 346.21 246.68 337.00 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.9132 1.8264 2.7396 3.6528 4.566 SE +/- 0.01381, N = 3 SE +/- 0.01801, N = 3 SE +/- 0.04801, N = 3 SE +/- 0.00807, N = 3 SE +/- 0.02633, N = 15 SE +/- 0.01443, N = 3 SE +/- 0.04072, N = 3 SE +/- 0.03338, N = 15 2.81739 2.15980 4.05881 3.06500 3.95819 2.88734 4.05270 2.97094 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.83, N = 3 SE +/- 0.14, N = 3 SE +/- 0.29, N = 3 SE +/- 0.24, N = 3 SE +/- 0.38, N = 3 SE +/- 0.03, N = 3 SE +/- 0.69, N = 3 SE +/- 0.45, N = 3 113.15 110.32 124.44 117.41 119.03 120.76 116.49 126.89
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: Li2_STO_ae Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.93, N = 3 SE +/- 1.03, N = 3 SE +/- 1.11, N = 3 SE +/- 0.89, N = 3 SE +/- 1.11, N = 5 SE +/- 1.04, N = 3 SE +/- 0.88, N = 3 SE +/- 1.20, N = 3 97.91 105.98 101.63 109.30 106.27 117.81 102.12 121.19 1. (CXX) g++ options: -fopenmp -foffload=disable -finline-limit=1000 -fstrict-aliasing -funroll-all-loops -ffast-math -march=native -O3 -lm -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: Material Tester Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 210.82 118.78 225.35 126.15 239.73 134.39 243.24 136.49
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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.39, N = 13 SE +/- 0.19, N = 3 SE +/- 0.36, N = 3 SE +/- 0.06, N = 3 SE +/- 0.18, N = 3 SE +/- 0.33, N = 3 SE +/- 0.21, N = 3 SE +/- 0.32, N = 15 26.09 15.04 27.59 18.93 29.44 23.66 24.65 21.22 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.51, N = 13 SE +/- 0.82, N = 3 SE +/- 0.48, N = 3 SE +/- 0.16, N = 3 SE +/- 0.21, N = 3 SE +/- 0.59, N = 3 SE +/- 0.34, N = 3 SE +/- 0.69, N = 15 38.42 66.52 36.25 52.82 33.96 42.27 40.57 47.27 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.39, N = 13 SE +/- 0.19, N = 3 SE +/- 0.36, N = 3 SE +/- 0.06, N = 3 SE +/- 0.18, N = 3 SE +/- 0.33, N = 3 SE +/- 0.21, N = 3 SE +/- 0.32, N = 15 26.09 15.04 27.59 18.93 29.44 23.66 24.65 21.22 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 SE +/- 0.36, N = 3 SE +/- 0.05, N = 3 SE +/- 0.11, N = 3 SE +/- 0.10, N = 3 SE +/- 0.32, N = 15 SE +/- 0.18, N = 15 SE +/- 0.22, N = 3 SE +/- 0.17, N = 3 42.13 23.73 36.15 19.85 15.37 16.32 15.98 15.40
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 3.1 Benchmark: particle_volume/ao/real_time Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.07, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 54.12 30.84 38.22 21.93 33.11 18.55 29.50 16.05
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 600 1200 1800 2400 3000 SE +/- 4.22, N = 3 SE +/- 5.57, N = 3 SE +/- 2.79, N = 3 SE +/- 0.67, N = 3 SE +/- 31.34, N = 3 SE +/- 8.14, N = 3 SE +/- 119.20, N = 12 SE +/- 12.69, N = 3 626.14 703.47 635.15 623.60 2776.52 1180.85 1966.88 1200.75 MIN: 600.37 MIN: 649.83 MIN: 612.85 MIN: 598.72 MIN: 2715.64 MIN: 1153.29 MIN: 1146.52 MIN: 1127.94 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -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: Read While Writing Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4M 8M 12M 16M 20M SE +/- 195912.00, N = 15 SE +/- 143570.67, N = 4 SE +/- 239363.92, N = 3 SE +/- 62510.72, N = 3 SE +/- 197596.27, N = 3 SE +/- 93731.24, N = 7 SE +/- 171807.74, N = 3 SE +/- 88257.34, N = 3 17585737 12168413 17297027 10695360 15354400 9988176 13859254 8795408 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
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: 1:100 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1000K 2000K 3000K 4000K 5000K SE +/- 35300.21, N = 3 SE +/- 35182.67, N = 3 SE +/- 50660.01, N = 15 SE +/- 25335.31, N = 3 SE +/- 24083.78, N = 3 SE +/- 52575.97, N = 3 SE +/- 26891.33, N = 3 SE +/- 21428.06, N = 3 3780240.56 4752673.79 3245908.40 4481667.95 3392842.38 4315467.17 3373323.92 4384040.52 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1600 3200 4800 6400 8000 SE +/- 52.55, N = 15 SE +/- 57.43, N = 15 SE +/- 31.29, N = 3 SE +/- 29.41, N = 3 SE +/- 48.60, N = 15 SE +/- 40.72, N = 15 SE +/- 48.92, N = 14 SE +/- 36.36, N = 15 7492 7321 4215 4115 4069 3884 4199 3887
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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.01358, N = 3 SE +/- 0.02572, N = 3 SE +/- 0.03225, N = 3 SE +/- 0.01810, N = 3 SE +/- 0.06816, N = 3 SE +/- 0.03815, N = 5 SE +/- 0.09530, N = 15 SE +/- 0.02187, N = 3 5.83954 3.18392 5.62517 3.17002 6.17796 3.37034 5.74491 3.02897 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 70 140 210 280 350 SE +/- 0.40, N = 3 SE +/- 2.52, N = 3 SE +/- 1.01, N = 3 SE +/- 1.80, N = 3 SE +/- 1.78, N = 3 SE +/- 3.31, N = 5 SE +/- 2.72, N = 15 SE +/- 2.40, N = 3 171.23 314.06 177.76 315.41 161.89 296.80 174.68 330.10 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.01358, N = 3 SE +/- 0.02572, N = 3 SE +/- 0.03225, N = 3 SE +/- 0.01810, N = 3 SE +/- 0.06816, N = 3 SE +/- 0.03815, N = 5 SE +/- 0.09530, N = 15 SE +/- 0.02187, N = 3 5.83954 3.18392 5.62517 3.17002 6.17796 3.37034 5.74491 3.02897 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: super-resolution-10 - Device: CPU - Executor: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1255 2.251 3.3765 4.502 5.6275 SE +/- 0.05342, N = 3 SE +/- 0.00044, N = 3 SE +/- 0.10855, N = 15 SE +/- 0.00313, N = 3 SE +/- 0.02558, N = 3 SE +/- 0.00532, N = 3 SE +/- 0.05347, N = 4 SE +/- 0.00187, N = 3 3.98013 3.85774 4.72530 4.39649 5.00222 4.67269 4.85818 4.58912 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 60 120 180 240 300 SE +/- 3.39, N = 3 SE +/- 0.03, N = 3 SE +/- 4.15, N = 15 SE +/- 0.16, N = 3 SE +/- 1.02, N = 3 SE +/- 0.25, N = 3 SE +/- 2.30, N = 4 SE +/- 0.09, N = 3 251.31 259.19 212.92 227.42 199.90 213.99 205.89 217.88 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1255 2.251 3.3765 4.502 5.6275 SE +/- 0.05342, N = 3 SE +/- 0.00044, N = 3 SE +/- 0.10855, N = 15 SE +/- 0.00313, N = 3 SE +/- 0.02558, N = 3 SE +/- 0.00532, N = 3 SE +/- 0.05347, N = 4 SE +/- 0.00187, N = 3 3.98013 3.85774 4.72530 4.39649 5.00222 4.67269 4.85818 4.58912 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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-152 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.19, N = 5 SE +/- 0.08, N = 3 SE +/- 0.16, N = 4 SE +/- 0.09, N = 3 SE +/- 0.13, N = 8 SE +/- 0.16, N = 3 SE +/- 0.18, N = 3 SE +/- 0.26, N = 3 17.57 27.21 15.77 19.41 14.86 18.07 17.91 19.99 MIN: 5.14 / MAX: 18.58 MIN: 13.91 / MAX: 28.31 MIN: 8.33 / MAX: 16.91 MIN: 14.68 / MAX: 22.3 MIN: 0.53 / MAX: 17.18 MIN: 0.27 / MAX: 21.93 MIN: 5.47 / MAX: 18.9 MIN: 0.34 / MAX: 23.05
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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.44, N = 3 SE +/- 0.04, N = 3 SE +/- 0.62, N = 3 SE +/- 0.17, N = 3 SE +/- 1.12, N = 15 SE +/- 0.23, N = 3 SE +/- 0.68, N = 3 SE +/- 0.44, N = 3 44.60 39.75 60.36 49.64 55.90 49.97 60.11 47.24 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.22, N = 3 SE +/- 0.02, N = 3 SE +/- 0.17, N = 3 SE +/- 0.07, N = 3 SE +/- 0.35, N = 15 SE +/- 0.09, N = 3 SE +/- 0.19, N = 3 SE +/- 0.20, N = 3 22.43 25.16 16.57 20.15 17.99 20.01 16.64 21.17 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: bertsquad-12 - Device: CPU - Executor: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.44, N = 3 SE +/- 0.04, N = 3 SE +/- 0.62, N = 3 SE +/- 0.17, N = 3 SE +/- 1.12, N = 15 SE +/- 0.23, N = 3 SE +/- 0.68, N = 3 SE +/- 0.44, N = 3 44.60 39.75 60.36 49.64 55.90 49.97 60.11 47.24 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: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.1103 0.2206 0.3309 0.4412 0.5515 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 15 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.46 0.49 0.29 0.43 0.30 0.37 0.28 0.33 MIN: 0.18 / MAX: 22.29 MIN: 0.18 / MAX: 16.07 MIN: 0.24 / MAX: 39.69 MIN: 0.23 / MAX: 17.03 MIN: 0.26 / MAX: 42.1 MIN: 0.28 / MAX: 40.85 MIN: 0.25 / MAX: 41.4 MIN: 0.26 / MAX: 19.2 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40K 80K 120K 160K 200K SE +/- 1158.60, N = 3 SE +/- 988.12, N = 3 SE +/- 2494.14, N = 3 SE +/- 75.30, N = 3 SE +/- 1422.88, N = 15 SE +/- 1761.33, N = 3 SE +/- 18.38, N = 3 SE +/- 383.22, N = 3 128566.81 81279.79 189893.31 115106.30 176900.32 140144.01 174862.56 135075.07 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: Standard Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.213 2.426 3.639 4.852 6.065 SE +/- 0.02371, N = 3 SE +/- 0.00472, N = 3 SE +/- 0.00316, N = 3 SE +/- 0.00659, N = 3 SE +/- 0.02608, N = 3 SE +/- 0.01103, N = 3 SE +/- 0.03283, N = 3 SE +/- 0.08827, N = 15 2.80263 2.16966 3.46571 2.79694 5.39123 4.64757 5.02267 4.57581 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100 200 300 400 500 SE +/- 3.01, N = 3 SE +/- 1.00, N = 3 SE +/- 0.26, N = 3 SE +/- 0.84, N = 3 SE +/- 0.89, N = 3 SE +/- 0.51, N = 3 SE +/- 1.31, N = 3 SE +/- 3.95, N = 15 356.76 460.78 288.47 357.36 185.46 215.13 199.08 219.53 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.213 2.426 3.639 4.852 6.065 SE +/- 0.02371, N = 3 SE +/- 0.00472, N = 3 SE +/- 0.00316, N = 3 SE +/- 0.00659, N = 3 SE +/- 0.02608, N = 3 SE +/- 0.01103, N = 3 SE +/- 0.03283, N = 3 SE +/- 0.08827, N = 15 2.80263 2.16966 3.46571 2.79694 5.39123 4.64757 5.02267 4.57581 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
Timed CPython Compilation This test times how long it takes to build the reference Python implementation, CPython, with optimizations and LTO enabled for a release build. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed CPython Compilation 3.10.6 Build Configuration: Released Build, PGO + LTO Optimized Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 60 120 180 240 300 244.95 244.82 279.69 277.39 276.23 273.52 276.46 275.39
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 129.96 119.80 130.62 120.75 137.00 128.60 137.46 132.13
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100 200 300 400 500 SE +/- 4.07, N = 3 SE +/- 0.74, N = 3 SE +/- 0.96, N = 3 SE +/- 0.94, N = 3 SE +/- 6.06, N = 3 SE +/- 2.21, N = 3 SE +/- 5.57, N = 13 SE +/- 1.73, N = 3 327.88 302.27 352.74 329.04 475.25 382.11 460.67 380.90 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.7444 1.4888 2.2332 2.9776 3.722 SE +/- 0.03826, N = 3 SE +/- 0.00805, N = 3 SE +/- 0.00768, N = 3 SE +/- 0.00871, N = 3 SE +/- 0.02652, N = 3 SE +/- 0.01517, N = 3 SE +/- 0.02702, N = 13 SE +/- 0.01189, N = 3 3.05085 3.30836 2.83498 3.03922 2.10480 2.61721 2.17463 2.62544 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100 200 300 400 500 SE +/- 4.07, N = 3 SE +/- 0.74, N = 3 SE +/- 0.96, N = 3 SE +/- 0.94, N = 3 SE +/- 6.06, N = 3 SE +/- 2.21, N = 3 SE +/- 5.57, N = 13 SE +/- 1.73, N = 3 327.88 302.27 352.74 329.04 475.25 382.11 460.67 380.90 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.22, N = 3 SE +/- 0.09, N = 3 SE +/- 0.27, N = 3 SE +/- 0.04, N = 3 SE +/- 0.22, N = 3 SE +/- 0.09, N = 3 SE +/- 0.23, N = 3 SE +/- 0.03, N = 3 86.11 78.71 81.30 79.54 80.32 81.72 78.27 83.44
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30K 60K 90K 120K 150K SE +/- 773.70, N = 3 SE +/- 840.59, N = 3 SE +/- 1197.41, N = 15 SE +/- 1953.08, N = 3 SE +/- 1131.21, N = 15 SE +/- 122.31, N = 3 SE +/- 1027.07, N = 15 SE +/- 36.42, N = 3 64181.29 153584.20 53914.93 147469.33 65663.82 134048.96 83665.76 117588.99 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.07, N = 3 SE +/- 0.08, N = 3 SE +/- 0.06, N = 3 SE +/- 0.26, N = 3 SE +/- 0.42, N = 3 SE +/- 0.12, N = 3 SE +/- 0.11, N = 3 SE +/- 0.53, N = 3 48.09 89.07 50.83 93.35 58.25 107.24 65.03 123.83 1. (CC) gcc options: -O2 -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: Tradebeans Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3K 6K 9K 12K 15K SE +/- 16.01, N = 3 SE +/- 226.65, N = 15 SE +/- 64.94, N = 3 SE +/- 79.99, N = 3 SE +/- 10.41, N = 3 SE +/- 54.76, N = 15 SE +/- 66.27, N = 3 SE +/- 51.68, N = 15 12713 11004 8050 7270 6672 5994 6681 6130
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.03, N = 15 SE +/- 0.05, N = 9 SE +/- 0.06, N = 15 SE +/- 0.05, N = 4 SE +/- 0.04, N = 15 SE +/- 0.10, N = 4 SE +/- 0.07, N = 3 SE +/- 0.09, N = 4 3.85 6.43 4.39 10.53 4.37 9.99 5.11 11.29
OpenBenchmarking.org images/sec, More Is Better TensorFlow 2.16.1 Device: CPU - Batch Size: 64 - Model: ResNet-50 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.70, N = 3 SE +/- 0.50, N = 3 SE +/- 1.12, N = 3 SE +/- 0.04, N = 3 SE +/- 1.28, N = 3 SE +/- 0.93, N = 3 SE +/- 0.39, N = 3 SE +/- 0.13, N = 3 94.66 91.51 115.47 120.93 107.30 103.56 111.71 104.23
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.06, N = 3 SE +/- 0.09, N = 3 SE +/- 0.26, N = 3 SE +/- 0.15, N = 3 SE +/- 0.20, N = 3 SE +/- 0.44, N = 3 SE +/- 0.08, N = 3 SE +/- 0.31, N = 3 41.85 80.29 46.38 87.92 51.29 97.80 55.26 106.96
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.03, N = 3 SE +/- 0.08, N = 3 SE +/- 0.05, N = 3 SE +/- 0.32, N = 3 SE +/- 0.32, N = 3 SE +/- 0.22, N = 3 SE +/- 0.29, N = 3 SE +/- 0.15, N = 3 60.53 64.48 67.38 69.93 68.44 73.06 69.65 76.92 1. (CXX) g++ options: -O3 -fPIC -lm
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: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 500 1000 1500 2000 2500 SE +/- 0.00, N = 3 SE +/- 1.45, N = 3 SE +/- 0.88, N = 3 SE +/- 0.88, N = 3 SE +/- 0.33, N = 3 SE +/- 1.20, N = 3 SE +/- 2.65, N = 3 SE +/- 1.20, N = 3 898 1709 1063 1977 1143 2123 1237 2357
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 110 220 330 440 550 SE +/- 2.68, N = 3 SE +/- 2.30, N = 3 SE +/- 4.87, N = 3 SE +/- 0.24, N = 3 SE +/- 4.80, N = 3 SE +/- 1.91, N = 3 SE +/- 2.29, N = 3 SE +/- 1.67, N = 3 501.39 423.59 528.93 448.46 449.91 386.00 447.50 375.73
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: Rotate Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.33, N = 3 SE +/- 0.67, N = 3 SE +/- 0.33, N = 3 SE +/- 1.93, N = 6 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.58, N = 3 SE +/- 0.33, N = 3 182 203 172 196 143 194 148 197 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
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: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400 800 1200 1600 2000 SE +/- 0.88, N = 3 SE +/- 0.88, N = 3 SE +/- 0.58, N = 3 SE +/- 0.58, N = 3 SE +/- 0.58, N = 3 SE +/- 0.58, N = 3 SE +/- 1.76, N = 3 SE +/- 0.33, N = 3 759 1441 894 1657 961 1792 1049 1988
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 80 160 240 320 400 SE +/- 0.33, N = 3 SE +/- 0.69, N = 3 SE +/- 0.00, N = 3 SE +/- 0.28, N = 3 SE +/- 0.35, N = 3 SE +/- 0.47, N = 3 SE +/- 0.40, N = 3 SE +/- 0.15, N = 3 234.24 224.64 283.11 356.38 278.52 257.50 256.63 236.52 MIN: 148.68 / MAX: 448.54 MIN: 153.29 / MAX: 553.64 MIN: 203.37 / MAX: 332.08 MIN: 109.11 / MAX: 395.69 MIN: 196.58 / MAX: 401.59 MIN: 187.62 / MAX: 295.06 MIN: 191.95 / MAX: 301.21 MIN: 172.47 / MAX: 265.77 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 120 240 360 480 600 SE +/- 0.77, N = 3 SE +/- 0.69, N = 3 SE +/- 0.01, N = 3 SE +/- 0.12, N = 3 SE +/- 0.54, N = 3 SE +/- 0.41, N = 3 SE +/- 0.58, N = 3 SE +/- 0.14, N = 3 545.21 284.64 423.18 167.83 401.45 217.11 373.50 202.59 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: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20K 40K 60K 80K 100K SE +/- 47.16, N = 3 SE +/- 162.52, N = 3 SE +/- 125.33, N = 3 SE +/- 58.68, N = 3 SE +/- 141.04, N = 3 SE +/- 65.97, N = 3 SE +/- 155.80, N = 3 SE +/- 41.33, N = 3 33391 59522 38813 68006 41971 73366 45211 80742
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 1.93, N = 15 SE +/- 0.03, N = 3 SE +/- 0.99, N = 3 SE +/- 0.31, N = 3 SE +/- 0.93, N = 3 SE +/- 0.96, N = 4 SE +/- 1.48, N = 3 SE +/- 0.31, N = 3 191.82 108.12 140.83 86.97 133.99 78.11 127.85 70.21 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1000 2000 3000 4000 5000 SE +/- 48.20, N = 15 SE +/- 0.70, N = 3 SE +/- 24.63, N = 3 SE +/- 7.77, N = 3 SE +/- 23.14, N = 3 SE +/- 23.93, N = 4 SE +/- 36.96, N = 3 SE +/- 7.81, N = 3 4795.46 2703.09 3520.71 2174.19 3349.76 1952.61 3196.12 1755.33 1. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 11 22 33 44 55 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.16, N = 3 SE +/- 0.03, N = 3 SE +/- 0.09, N = 3 SE +/- 0.01, N = 3 48.76 27.10 42.82 24.64 39.34 22.36 35.93 19.91
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: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15K 30K 45K 60K 75K SE +/- 16.05, N = 3 SE +/- 90.37, N = 3 SE +/- 64.24, N = 3 SE +/- 102.40, N = 3 SE +/- 13.86, N = 3 SE +/- 64.91, N = 3 SE +/- 37.55, N = 3 SE +/- 134.33, N = 3 24249 50798 33411 58026 36179 62540 38926 68973
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 1.0 Camera: 1 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8K 16K 24K 32K 40K SE +/- 7.13, N = 3 SE +/- 37.49, N = 3 SE +/- 34.10, N = 3 SE +/- 62.46, N = 3 SE +/- 51.39, N = 3 SE +/- 83.47, N = 3 SE +/- 72.42, N = 3 SE +/- 28.06, N = 3 12124 23125 14138 31380 15357 33790 16645 37055
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: Weld Porosity Detection FP16-INT8 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.8055 1.611 2.4165 3.222 4.0275 SE +/- 0.00, N = 3 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 SE +/- 0.01, N = 3 SE +/- 0.03, N = 4 2.48 2.36 2.50 3.58 2.79 2.67 2.53 2.50 MIN: 1.98 / MAX: 34.01 MIN: 2.02 / MAX: 21.33 MIN: 2.28 / MAX: 40.8 MIN: 1.4 / MAX: 12.71 MIN: 2.44 / MAX: 34.21 MIN: 2.39 / MAX: 34.22 MIN: 2.23 / MAX: 31.44 MIN: 2.22 / MAX: 27.89 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10K 20K 30K 40K 50K SE +/- 39.08, N = 3 SE +/- 47.41, N = 3 SE +/- 53.98, N = 3 SE +/- 2.71, N = 3 SE +/- 233.95, N = 3 SE +/- 21.27, N = 3 SE +/- 80.58, N = 3 SE +/- 234.43, N = 4 48122.16 26943.42 47654.95 15943.17 40032.81 20910.53 37710.60 19183.83 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.00, N = 3 SE +/- 0.09, N = 3 SE +/- 0.02, N = 3 SE +/- 0.12, N = 3 SE +/- 0.04, N = 3 13.15 9.40 15.96 13.95 18.72 12.72 17.15 11.37 MIN: 7.28 / MAX: 47.97 MIN: 7.66 / MAX: 25.07 MIN: 10.38 / MAX: 43.64 MIN: 8.67 / MAX: 65.92 MIN: 12.15 / MAX: 64.81 MIN: 9.56 / MAX: 30.66 MIN: 10.81 / MAX: 59.34 MIN: 9.17 / MAX: 32.12 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2K 4K 6K 8K 10K SE +/- 24.76, N = 3 SE +/- 10.14, N = 3 SE +/- 13.31, N = 3 SE +/- 1.87, N = 3 SE +/- 28.98, N = 3 SE +/- 7.60, N = 3 SE +/- 37.79, N = 3 SE +/- 15.13, N = 3 9360.89 6757.26 7356.63 4209.10 5933.65 4386.48 5565.23 4208.49 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: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9K 18K 27K 36K 45K SE +/- 8.00, N = 3 SE +/- 26.00, N = 3 SE +/- 29.96, N = 3 SE +/- 61.20, N = 3 SE +/- 10.97, N = 3 SE +/- 74.33, N = 3 SE +/- 25.12, N = 3 SE +/- 36.47, N = 3 14358 31993 16837 36649 18111 39238 19701 42837
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.4698 0.9396 1.4094 1.8792 2.349 SE +/- 0.00701, N = 3 SE +/- 0.01491, N = 4 SE +/- 0.01492, N = 3 SE +/- 0.00936, N = 3 SE +/- 0.00871, N = 3 SE +/- 0.01053, N = 3 SE +/- 0.00720, N = 3 SE +/- 0.00884, N = 3 1.64919 1.20298 1.76353 1.27368 2.08780 1.47249 2.02659 1.45280 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200 400 600 800 1000 SE +/- 2.57, N = 3 SE +/- 10.06, N = 4 SE +/- 4.81, N = 3 SE +/- 5.72, N = 3 SE +/- 1.98, N = 3 SE +/- 4.87, N = 3 SE +/- 1.77, N = 3 SE +/- 4.16, N = 3 605.59 830.73 566.38 784.30 478.47 678.48 492.89 687.67 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.4698 0.9396 1.4094 1.8792 2.349 SE +/- 0.00701, N = 3 SE +/- 0.01491, N = 4 SE +/- 0.01492, N = 3 SE +/- 0.00936, N = 3 SE +/- 0.00871, N = 3 SE +/- 0.01053, N = 3 SE +/- 0.00720, N = 3 SE +/- 0.00884, N = 3 1.64919 1.20298 1.76353 1.27368 2.08780 1.47249 2.02659 1.45280 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 Vehicle Bike Detection FP16 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 11.69 11.17 15.04 16.50 14.42 13.59 13.34 12.54 MIN: 10.28 / MAX: 42.12 MIN: 10.18 / MAX: 19.36 MIN: 12.32 / MAX: 32.87 MIN: 11.95 / MAX: 28.87 MIN: 9.74 / MAX: 43.95 MIN: 9.53 / MAX: 21.61 MIN: 9.67 / MAX: 39.84 MIN: 9.17 / MAX: 21.01 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2K 4K 6K 8K 10K SE +/- 3.37, N = 3 SE +/- 4.97, N = 3 SE +/- 8.98, N = 3 SE +/- 4.52, N = 3 SE +/- 3.66, N = 3 SE +/- 9.37, N = 3 SE +/- 4.05, N = 3 SE +/- 7.17, N = 3 10934.17 5725.03 7862.58 3606.12 7689.32 4093.73 7127.28 3802.61 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 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 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 6.95 6.72 7.72 9.79 7.97 7.76 7.32 7.13 MIN: 6.28 / MAX: 20.42 MIN: 6.22 / MAX: 11.15 MIN: 7.16 / MAX: 18.31 MIN: 4.37 / MAX: 18.18 MIN: 7.04 / MAX: 21.51 MIN: 6.86 / MAX: 12.92 MIN: 6.78 / MAX: 18.72 MIN: 6.64 / MAX: 12.22 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: Person Re-Identification Retail FP16 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4K 8K 12K 16K 20K SE +/- 7.74, N = 3 SE +/- 5.57, N = 3 SE +/- 4.14, N = 3 SE +/- 3.97, N = 3 SE +/- 7.22, N = 3 SE +/- 0.46, N = 3 SE +/- 3.67, N = 3 SE +/- 3.72, N = 3 18375.63 9515.86 15519.92 5943.66 14033.94 7207.74 13102.28 6717.65 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.28, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.14, N = 3 SE +/- 0.06, N = 3 SE +/- 0.21, N = 3 SE +/- 0.05, N = 3 SE +/- 0.15, N = 3 57.13 55.07 65.02 55.99 62.11 62.53 56.71 55.74 MIN: 39.58 / MAX: 104.91 MIN: 46.78 / MAX: 80.73 MIN: 55.15 / MAX: 104.43 MIN: 27.63 / MAX: 76.45 MIN: 47.82 / MAX: 101.26 MIN: 48.23 / MAX: 78.43 MIN: 45.43 / MAX: 89.46 MIN: 44.83 / MAX: 69.99 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 500 1000 1500 2000 2500 SE +/- 11.53, N = 3 SE +/- 0.47, N = 3 SE +/- 0.66, N = 3 SE +/- 2.65, N = 3 SE +/- 1.74, N = 3 SE +/- 2.94, N = 3 SE +/- 1.56, N = 3 SE +/- 2.29, N = 3 2237.69 1159.66 1843.61 1068.43 1801.89 894.87 1691.79 860.64 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 80K 160K 240K 320K 400K SE +/- 248.00, N = 3 SE +/- 63.14, N = 3 SE +/- 362.38, N = 3 SE +/- 105.60, N = 3 SE +/- 160.59, N = 3 SE +/- 40.48, N = 3 SE +/- 286.86, N = 3 SE +/- 155.19, N = 3 380867.5 195648.8 350187.5 179647.5 316359.2 162727.4 288129.1 143755.3 1. (CXX) g++ options: -O3 -march=native -fPIE -pie
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.09, N = 3 SE +/- 0.09, N = 3 SE +/- 0.12, N = 3 SE +/- 0.04, N = 3 SE +/- 0.33, N = 3 SE +/- 0.31, N = 3 SE +/- 0.40, N = 3 SE +/- 0.11, N = 3 28.85 29.02 29.75 29.41 36.00 33.99 36.23 33.33 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.11, N = 3 SE +/- 0.10, N = 3 SE +/- 0.13, N = 3 SE +/- 0.05, N = 3 SE +/- 0.26, N = 3 SE +/- 0.27, N = 3 SE +/- 0.31, N = 3 SE +/- 0.09, N = 3 34.66 34.46 33.61 34.00 27.78 29.43 27.61 30.00 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.09, N = 3 SE +/- 0.09, N = 3 SE +/- 0.12, N = 3 SE +/- 0.04, N = 3 SE +/- 0.33, N = 3 SE +/- 0.31, N = 3 SE +/- 0.40, N = 3 SE +/- 0.11, N = 3 28.85 29.02 29.75 29.41 36.00 33.99 36.23 33.33 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: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 SE +/- 0.32, N = 3 SE +/- 0.08, N = 3 SE +/- 0.55, N = 3 SE +/- 0.13, N = 3 31.53 29.96 33.29 30.92 40.52 35.11 40.45 35.07 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: Faster R-CNN R-50-FPN-int8 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 SE +/- 0.19, N = 3 SE +/- 0.07, N = 3 SE +/- 0.34, N = 3 SE +/- 0.10, N = 3 31.71 33.38 30.04 32.34 24.68 28.48 24.73 28.52 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 SE +/- 0.32, N = 3 SE +/- 0.08, N = 3 SE +/- 0.55, N = 3 SE +/- 0.13, N = 3 31.53 29.96 33.29 30.92 40.52 35.11 40.45 35.07 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 11 22 33 44 55 SE +/- 0.67, N = 12 SE +/- 0.53, N = 15 SE +/- 0.06, N = 3 SE +/- 0.35, N = 3 SE +/- 0.36, N = 5 SE +/- 0.21, N = 3 SE +/- 0.01, N = 3 SE +/- 0.34, N = 8 42.97 48.36 42.04 43.20 34.63 43.63 39.59 39.10 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.09, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.10, N = 3 SE +/- 0.10, N = 3 SE +/- 0.18, N = 3 43.41 39.62 45.78 100.39 64.91 59.69 42.16 41.93 MIN: 27.3 / MAX: 82.85 MIN: 26.71 / MAX: 62.24 MIN: 33.94 / MAX: 85.21 MIN: 41.76 / MAX: 139.86 MIN: 49.88 / MAX: 107.03 MIN: 42.86 / MAX: 84.25 MIN: 32.47 / MAX: 70.84 MIN: 27.14 / MAX: 54.05 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 160 320 480 640 800 SE +/- 1.42, N = 3 SE +/- 0.95, N = 3 SE +/- 0.69, N = 3 SE +/- 0.04, N = 3 SE +/- 0.55, N = 3 SE +/- 0.52, N = 3 SE +/- 1.28, N = 3 SE +/- 1.23, N = 3 736.23 403.20 654.84 298.56 569.60 301.30 568.81 285.95 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: GPT-2 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.02114, N = 3 SE +/- 0.00434, N = 3 SE +/- 0.03717, N = 3 SE +/- 0.01428, N = 3 SE +/- 0.01417, N = 3 SE +/- 0.07483, N = 3 SE +/- 0.08283, N = 3 SE +/- 0.02823, N = 3 5.10948 4.01687 6.90619 5.21520 7.30822 5.29027 7.34848 5.22781 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 SE +/- 0.81, N = 3 SE +/- 0.27, N = 3 SE +/- 0.78, N = 3 SE +/- 0.53, N = 3 SE +/- 0.26, N = 3 SE +/- 2.63, N = 3 SE +/- 1.53, N = 3 SE +/- 1.04, N = 3 195.45 248.68 144.64 191.56 136.69 188.93 135.98 191.11 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.02114, N = 3 SE +/- 0.00434, N = 3 SE +/- 0.03717, N = 3 SE +/- 0.01428, N = 3 SE +/- 0.01417, N = 3 SE +/- 0.07483, N = 3 SE +/- 0.08283, N = 3 SE +/- 0.02823, N = 3 5.10948 4.01687 6.90619 5.21520 7.30822 5.29027 7.34848 5.22781 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: Machine Translation EN To DE FP16 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.16, N = 3 SE +/- 0.03, N = 3 SE +/- 0.05, N = 3 SE +/- 0.34, N = 3 SE +/- 0.07, N = 3 SE +/- 0.12, N = 3 28.17 26.85 37.58 62.46 41.57 38.50 29.98 28.73 MIN: 20.12 / MAX: 261.31 MIN: 20.18 / MAX: 307.38 MIN: 20.58 / MAX: 347.67 MIN: 46.12 / MAX: 79.06 MIN: 28.11 / MAX: 155.06 MIN: 25.18 / MAX: 161.13 MIN: 20.48 / MAX: 222.7 MIN: 19.52 / MAX: 209.26 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200 400 600 800 1000 SE +/- 0.85, N = 3 SE +/- 0.54, N = 3 SE +/- 3.48, N = 3 SE +/- 0.17, N = 3 SE +/- 1.13, N = 3 SE +/- 4.09, N = 3 SE +/- 1.99, N = 3 SE +/- 1.21, N = 3 1128.78 592.20 793.64 477.77 889.18 467.08 795.63 415.96 1. (CXX) g++ options: -fPIC -fsigned-char -ffunction-sections -fdata-sections -O3 -fno-strict-overflow -fwrapv -shared -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.11, N = 3 SE +/- 0.11, N = 3 17.02 17.42 15.49 15.76 14.51 15.62 14.99 15.59
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 14.67 13.73 21.73 17.84 20.24 19.44 17.19 16.33 MIN: 12.73 / MAX: 41.28 MIN: 12.23 / MAX: 23.01 MIN: 15.08 / MAX: 39.82 MIN: 9.08 / MAX: 37.11 MIN: 13.69 / MAX: 36.9 MIN: 15.1 / MAX: 32.19 MIN: 13.16 / MAX: 42.02 MIN: 13.49 / MAX: 26.98 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2K 4K 6K 8K 10K SE +/- 6.25, N = 3 SE +/- 0.73, N = 3 SE +/- 11.51, N = 3 SE +/- 0.81, N = 3 SE +/- 8.70, N = 3 SE +/- 5.68, N = 3 SE +/- 4.11, N = 3 SE +/- 8.09, N = 3 8719.37 4656.92 5483.81 3325.90 5525.56 2877.32 5576.86 2935.47 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: yolov4 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.57, N = 3 SE +/- 0.30, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.13, N = 3 SE +/- 0.01, N = 3 SE +/- 0.74, N = 3 SE +/- 0.22, N = 3 70.44 47.90 78.40 55.12 79.38 61.01 76.42 61.83 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: yolov4 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.11, N = 3 SE +/- 0.13, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.13, N = 3 SE +/- 0.06, N = 3 14.20 20.88 12.75 18.14 12.60 16.39 13.09 16.17 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: yolov4 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.57, N = 3 SE +/- 0.30, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.13, N = 3 SE +/- 0.01, N = 3 SE +/- 0.74, N = 3 SE +/- 0.22, N = 3 70.44 47.90 78.40 55.12 79.38 61.01 76.42 61.83 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: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 16 32 48 64 80 SE +/- 0.12, N = 3 SE +/- 0.57, N = 3 SE +/- 0.11, N = 3 SE +/- 0.18, N = 3 SE +/- 0.23, N = 3 SE +/- 0.31, N = 3 SE +/- 0.50, N = 3 SE +/- 0.23, N = 3 64.79 50.83 68.83 62.13 73.97 63.54 72.80 61.25 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.03, N = 3 SE +/- 0.22, N = 3 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.08, N = 3 SE +/- 0.09, N = 3 SE +/- 0.06, N = 3 15.43 19.68 14.53 16.10 13.52 15.74 13.74 16.33 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: bertsquad-12 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 16 32 48 64 80 SE +/- 0.12, N = 3 SE +/- 0.57, N = 3 SE +/- 0.11, N = 3 SE +/- 0.18, N = 3 SE +/- 0.23, N = 3 SE +/- 0.31, N = 3 SE +/- 0.50, N = 3 SE +/- 0.23, N = 3 64.79 50.83 68.83 62.13 73.97 63.54 72.80 61.25 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-INT8 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 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 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 5.18 5.19 5.72 6.43 6.02 5.77 5.55 5.27 MIN: 4.74 / MAX: 26.35 MIN: 4.72 / MAX: 11.02 MIN: 5.1 / MAX: 17.93 MIN: 3.84 / MAX: 15.83 MIN: 5.37 / MAX: 27.77 MIN: 5.26 / MAX: 13.86 MIN: 5.12 / MAX: 24.99 MIN: 4.98 / MAX: 13.16 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5K 10K 15K 20K 25K SE +/- 17.69, N = 3 SE +/- 4.30, N = 3 SE +/- 12.64, N = 3 SE +/- 5.45, N = 3 SE +/- 3.89, N = 3 SE +/- 7.39, N = 3 SE +/- 12.64, N = 3 SE +/- 18.13, N = 3 24676.86 12317.94 20937.58 9004.23 18561.91 9686.39 17252.27 9087.12 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: ArcFace ResNet-100 - Device: CPU - Executor: Parallel Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.28, N = 3 SE +/- 0.07, N = 3 SE +/- 0.11, N = 3 SE +/- 0.04, N = 3 SE +/- 0.13, N = 3 SE +/- 0.12, N = 3 SE +/- 0.11, N = 3 SE +/- 0.17, N = 3 31.79 16.55 34.31 21.34 34.25 23.66 30.47 24.55 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.27, N = 3 SE +/- 0.27, N = 3 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 SE +/- 0.11, N = 3 SE +/- 0.21, N = 3 SE +/- 0.12, N = 3 SE +/- 0.28, N = 3 31.46 60.41 29.15 46.85 29.20 42.27 32.81 40.73 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.28, N = 3 SE +/- 0.07, N = 3 SE +/- 0.11, N = 3 SE +/- 0.04, N = 3 SE +/- 0.13, N = 3 SE +/- 0.12, N = 3 SE +/- 0.11, N = 3 SE +/- 0.17, N = 3 31.79 16.55 34.31 21.34 34.25 23.66 30.47 24.55 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: Handwritten English Recognition FP16-INT8 - Device: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 22.96 21.87 25.12 36.20 25.72 24.94 24.00 23.12 MIN: 20.8 / MAX: 34.86 MIN: 20.16 / MAX: 32.36 MIN: 19.86 / MAX: 59.96 MIN: 20.73 / MAX: 45.28 MIN: 20.7 / MAX: 44.98 MIN: 19.75 / MAX: 33.18 MIN: 19.19 / MAX: 45.88 MIN: 18.74 / MAX: 31.08 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1200 2400 3600 4800 6000 SE +/- 3.43, N = 3 SE +/- 0.73, N = 3 SE +/- 2.46, N = 3 SE +/- 1.46, N = 3 SE +/- 2.81, N = 3 SE +/- 4.55, N = 3 SE +/- 1.09, N = 3 SE +/- 3.41, N = 3 5570.63 2924.66 4744.05 1643.43 4351.59 2243.84 3998.02 2074.30 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: Sharpen Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 60 120 180 240 300 SE +/- 1.67, N = 3 SE +/- 0.00, N = 3 SE +/- 1.67, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 SE +/- 1.20, N = 3 SE +/- 0.00, N = 3 284 167 265 157 227 130 199 114 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.88, N = 3 SE +/- 0.00, N = 3 SE +/- 1.15, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.58, N = 3 SE +/- 0.33, N = 3 194 171 182 158 164 158 150 150 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: Resizing Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.58, N = 3 SE +/- 0.33, N = 3 SE +/- 0.58, N = 3 96 230 95 229 97 213 108 245 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.06875, N = 3 SE +/- 0.02630, N = 3 SE +/- 0.02968, N = 3 SE +/- 0.00544, N = 3 SE +/- 0.03080, N = 3 SE +/- 0.03023, N = 3 SE +/- 0.01680, N = 3 SE +/- 0.01698, N = 3 6.49693 3.23780 6.39031 3.19298 6.51744 3.32030 6.00525 3.13366 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 70 140 210 280 350 SE +/- 1.63, N = 3 SE +/- 2.50, N = 3 SE +/- 0.73, N = 3 SE +/- 0.53, N = 3 SE +/- 0.73, N = 3 SE +/- 2.72, N = 3 SE +/- 0.47, N = 3 SE +/- 1.74, N = 3 153.92 308.80 156.46 313.10 153.41 301.14 166.49 319.04 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.06875, N = 3 SE +/- 0.02630, N = 3 SE +/- 0.02968, N = 3 SE +/- 0.00544, N = 3 SE +/- 0.03080, N = 3 SE +/- 0.03023, N = 3 SE +/- 0.01680, N = 3 SE +/- 0.01698, N = 3 6.49693 3.23780 6.39031 3.19298 6.51744 3.32030 6.00525 3.13366 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 60 120 180 240 300 SE +/- 0.33, N = 3 SE +/- 0.67, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 2.40, N = 3 SE +/- 0.58, N = 3 SE +/- 1.20, N = 3 SE +/- 1.00, N = 3 256 285 236 265 207 277 193 286 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: Enhanced Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 80 160 240 320 400 SE +/- 1.15, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 SE +/- 1.00, N = 3 SE +/- 0.00, N = 3 SE +/- 2.60, N = 3 SE +/- 0.00, N = 3 374 228 342 216 301 181 269 160 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.3244 2.6488 3.9732 5.2976 6.622 SE +/- 0.02988, N = 3 SE +/- 0.03569, N = 3 SE +/- 0.02062, N = 3 SE +/- 0.05586, N = 3 SE +/- 0.02155, N = 3 SE +/- 0.01694, N = 3 SE +/- 0.03854, N = 3 SE +/- 0.02534, N = 3 4.65431 4.26352 5.29287 4.86604 5.88628 5.26075 5.74921 5.10914 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50 100 150 200 250 SE +/- 1.38, N = 3 SE +/- 1.98, N = 3 SE +/- 0.74, N = 3 SE +/- 2.34, N = 3 SE +/- 0.62, N = 3 SE +/- 0.61, N = 3 SE +/- 1.17, N = 3 SE +/- 0.98, N = 3 214.81 234.54 188.89 205.53 169.85 190.06 173.91 195.71 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.3244 2.6488 3.9732 5.2976 6.622 SE +/- 0.02988, N = 3 SE +/- 0.03569, N = 3 SE +/- 0.02062, N = 3 SE +/- 0.05586, N = 3 SE +/- 0.02155, N = 3 SE +/- 0.01694, N = 3 SE +/- 0.03854, N = 3 SE +/- 0.02534, N = 3 4.65431 4.26352 5.29287 4.86604 5.88628 5.26075 5.74921 5.10914 1. (CXX) g++ options: -O3 -march=native -ffunction-sections -fdata-sections -mtune=native -flto=auto -fno-fat-lto-objects -ldl -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 150 300 450 600 750 SE +/- 0.58, N = 3 SE +/- 0.88, N = 3 SE +/- 1.76, N = 3 SE +/- 1.15, N = 3 SE +/- 0.67, N = 3 SE +/- 0.67, N = 3 SE +/- 1.33, N = 3 SE +/- 0.58, N = 3 689 529 617 502 558 432 509 394 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 Read Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 120M 240M 360M 480M 600M SE +/- 82996.27, N = 3 SE +/- 1119855.47, N = 3 SE +/- 1411218.60, N = 3 SE +/- 3338569.32, N = 3 SE +/- 785893.13, N = 3 SE +/- 86315.43, N = 3 SE +/- 1120857.00, N = 3 SE +/- 81327.78, N = 3 578392800 297517946 525049998 265366170 473755946 234469939 405533173 205346154 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.22, N = 3 SE +/- 0.11, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.25, N = 3 SE +/- 0.40, N = 3 SE +/- 0.38, N = 3 SE +/- 0.20, N = 3 33.61 63.61 35.91 68.61 40.38 76.87 44.71 86.67
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: 256 - Model: ResNet-50 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.29, N = 3 SE +/- 0.64, N = 3 SE +/- 0.49, N = 3 SE +/- 0.52, N = 3 SE +/- 0.45, N = 3 SE +/- 0.62, N = 3 SE +/- 0.52, N = 3 SE +/- 0.76, N = 3 41.45 63.60 37.44 54.84 37.91 51.56 41.68 52.97 MIN: 38.42 / MAX: 42.93 MIN: 37.79 / MAX: 65.67 MIN: 26.88 / MAX: 41.26 MIN: 46.5 / MAX: 58.5 MIN: 16.15 / MAX: 40.53 MIN: 30.49 / MAX: 54.7 MIN: 37.38 / MAX: 44.02 MIN: 33.28 / MAX: 55.78
OpenBenchmarking.org batches/sec, More Is Better PyTorch 2.2.1 Device: CPU - Batch Size: 64 - Model: ResNet-50 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.12, N = 3 SE +/- 0.12, N = 3 SE +/- 0.50, N = 3 SE +/- 0.24, N = 3 SE +/- 0.46, N = 4 SE +/- 0.62, N = 3 SE +/- 0.51, N = 3 SE +/- 0.71, N = 3 40.77 63.06 37.08 56.01 38.30 51.35 42.35 51.55 MIN: 16.88 / MAX: 42.06 MIN: 33.2 / MAX: 65.54 MIN: 26.46 / MAX: 40.15 MIN: 51.84 / MAX: 59.12 MIN: 27.11 / MAX: 40.39 MIN: 41.6 / MAX: 53.81 MIN: 36.76 / MAX: 44.4 MIN: 37.49 / MAX: 54.33
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 SE +/- 0.45, N = 3 SE +/- 0.03, N = 3 SE +/- 0.24, N = 3 SE +/- 0.02, N = 3 SE +/- 0.40, N = 4 SE +/- 0.17, N = 3 SE +/- 0.31, N = 15 SE +/- 0.15, N = 3 31.90 39.38 26.58 39.48 32.44 42.22 32.25 41.43 1. (F9X) gfortran options: -O3 -march=native -funroll-loops -fopenmp
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 0.7 Test: llava-v1.5-7b-q4 - Acceleration: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 1.08, N = 12 SE +/- 0.17, N = 4 SE +/- 0.09, N = 15 SE +/- 0.15, N = 4 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 SE +/- 0.10, N = 3 7.49 16.61 10.44 20.95 7.26 8.13 7.77 7.67
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7K 14K 21K 28K 35K SE +/- 356.81, N = 15 SE +/- 298.34, N = 15 SE +/- 312.17, N = 3 SE +/- 236.27, N = 4 SE +/- 490.54, N = 15 SE +/- 281.72, N = 15 SE +/- 175.78, N = 4 SE +/- 938.95, N = 15 30547.6 26378.1 30108.7 24688.5 11866.7 15081.0 30236.0 21286.6 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -mavx512f -mavx512cd -mavx512bw -mavx512dq -fno-trapping-math -fno-math-errno -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13 26 39 52 65 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.08, N = 3 SE +/- 0.17, N = 3 SE +/- 0.19, N = 3 SE +/- 0.08, N = 3 SE +/- 0.10, N = 3 SE +/- 0.07, N = 3 52.69 42.78 53.76 46.76 56.78 47.09 55.44 46.96 1. RawTherapee, version 5.10, command line.
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.10, N = 3 SE +/- 0.12, N = 3 SE +/- 0.55, N = 3 SE +/- 0.24, N = 3 SE +/- 0.27, N = 3 SE +/- 0.08, N = 3 SE +/- 0.08, N = 3 SE +/- 0.03, N = 3 38.05 43.05 41.56 47.78 43.91 48.49 42.87 54.38 1. (CC) gcc options: -shared -lxc -lblas -lmpi
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.12, N = 3 SE +/- 0.08, N = 3 SE +/- 0.16, N = 3 SE +/- 0.02, N = 3 SE +/- 0.13, N = 3 SE +/- 0.13, N = 3 SE +/- 0.11, N = 3 SE +/- 0.05, N = 3 13.26 15.97 12.54 15.06 11.97 14.36 11.90 14.10 1. (CXX) g++ options: -O3 -flto=auto -fno-fat-lto-objects
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1M 2.2M 3.3M 4.4M 5.5M SE +/- 9700.21, N = 3 SE +/- 4635.95, N = 3 SE +/- 12743.41, N = 3 SE +/- 9256.47, N = 3 SE +/- 20649.43, N = 3 SE +/- 5202.86, N = 3 SE +/- 7219.73, N = 3 SE +/- 1747.99, N = 3 4966250.78 2369998.36 3902076.24 2194630.60 3279262.53 1855135.33 2993136.60 1614763.70 1. (CC) gcc options: -O2 -lrt" -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10K 20K 30K 40K 50K SE +/- 1633.92, N = 15 SE +/- 135.77, N = 4 SE +/- 467.78, N = 4 SE +/- 11.14, N = 4 SE +/- 283.44, N = 3 SE +/- 572.00, N = 12 SE +/- 1097.75, N = 15 SE +/- 601.03, N = 12 42114.6 31830.2 46877.5 35230.1 21583.2 18621.2 19876.5 20091.1 1. (CXX) g++ options: -O3 -fopenmp -lmpi_cxx -lmpi
Timed PHP Compilation This test times how long it takes to build PHP. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed PHP Compilation 8.3.4 Time To Compile Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 11 22 33 44 55 SE +/- 0.09, N = 3 SE +/- 0.31, N = 3 SE +/- 0.07, N = 3 SE +/- 0.16, N = 3 SE +/- 0.07, N = 3 SE +/- 0.16, N = 3 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 40.38 41.85 46.64 48.01 46.57 47.95 46.87 48.71
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200K 400K 600K 800K 1000K SE +/- 15579.94, N = 15 SE +/- 864.17, N = 3 SE +/- 2519.00, N = 3 SE +/- 1599.31, N = 3 SE +/- 788.50, N = 3 SE +/- 1274.00, N = 3 SE +/- 5122.51, N = 3 SE +/- 987.33, N = 3 797093 428116 793253 401278 657889 334076 562875 286552 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -lm -lrt -lz -ldl -lcrypt -lbz2
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 120K 240K 360K 480K 600K SE +/- 28320.74, N = 3 SE +/- 119.21, N = 3 SE +/- 185.14, N = 3 SE +/- 61.74, N = 3 SE +/- 6276.88, N = 3 SE +/- 374.59, N = 3 SE +/- 2598.93, N = 3 SE +/- 916.22, N = 3 573410 338146 557054 287862 469131 243347 401707 207738 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 22.01 Test: Compression Rating Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 140K 280K 420K 560K 700K SE +/- 709.87, N = 3 SE +/- 1982.97, N = 3 SE +/- 4200.62, N = 3 SE +/- 1535.53, N = 3 SE +/- 3924.87, N = 3 SE +/- 2258.81, N = 3 SE +/- 5127.15, N = 3 SE +/- 1836.55, N = 3 674239 441636 592729 388190 525514 342363 491729 314716 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.04, N = 3 SE +/- 0.15, N = 4 SE +/- 0.10, N = 15 SE +/- 0.05, N = 4 SE +/- 0.03, N = 3 SE +/- 0.06, N = 4 SE +/- 0.07, N = 15 SE +/- 0.12, N = 7 7.22 14.28 5.96 13.54 6.51 12.97 9.64 12.80 MIN: 7.13 / MAX: 7.4 MIN: 12.85 / MAX: 14.54 MIN: 5.5 / MAX: 7.04 MIN: 13.42 / MAX: 13.68 MIN: 6.27 / MAX: 6.67 MIN: 12.11 / MAX: 13.11 MIN: 8.98 / MAX: 10.13 MIN: 11.39 / MAX: 13.21
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4M 8M 12M 16M 20M SE +/- 162096.41, N = 15 SE +/- 9492.69, N = 3 SE +/- 38670.98, N = 3 SE +/- 8685.88, N = 3 SE +/- 7571.88, N = 3 SE +/- 30179.10, N = 3 SE +/- 17457.89, N = 3 SE +/- 5238.74, N = 3 18200533 18774333 16843333 16882333 16884000 16885333 16904667 16906333 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 SE +/- 0.23, N = 7 SE +/- 0.39, N = 4 SE +/- 0.43, N = 3 SE +/- 0.55, N = 3 SE +/- 0.31, N = 5 SE +/- 0.57, N = 3 SE +/- 0.40, N = 4 SE +/- 0.59, N = 3 25.87 34.59 29.98 39.45 30.53 40.50 32.14 43.42
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.135 0.27 0.405 0.54 0.675 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 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.59 0.60 0.51 0.52 0.54 0.55 0.54 0.55 1. (CC) gcc options: -fvisibility=hidden -O2 -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: 512 - Model: AlexNet Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400 800 1200 1600 2000 SE +/- 19.23, N = 3 SE +/- 1.84, N = 3 SE +/- 18.52, N = 3 SE +/- 0.51, N = 3 SE +/- 8.55, N = 3 SE +/- 5.13, N = 3 SE +/- 17.88, N = 4 SE +/- 3.89, N = 3 2039.63 1486.12 1899.64 1278.45 1677.62 1108.95 1604.11 1047.96
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 11 22 33 44 55 SE +/- 0.50, N = 3 SE +/- 0.49, N = 14 SE +/- 0.12, N = 3 SE +/- 0.17, N = 3 SE +/- 0.32, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 39.52 46.32 40.08 42.91 33.27 40.35 34.83 43.36 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 SE +/- 0.07, N = 3 25.02 48.41 27.11 51.82 30.20 58.16 33.92 67.32 1. (CXX) g++ options: -O3
Graph500 This is a benchmark of the reference implementation of Graph500, an HPC benchmark focused on data intensive loads and commonly tested on supercomputers for complex data problems. Graph500 primarily stresses the communication subsystem of the hardware under test. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org sssp median_TEPS, More Is Better Graph500 3.0 Scale: 26 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ 100M 200M 300M 400M 500M 484977000 416523000 1. (CC) gcc options: -fcommon -O3 -lpthread -lm -lmpi Scale: 26
Xeon Platinum 8490H: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Platinum 8490H 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
OpenBenchmarking.org sssp max_TEPS, More Is Better Graph500 3.0 Scale: 26 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ 140M 280M 420M 560M 700M 667365000 539613000 1. (CC) gcc options: -fcommon -O3 -lpthread -lm -lmpi Scale: 26
Xeon Platinum 8490H: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Platinum 8490H 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
OpenBenchmarking.org bfs median_TEPS, More Is Better Graph500 3.0 Scale: 26 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ 300M 600M 900M 1200M 1500M 1204520000 1244300000 1. (CC) gcc options: -fcommon -O3 -lpthread -lm -lmpi Scale: 26
Xeon Platinum 8490H: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Platinum 8490H 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
OpenBenchmarking.org bfs max_TEPS, More Is Better Graph500 3.0 Scale: 26 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ 300M 600M 900M 1200M 1500M 1267480000 1299170000 1. (CC) gcc options: -fcommon -O3 -lpthread -lm -lmpi Scale: 26
Xeon Platinum 8490H: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Platinum 8490H 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9468 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
Xeon Max 9480 2P: The test quit with a non-zero exit status. E: AML: Fatal: non power2 groupsize unsupported. Define macro PROCS_PER_NODE_NOT_POWER_OF_TWO to override
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: 57 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 16M 32M 48M 64M 80M SE +/- 68119.99, N = 3 SE +/- 599433.09, N = 9 SE +/- 3844.19, N = 3 SE +/- 666.67, N = 3 SE +/- 822374.71, N = 4 SE +/- 5364.49, N = 3 SE +/- 4163.33, N = 3 SE +/- 4910.31, N = 3 70499000 73140667 66991333 67074667 66266000 67127667 67064000 67130333 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1200 2400 3600 4800 6000 SE +/- 0.12, N = 3 SE +/- 18.09, N = 3 SE +/- 47.68, N = 6 SE +/- 18.43, N = 3 SE +/- 0.15, N = 3 SE +/- 18.30, N = 3 SE +/- 26.47, N = 3 SE +/- 0.09, N = 3 5493.4 3088.7 4853.6 2616.9 4386.6 2514.0 3787.0 2170.7 MIN: 3703.4 / MAX: 5493.6 MIN: 1852.6 / MAX: 3124.9 MIN: 3181.8 / MAX: 4929.4 MIN: 1591.7 / MAX: 2653.8 MIN: 2970.2 / MAX: 4386.9 MIN: 1550.4 / MAX: 2532.5 MIN: 2545.7 / MAX: 3839.9 MIN: 1388.2 / MAX: 2170.8 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -mavx512f -mavx512cd -mavx512bw -mavx512dq -fno-trapping-math -fno-math-errno -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 SE +/- 0.16, N = 3 SE +/- 0.04, N = 3 SE +/- 0.06, N = 3 SE +/- 0.31, N = 3 SE +/- 0.22, N = 3 SE +/- 0.21, N = 3 33.65 34.79 37.31 37.75 38.14 39.88 38.65 41.13 1. (CXX) g++ options: -O3 -fPIC -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15K 30K 45K 60K 75K SE +/- 1080.03, N = 3 SE +/- 219.20, N = 3 SE +/- 412.68, N = 4 SE +/- 76.67, N = 3 SE +/- 802.88, N = 3 SE +/- 0.00, N = 3 SE +/- 194.30, N = 3 SE +/- 99.83, N = 3 72004.0 37265.9 41773.0 21314.2 45195.1 23250.2 42939.2 21766.5 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
OpenBenchmarking.org Mpix/sec, More Is Better ASKAP 1.0 Test: tConvolve MPI - Degridding Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13K 26K 39K 52K 65K SE +/- 480.95, N = 3 SE +/- 153.57, N = 3 SE +/- 378.22, N = 4 SE +/- 93.91, N = 3 SE +/- 0.00, N = 3 SE +/- 76.47, N = 3 SE +/- 143.60, N = 3 SE +/- 0.00, N = 3 59139.2 31177.2 34614.1 17891.6 40368.6 20561.4 36900.5 18522.0 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.32, N = 3 SE +/- 0.18, N = 3 SE +/- 0.10, N = 3 SE +/- 0.08, N = 3 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.08, N = 3 26.88 27.90 28.22 31.49 27.63 31.92 27.51 33.75
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: 64 - Buffer Length: 256 - Filter Length: 512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200M 400M 600M 800M 1000M SE +/- 7562646.32, N = 8 SE +/- 1515552.70, N = 3 SE +/- 3717554.45, N = 3 SE +/- 940265.92, N = 3 SE +/- 6141471.96, N = 3 SE +/- 1289603.73, N = 3 SE +/- 4432294.36, N = 3 SE +/- 1611276.24, N = 3 883701250 783770000 799496667 744410000 757176667 689003333 732560000 632396667 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9M 18M 27M 36M 45M SE +/- 419600.12, N = 5 SE +/- 417042.80, N = 5 SE +/- 1154.70, N = 3 SE +/- 1527.53, N = 3 SE +/- 1452.97, N = 3 SE +/- 3000.00, N = 3 SE +/- 1201.85, N = 3 SE +/- 1666.67, N = 3 38950600 40244000 36461000 36506000 36521333 36539000 36497667 36537667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.11, N = 3 SE +/- 0.03, N = 3 SE +/- 0.17, N = 3 SE +/- 0.16, N = 3 SE +/- 0.12, N = 3 SE +/- 0.16, N = 3 30.47 30.17 35.46 34.93 35.70 35.20 36.20 35.66 1. (CC) gcc options: -m64 -ldl -lgcc_s -lutil -lrt -lpthread -lm -lc -pie -nodefaultlibs
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.52, N = 3 SE +/- 0.16, N = 3 SE +/- 0.49, N = 3 SE +/- 0.03, N = 3 SE +/- 0.44, N = 3 SE +/- 0.36, N = 3 SE +/- 0.38, N = 8 SE +/- 0.28, N = 3 45.69 77.52 35.53 51.46 35.46 48.59 46.49 50.99 MIN: 13.24 / MAX: 48.92 MIN: 36.51 / MAX: 79.65 MIN: 16.36 / MAX: 38.41 MIN: 35.78 / MAX: 55.29 MIN: 0.7 / MAX: 41 MIN: 4.58 / MAX: 55.2 MIN: 15.6 / MAX: 50.65 MIN: 11.73 / MAX: 59.01
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13 26 39 52 65 43.70 48.43 48.56 53.49 49.14 56.40 51.25 58.38
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.0007, N = 3 SE +/- 0.0294, N = 3 SE +/- 0.0379, N = 3 SE +/- 0.0054, N = 3 SE +/- 0.0022, N = 3 SE +/- 0.0016, N = 3 SE +/- 0.0055, N = 3 SE +/- 0.0016, N = 3 14.6334 7.6240 13.7518 7.0885 11.8875 5.9702 10.2051 5.1342 1. (CXX) g++ options: -O3 -flto -pthread
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: 64 - Buffer Length: 256 - Filter Length: 57 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 600M 1200M 1800M 2400M 3000M SE +/- 32405829.38, N = 3 SE +/- 11426868.92, N = 3 SE +/- 28500058.48, N = 3 SE +/- 7076957.92, N = 3 SE +/- 27382304.50, N = 5 SE +/- 5493127.02, N = 3 SE +/- 28812627.44, N = 4 SE +/- 2629533.12, N = 3 2620066667 2475500000 2655500000 2539200000 2491340000 2308966667 2494050000 2138766667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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: Slow Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 27.11 16.06 24.71 15.71 22.14 13.89 19.75 13.36
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: Exhaustive Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.0157, N = 3 SE +/- 0.0072, N = 3 SE +/- 0.0164, N = 3 SE +/- 0.0024, N = 3 SE +/- 0.0016, N = 3 SE +/- 0.0011, N = 3 SE +/- 0.0033, N = 3 SE +/- 0.0003, N = 3 8.9091 4.6303 8.3701 4.2689 7.2075 3.6276 6.2126 3.1204 1. (CXX) g++ options: -O3 -flto -pthread
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: 128 - Buffer Length: 256 - Filter Length: 512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 300M 600M 900M 1200M 1500M SE +/- 4790383.89, N = 3 SE +/- 2313246.88, N = 3 SE +/- 6842108.19, N = 3 SE +/- 1740596.96, N = 3 SE +/- 2123152.79, N = 3 SE +/- 1345523.44, N = 3 SE +/- 11560043.25, N = 5 SE +/- 681281.47, N = 3 1493833333 1152933333 1126466667 918566667 1206066667 882200000 1091460000 800186667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 2.84, N = 15 SE +/- 2.83, N = 15 SE +/- 2.18, N = 15 SE +/- 2.31, N = 15 SE +/- 1.67, N = 15 SE +/- 1.62, N = 15 SE +/- 1.63, N = 15 SE +/- 1.44, N = 15 162.98 174.83 154.50 162.47 124.26 150.00 122.62 144.70 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 800 1600 2400 3200 4000 SE +/- 30.36, N = 12 SE +/- 33.58, N = 3 SE +/- 25.40, N = 3 SE +/- 28.42, N = 3 SE +/- 28.57, N = 3 SE +/- 25.35, N = 3 SE +/- 28.11, N = 3 SE +/- 22.17, N = 3 3586.3 3600.5 3210.1 3209.6 3222.8 3229.1 3207.0 3225.7 1. (CXX) g++ options: -O3 -march=native -fPIE -pie
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: 128 - Buffer Length: 256 - Filter Length: 57 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1000M 2000M 3000M 4000M 5000M SE +/- 51282832.98, N = 4 SE +/- 5260228.13, N = 3 SE +/- 13655808.69, N = 3 SE +/- 3116800.35, N = 3 SE +/- 14772609.79, N = 3 SE +/- 3119294.79, N = 3 SE +/- 25888414.40, N = 3 SE +/- 2030052.00, N = 3 4450525000 3503200000 3723666667 3181266667 3896500000 2972400000 3596400000 2670033333 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Max 9480 90K 180K 270K 360K 450K SE +/- 576.43, N = 2 SE +/- 142.71, N = 3 SE +/- 125.38, N = 3 381352.69 399698.78 367199.06 1. (CC) gcc options: -lluajit-5.1 -lm -lssl -lcrypto -lpthread -ldl -std=c99 -O2 Connections: 1000
Xeon Platinum 8490H: The test quit with a non-zero exit status.
Xeon Platinum 8490H 2P: The test quit with a non-zero exit status.
Xeon Max 9468: The test quit with a non-zero exit status.
Xeon Max 9468 2P: The test quit with a non-zero exit status.
Xeon Max 9480 2P: The test quit with a non-zero exit status.
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: 256 - Buffer Length: 256 - Filter Length: 512 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400M 800M 1200M 1600M 2000M SE +/- 14465476.14, N = 3 SE +/- 240370.09, N = 3 SE +/- 5630373.98, N = 3 SE +/- 1722723.94, N = 3 SE +/- 6728628.72, N = 3 SE +/- 740202.67, N = 3 SE +/- 4238841.56, N = 3 SE +/- 569766.03, N = 3 2027000000 1178766667 1616966667 1020833333 1518933333 944550000 1444066667 838530000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 SE +/- 0.04, N = 3 SE +/- 0.24, N = 3 SE +/- 0.19, N = 3 SE +/- 0.10, N = 3 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 18.31 32.08 21.15 35.96 23.45 39.17 24.99 42.85
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.D Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3K 6K 9K 12K 15K SE +/- 30.19, N = 4 SE +/- 22.33, N = 3 SE +/- 100.23, N = 4 SE +/- 6.77, N = 3 SE +/- 10.93, N = 4 SE +/- 6.53, N = 3 SE +/- 4.83, N = 4 SE +/- 155.10, N = 12 16004.07 8089.14 14946.76 7562.80 12793.01 6445.09 10970.45 5366.69 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
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: bcrypt Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40K 80K 120K 160K 200K SE +/- 288.38, N = 3 SE +/- 46.28, N = 3 SE +/- 292.24, N = 3 SE +/- 10.67, N = 3 SE +/- 576.80, N = 3 SE +/- 8.67, N = 3 SE +/- 866.68, N = 3 SE +/- 8.84, N = 3 206778 105781 183603 94551 155959 79144 133866 67875 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40K 80K 120K 160K 200K SE +/- 649.08, N = 3 SE +/- 79.75, N = 3 SE +/- 348.45, N = 3 SE +/- 11.39, N = 3 SE +/- 169.28, N = 3 SE +/- 8.67, N = 3 SE +/- 116.69, N = 3 SE +/- 17.33, N = 3 205499 105831 183583 94513 156658 79152 134562 67893 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: 256 - Buffer Length: 256 - Filter Length: 57 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1300M 2600M 3900M 5200M 6500M SE +/- 61971606.40, N = 3 SE +/- 3555434.03, N = 3 SE +/- 2331189.49, N = 3 SE +/- 5292237.50, N = 3 SE +/- 28453880.81, N = 3 SE +/- 1976810.00, N = 3 SE +/- 17558125.69, N = 3 SE +/- 4088330.28, N = 3 6284300000 3806833333 5361266667 3500833333 5036000000 3222533333 4825133333 2848933333 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
OpenBenchmarking.org samples/s, More Is Better Liquid-DSP 1.6 Threads: 256 - Buffer Length: 256 - Filter Length: 32 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1300M 2600M 3900M 5200M 6500M SE +/- 6711433.03, N = 3 SE +/- 2516611.48, N = 3 SE +/- 10496401.50, N = 3 SE +/- 4021608.30, N = 3 SE +/- 895048.11, N = 3 SE +/- 1338323.99, N = 3 SE +/- 2447674.63, N = 3 SE +/- 1320353.49, N = 3 6077300000 3167200000 6116566667 3243800000 5367033333 2769066667 4670966667 2381800000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.05, N = 3 SE +/- 0.16, N = 3 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 30.37 16.64 27.65 17.85 24.65 15.68 22.25 15.17
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: 128 - Buffer Length: 256 - Filter Length: 32 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 800M 1600M 2400M 3200M 4000M SE +/- 13813198.20, N = 3 SE +/- 2236316.42, N = 3 SE +/- 26633124.74, N = 3 SE +/- 1858314.65, N = 3 SE +/- 2403700.85, N = 3 SE +/- 1342054.81, N = 3 SE +/- 4598671.31, N = 3 SE +/- 400000.00, N = 3 3520766667 3148233333 3558600000 3200600000 3337066667 2739633333 3254033333 2351900000 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
Google SynthMark SynthMark is a cross platform tool for benchmarking CPU performance under a variety of real-time audio workloads. It uses a polyphonic synthesizer model to provide standardized tests for latency, jitter and computational throughput. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Voices, More Is Better Google SynthMark 20201109 Test: VoiceMark_100 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200 400 600 800 1000 SE +/- 0.38, N = 3 SE +/- 0.13, N = 3 SE +/- 1.51, N = 3 SE +/- 0.22, N = 3 SE +/- 0.48, N = 3 SE +/- 0.24, N = 3 SE +/- 0.30, N = 3 SE +/- 0.26, N = 3 799.49 799.72 712.70 711.38 712.09 712.01 710.92 711.10 1. (CXX) g++ options: -lm -lpthread -std=c++11 -Ofast
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: 64 - Buffer Length: 256 - Filter Length: 32 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 500M 1000M 1500M 2000M 2500M SE +/- 17236814.61, N = 3 SE +/- 27191440.65, N = 3 SE +/- 3300505.01, N = 3 SE +/- 3167192.94, N = 3 SE +/- 3017909.50, N = 3 SE +/- 692820.32, N = 3 SE +/- 2206304.10, N = 3 SE +/- 688799.28, N = 3 2154633333 1927766667 2025000000 1890566667 2090466667 1679700000 1963233333 1640266667 1. (CC) gcc options: -O3 -pthread -lm -lc -lliquid
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 SE +/- 0.04, N = 3 SE +/- 0.06, N = 3 SE +/- 0.08, N = 3 SE +/- 0.11, N = 3 SE +/- 0.18, N = 3 SE +/- 0.04, N = 3 SE +/- 0.13, N = 3 SE +/- 0.17, N = 3 17.29 31.91 19.13 34.83 21.45 38.48 23.17 42.78
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 SE +/- 0.12, N = 6 SE +/- 0.21, N = 6 SE +/- 0.23, N = 15 SE +/- 0.19, N = 6 SE +/- 0.23, N = 12 SE +/- 0.27, N = 6 SE +/- 0.20, N = 15 SE +/- 0.27, N = 7 12.69 21.67 13.50 33.56 14.67 31.47 18.24 36.13
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 800 1600 2400 3200 4000 SE +/- 12.44, N = 3 SE +/- 18.19, N = 3 SE +/- 41.09, N = 5 SE +/- 11.79, N = 3 SE +/- 11.86, N = 3 SE +/- 42.24, N = 4 SE +/- 10.84, N = 3 SE +/- 33.41, N = 3 3345 3242 3689 3581 3537 3396 3643 3378
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.29, N = 4 SE +/- 0.35, N = 3 SE +/- 0.22, N = 4 SE +/- 0.09, N = 3 SE +/- 0.89, N = 15 SE +/- 0.09, N = 3 SE +/- 0.76, N = 3 SE +/- 0.08, N = 3 173.22 97.19 132.28 79.38 112.72 71.83 104.84 64.96 MIN: 164.07 / MAX: 187.8 MIN: 92.96 / MAX: 102.04 MIN: 125 / MAX: 146.52 MIN: 76.07 / MAX: 85.98 MIN: 102 / MAX: 125.96 MIN: 69.55 / MAX: 76.73 MIN: 101.05 / MAX: 112.02 MIN: 63.12 / MAX: 67.81
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.046, N = 3 SE +/- 0.009, N = 3 SE +/- 0.019, N = 3 SE +/- 0.010, N = 3 SE +/- 0.032, N = 3 SE +/- 0.007, N = 3 SE +/- 0.025, N = 3 SE +/- 0.019, N = 3 18.540 10.477 15.450 8.762 14.127 8.282 13.392 7.545 1. (CXX) g++ options: -O3 -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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.044, N = 3 SE +/- 0.005, N = 3 SE +/- 0.049, N = 3 SE +/- 0.030, N = 3 SE +/- 0.042, N = 3 SE +/- 0.047, N = 3 SE +/- 0.015, N = 3 SE +/- 0.027, N = 3 7.360 7.876 6.698 7.061 6.052 6.572 5.991 6.490 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 80 160 240 320 400 SE +/- 0.21, N = 3 SE +/- 1.12, N = 3 SE +/- 2.37, N = 3 SE +/- 1.62, N = 3 SE +/- 2.38, N = 3 SE +/- 0.83, N = 3 SE +/- 3.52, N = 5 SE +/- 1.94, N = 3 291.37 319.76 331.76 380.14 308.95 335.58 341.51 333.73
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 11K 22K 33K 44K 55K SE +/- 1365.55, N = 15 SE +/- 131.12, N = 15 SE +/- 948.80, N = 15 SE +/- 677.85, N = 15 SE +/- 955.08, N = 5 SE +/- 69.31, N = 5 SE +/- 0.00, N = 4 SE +/- 41.81, N = 5 48686.80 16936.80 51476.20 30471.50 7304.58 7046.56 4437.60 6758.81 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4K 8K 12K 16K 20K SE +/- 221.77, N = 15 SE +/- 211.64, N = 15 SE +/- 139.77, N = 15 SE +/- 195.05, N = 15 SE +/- 13.50, N = 5 SE +/- 33.09, N = 5 SE +/- 40.14, N = 4 SE +/- 13.21, N = 5 20204.90 19141.80 17010.80 19798.50 3836.73 4438.58 3916.78 4173.46 1. (CXX) g++ options: -O3 -fstrict-aliasing -fopenmp
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: Slow Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.04, N = 4 SE +/- 0.07, N = 3 SE +/- 0.08, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.03, N = 3 40.92 24.23 31.24 20.23 28.96 18.24 26.00 17.52 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 4K - Video Preset: Medium Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.27, N = 4 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.02, N = 3 SE +/- 0.08, N = 3 SE +/- 0.04, N = 3 41.58 24.80 32.04 20.97 29.62 18.84 26.69 18.13 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.79, N = 5 SE +/- 0.22, N = 5 SE +/- 0.46, N = 4 SE +/- 0.20, N = 5 SE +/- 0.44, N = 15 SE +/- 0.29, N = 4 SE +/- 0.58, N = 5 SE +/- 0.26, N = 4 70.43 79.44 65.22 70.91 55.98 65.36 54.82 62.75 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 4K - Video Preset: Ultra Fast Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 14 28 42 56 70 SE +/- 0.17, N = 5 SE +/- 0.04, N = 4 SE +/- 0.58, N = 5 SE +/- 0.07, N = 5 SE +/- 0.36, N = 15 SE +/- 0.13, N = 4 SE +/- 0.50, N = 5 SE +/- 0.06, N = 4 57.13 39.31 55.26 60.87 45.79 50.09 46.13 48.91
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 400 800 1200 1600 2000 SE +/- 1.86, N = 3 SE +/- 4.27, N = 3 SE +/- 5.06, N = 3 SE +/- 0.54, N = 3 SE +/- 5.81, N = 3 SE +/- 6.82, N = 3 SE +/- 1.15, N = 3 SE +/- 1.21, N = 3 1681.44 1277.46 1730.25 1205.30 1528.41 1058.99 1479.53 986.04
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 1.02, N = 8 SE +/- 0.20, N = 6 SE +/- 0.81, N = 7 SE +/- 0.61, N = 5 SE +/- 1.85, N = 12 SE +/- 0.84, N = 15 SE +/- 1.28, N = 12 SE +/- 0.12, N = 4 199.73 106.40 133.63 82.85 114.90 70.84 108.40 63.62 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1100 2200 3300 4400 5500 SE +/- 25.44, N = 8 SE +/- 5.07, N = 6 SE +/- 20.15, N = 7 SE +/- 15.23, N = 5 SE +/- 46.19, N = 12 SE +/- 21.10, N = 15 SE +/- 32.01, N = 12 SE +/- 2.91, N = 4 4993.34 2659.88 3340.69 2071.19 2872.56 1771.07 2709.93 1590.50 1. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -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: 1 - Model: AlexNet Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13 26 39 52 65 SE +/- 0.45, N = 15 SE +/- 0.66, N = 15 SE +/- 0.44, N = 15 SE +/- 0.08, N = 8 SE +/- 0.52, N = 15 SE +/- 1.22, N = 15 SE +/- 0.57, N = 15 SE +/- 0.76, N = 15 30.76 35.50 37.57 56.53 33.21 43.67 37.76 48.28
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.03, N = 4 SE +/- 0.04, N = 3 SE +/- 0.08, N = 4 SE +/- 0.15, N = 3 SE +/- 0.08, N = 4 SE +/- 0.24, N = 3 SE +/- 0.03, N = 3 SE +/- 0.11, N = 3 12.76 23.72 13.74 25.89 15.29 28.49 16.55 31.66
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.010, N = 6 SE +/- 0.030, N = 5 SE +/- 0.007, N = 6 SE +/- 0.005, N = 5 SE +/- 0.142, N = 12 SE +/- 0.026, N = 5 SE +/- 0.122, N = 15 SE +/- 0.045, N = 5 5.110 6.701 5.008 7.047 6.266 7.959 6.334 8.300
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 10 20 30 40 50 23.56 33.07 28.56 42.22 27.58 41.75 28.99 45.95 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfiniteVolume -lmeshTools -lparallel -llagrangian -lregionModels -lgenericPatchFields -lOpenFOAM -ldl -lm
OpenBenchmarking.org Seconds, Fewer Is Better OpenFOAM 10 Input: drivaerFastback, Small Mesh Size - Mesh Time Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 8 16 24 32 40 29.33 31.70 35.60 31.38 28.84 32.00 32.48 31.78 1. (CXX) g++ options: -std=c++14 -m64 -O3 -ftemplate-depth-100 -fPIC -fuse-ld=bfd -Xlinker --add-needed --no-as-needed -lfiniteVolume -lmeshTools -lparallel -llagrangian -lregionModels -lgenericPatchFields -lOpenFOAM -ldl -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: IS.D Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1600 3200 4800 6400 8000 SE +/- 155.00, N = 15 SE +/- 8.92, N = 4 SE +/- 23.92, N = 4 SE +/- 6.56, N = 3 SE +/- 4.96, N = 4 SE +/- 18.08, N = 4 SE +/- 12.99, N = 4 SE +/- 15.55, N = 4 7437.81 4059.70 3960.28 2865.72 3819.83 3076.66 3816.70 3050.16 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.21, N = 5 SE +/- 0.09, N = 4 SE +/- 0.50, N = 5 SE +/- 0.07, N = 5 SE +/- 0.49, N = 15 SE +/- 0.20, N = 5 SE +/- 0.49, N = 5 SE +/- 0.12, N = 5 66.78 39.20 69.05 66.74 54.19 54.47 52.63 53.13 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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 800M 1600M 2400M 3200M 4000M SE +/- 9014927.20, N = 3 SE +/- 4000850.45, N = 3 SE +/- 2438223.49, N = 3 SE +/- 468935.32, N = 3 SE +/- 9164253.67, N = 3 SE +/- 1778769.08, N = 3 SE +/- 7233468.75, N = 3 SE +/- 3359086.30, N = 3 3609646333 1814760000 3124402667 1605097000 3205161000 1597993333 3156261667 1588496667 1. (CC) gcc options: -lparcsr_ls -lparcsr_mv -lseq_mv -lIJ_mv -lkrylov -lHYPRE_utilities -lm -fopenmp -lmpi
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.03329814, N = 5 SE +/- 0.02923582, N = 4 SE +/- 0.04312801, N = 5 SE +/- 0.01004100, N = 4 SE +/- 0.05695382, N = 15 SE +/- 0.01708632, N = 4 SE +/- 0.01741826, N = 5 SE +/- 0.04445674, N = 4 5.39125967 10.19670660 6.39169817 12.92934130 6.66387510 12.79975820 7.29349117 13.96923760 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
Timed FFmpeg Compilation This test times how long it takes to build the FFmpeg multimedia library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed FFmpeg Compilation 7.0 Time To Compile Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 6 12 18 24 30 SE +/- 0.05, N = 3 SE +/- 0.07, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 SE +/- 0.09, N = 3 SE +/- 0.09, N = 3 SE +/- 0.03, N = 3 17.39 21.36 19.51 24.06 20.01 24.84 20.79 26.46
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.54 1.08 1.62 2.16 2.7 SE +/- 0.01, N = 4 SE +/- 0.00, N = 3 SE +/- 0.01, N = 4 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 2.40 1.58 2.18 1.29 1.74 1.14 1.72 1.08
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.32768, N = 3 SE +/- 0.08324, N = 3 SE +/- 0.10023, N = 3 SE +/- 0.00627, N = 3 SE +/- 0.09318, N = 3 SE +/- 0.02700, N = 3 SE +/- 0.10586, N = 3 SE +/- 0.00876, N = 3 38.08130 6.69692 16.55070 6.15856 16.09010 6.09747 13.45000 5.30953 MIN: 9.76 MIN: 3.74 MIN: 14.58 MIN: 9.89 MIN: 11.97 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -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 3.6 Video Input: Bosphorus 4K Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 9 18 27 36 45 SE +/- 0.09, N = 4 SE +/- 0.06, N = 4 SE +/- 0.20, N = 4 SE +/- 0.11, N = 4 SE +/- 0.10, N = 3 SE +/- 0.39, N = 3 SE +/- 0.12, N = 3 SE +/- 0.08, N = 3 39.31 38.93 37.04 37.86 27.90 31.40 27.26 29.05 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.16, N = 5 SE +/- 0.08, N = 3 SE +/- 0.50, N = 5 SE +/- 0.01, N = 4 SE +/- 0.38, N = 9 SE +/- 0.02, N = 4 SE +/- 0.39, N = 4 SE +/- 0.12, N = 4 53.66 35.10 54.88 49.42 46.25 40.82 44.75 39.56
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15K 30K 45K 60K 75K SE +/- 359.29, N = 3 SE +/- 437.51, N = 4 SE +/- 272.44, N = 4 SE +/- 182.38, N = 15 SE +/- 207.54, N = 5 SE +/- 217.67, N = 6 SE +/- 206.20, N = 5 SE +/- 211.12, N = 6 70826.73 38433.97 43724.06 23072.30 53269.47 25831.84 52102.67 23848.54 1. (CXX) g++ options: -O3 -fopenmp -lm -lmpi_cxx -lmpi
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.1762 0.3524 0.5286 0.7048 0.881 SE +/- 0.004443, N = 4 SE +/- 0.000767, N = 4 SE +/- 0.004208, N = 4 SE +/- 0.000915, N = 4 SE +/- 0.003500, N = 4 SE +/- 0.000666, N = 4 SE +/- 0.004822, N = 4 SE +/- 0.001869, N = 4 0.580664 0.398446 0.783180 0.539467 0.759434 0.617190 0.662387 0.725974 MIN: 0.51 MIN: 0.37 MIN: 0.69 MIN: 0.5 MIN: 0.67 MIN: 0.55 MIN: 0.58 MIN: 0.67 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 200 400 600 800 1000 SE +/- 2.22, N = 5 SE +/- 6.22, N = 5 SE +/- 7.98, N = 5 SE +/- 0.51, N = 5 SE +/- 7.35, N = 5 SE +/- 2.31, N = 5 SE +/- 6.39, N = 13 SE +/- 1.95, N = 5 858.84 819.89 996.47 969.62 940.40 852.36 932.59 841.68
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50K 100K 150K 200K 250K SE +/- 1528.09, N = 5 SE +/- 31.76, N = 4 SE +/- 560.20, N = 4 SE +/- 150.79, N = 3 SE +/- 420.81, N = 4 SE +/- 9.58, N = 3 SE +/- 679.14, N = 4 SE +/- 67.02, N = 3 228184.68 109416.42 147311.14 71245.26 141954.15 69776.41 135939.25 58815.36 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.018471, N = 4 SE +/- 0.002400, N = 3 SE +/- 0.016400, N = 4 SE +/- 0.002987, N = 3 SE +/- 0.003253, N = 4 SE +/- 0.013353, N = 3 SE +/- 0.003237, N = 4 SE +/- 0.001847, N = 3 9.964526 16.969100 11.284380 16.030800 12.106900 19.882460 13.211250 21.021120 1. (CXX) g++ options: -fopenmp -lmpi_cxx -lmpi
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.39, N = 6 SE +/- 0.17, N = 4 SE +/- 0.54, N = 6 SE +/- 0.75, N = 15 SE +/- 0.56, N = 5 SE +/- 0.32, N = 5 SE +/- 0.55, N = 5 SE +/- 0.13, N = 5 73.81 47.07 72.01 74.57 58.14 63.99 54.98 63.10 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.21, N = 5 SE +/- 0.07, N = 3 SE +/- 0.16, N = 5 SE +/- 0.05, N = 5 SE +/- 0.41, N = 7 SE +/- 0.15, N = 4 SE +/- 0.43, N = 4 SE +/- 0.15, N = 4 54.95 35.56 55.40 55.54 44.40 45.32 45.09 44.05
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 800 1600 2400 3200 4000 SE +/- 9.71, N = 4 SE +/- 21.99, N = 4 SE +/- 24.09, N = 4 SE +/- 7.09, N = 4 SE +/- 24.34, N = 4 SE +/- 38.86, N = 5 SE +/- 28.30, N = 4 SE +/- 20.55, N = 4 3466 3424 3892 3801 3921 3707 3926 3785
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: 100 Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.08, N = 4 SE +/- 0.07, N = 4 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 30.97 31.35 27.66 28.10 26.99 27.09 26.84 26.25 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.05, N = 4 SE +/- 0.07, N = 3 SE +/- 0.12, N = 3 SE +/- 0.07, N = 3 SE +/- 0.17, N = 3 SE +/- 0.08, N = 3 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 16.18 17.26 18.29 19.23 18.14 19.65 18.47 20.18
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 170 340 510 680 850 SE +/- 0.48, N = 4 SE +/- 0.41, N = 4 SE +/- 1.15, N = 3 SE +/- 0.33, N = 3 SE +/- 0.67, N = 3 SE +/- 0.58, N = 3 SE +/- 0.67, N = 3 SE +/- 0.33, N = 3 692 691 767 765 763 762 763 763
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1200 2400 3600 4800 6000 SE +/- 10.11, N = 3 SE +/- 31.50, N = 3 SE +/- 1.76, N = 3 SE +/- 1.20, N = 3 SE +/- 4.06, N = 3 SE +/- 3.51, N = 3 SE +/- 2.03, N = 3 SE +/- 6.44, N = 3 5611 5560 5113 5106 5088 5102 5101 5108
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.3375 0.675 1.0125 1.35 1.6875 SE +/- 0.00, N = 4 SE +/- 0.00, N = 4 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 SE +/- 0.00, N = 3 1.50 1.50 1.30 1.31 1.39 1.39 1.39 1.40 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
Timed ImageMagick Compilation This test times how long it takes to build ImageMagick. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed ImageMagick Compilation 6.9.0 Time To Compile Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 4 8 12 16 20 SE +/- 0.03, N = 4 SE +/- 0.10, N = 4 SE +/- 0.13, N = 4 SE +/- 0.10, N = 4 SE +/- 0.07, N = 4 SE +/- 0.06, N = 4 SE +/- 0.11, N = 4 SE +/- 0.15, N = 4 11.49 12.60 13.26 14.72 14.09 15.38 14.40 16.08
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 7 14 21 28 35 SE +/- 0.07, N = 4 SE +/- 0.04, N = 4 SE +/- 0.08, N = 3 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 31.61 32.24 28.16 28.75 27.26 27.75 27.33 26.86 1. (CXX) g++ options: -fno-rtti -O3 -fPIE -pie -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20 40 60 80 100 SE +/- 0.55, N = 5 SE +/- 0.05, N = 4 SE +/- 0.19, N = 5 SE +/- 0.01, N = 4 SE +/- 0.03, N = 5 SE +/- 0.03, N = 4 SE +/- 0.05, N = 4 SE +/- 0.01, N = 3 97.94 54.45 86.98 51.30 83.34 42.98 72.33 37.01 1. (CXX) g++ options: -O3 -flto -pthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 100K 200K 300K 400K 500K SE +/- 726.00, N = 5 SE +/- 213.44, N = 5 SE +/- 398.40, N = 5 SE +/- 130.20, N = 4 SE +/- 411.20, N = 5 SE +/- 37.04, N = 4 SE +/- 441.66, N = 5 SE +/- 71.37, N = 3 455312.09 232012.77 261127.09 135125.15 257384.55 131606.86 246822.52 126171.52 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 SE +/- 0.028, N = 6 SE +/- 0.030, N = 4 SE +/- 0.021, N = 6 SE +/- 0.025, N = 4 SE +/- 0.028, N = 5 SE +/- 0.035, N = 3 SE +/- 0.021, N = 5 SE +/- 0.031, N = 3 7.402 14.257 7.881 15.211 9.285 18.102 10.840 21.030 1. (CXX) g++ options: -fopenmp -O2 -march=native
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 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 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 129.5 129.5 129.7 129.7 129.7 129.7 129.7 129.7 MIN: 105.7 MIN: 105.7 MIN: 105.9 MIN: 105.9 MIN: 105.9 MIN: 105.9 MIN: 105.9 MIN: 105.9 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -mavx512f -mavx512cd -mavx512bw -mavx512dq -fno-trapping-math -fno-math-errno -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.8507 1.7014 2.5521 3.4028 4.2535 SE +/- 0.004, N = 7 SE +/- 0.008, N = 7 SE +/- 0.002, N = 8 SE +/- 0.003, N = 7 SE +/- 0.074, N = 15 SE +/- 0.023, N = 15 SE +/- 0.064, N = 15 SE +/- 0.010, N = 7 2.659 3.391 2.364 3.267 2.961 3.704 3.110 3.781
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 20K 40K 60K 80K 100K SE +/- 271.54, N = 6 SE +/- 328.18, N = 8 SE +/- 178.23, N = 8 SE +/- 243.41, N = 12 SE +/- 560.41, N = 15 SE +/- 260.16, N = 15 SE +/- 545.92, N = 15 SE +/- 313.88, N = 7 98764.93 63760.45 58128.24 33811.63 66492.44 35234.06 65396.22 35116.55 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 15 30 45 60 75 SE +/- 0.36, N = 5 SE +/- 0.04, N = 4 SE +/- 0.45, N = 5 SE +/- 0.08, N = 4 SE +/- 0.29, N = 4 SE +/- 0.08, N = 4 SE +/- 0.33, N = 4 SE +/- 0.06, N = 4 65.38 38.02 67.64 50.03 52.23 41.34 51.91 40.23 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.26, N = 8 SE +/- 0.13, N = 6 SE +/- 0.13, N = 7 SE +/- 0.05, N = 5 SE +/- 1.20, N = 15 SE +/- 0.11, N = 5 SE +/- 0.63, N = 6 SE +/- 0.04, N = 5 199.61 112.66 152.75 91.99 129.75 83.49 125.75 75.47 MIN: 185.98 / MAX: 216.64 MIN: 107.88 / MAX: 117.19 MIN: 143.8 / MAX: 168.98 MIN: 88.57 / MAX: 98.54 MIN: 110.67 / MAX: 151.91 MIN: 80.58 / MAX: 88.97 MIN: 120.06 / MAX: 135.13 MIN: 72.93 / MAX: 78.77
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.019, N = 6 SE +/- 0.020, N = 6 SE +/- 0.019, N = 6 SE +/- 0.015, N = 6 SE +/- 0.082, N = 5 SE +/- 0.024, N = 6 SE +/- 0.096, N = 5 SE +/- 0.015, N = 6 6.856 6.458 8.036 7.532 12.578 7.748 12.617 7.622
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1318 2.2636 3.3954 4.5272 5.659 SE +/- 0.03, N = 6 SE +/- 0.00, N = 5 SE +/- 0.03, N = 6 SE +/- 0.00, N = 4 SE +/- 0.04, N = 5 SE +/- 0.00, N = 4 SE +/- 0.01, N = 5 SE +/- 0.00, N = 4 5.03 3.31 4.41 2.71 3.65 2.38 3.51 2.26
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.3345 2.669 4.0035 5.338 6.6725 SE +/- 0.013568, N = 6 SE +/- 0.022698, N = 7 SE +/- 0.019069, N = 15 SE +/- 0.022393, N = 7 SE +/- 0.008354, N = 7 SE +/- 0.023904, N = 7 SE +/- 0.010204, N = 7 SE +/- 0.025580, N = 6 2.159459 4.182707 2.529807 4.955865 2.527363 5.114161 2.881099 5.930957 1. (CXX) g++ options: -fopenmp -lmpi_cxx -lmpi
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.ldr_alb_nrm.3840x2160 - Device: CPU-Only Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1453 2.2906 3.4359 4.5812 5.7265 SE +/- 0.03, N = 6 SE +/- 0.00, N = 5 SE +/- 0.01, N = 6 SE +/- 0.00, N = 4 SE +/- 0.02, N = 5 SE +/- 0.00, N = 4 SE +/- 0.02, N = 5 SE +/- 0.00, N = 4 5.09 3.32 4.48 2.71 3.66 2.39 3.49 2.27
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 30 60 90 120 150 SE +/- 0.07, N = 7 SE +/- 0.05, N = 5 SE +/- 0.12, N = 6 SE +/- 0.07, N = 5 SE +/- 0.48, N = 6 SE +/- 0.02, N = 4 SE +/- 0.14, N = 6 SE +/- 0.05, N = 4 153.16 82.97 124.71 70.05 116.61 65.33 107.63 58.42 MIN: 143.78 / MAX: 166.08 MIN: 79.09 / MAX: 87.11 MIN: 115.3 / MAX: 143.82 MIN: 63.26 / MAX: 79.13 MIN: 107.47 / MAX: 129.72 MIN: 60.89 / MAX: 69.91 MIN: 102.22 / MAX: 114.8 MIN: 55.81 / MAX: 61.43
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.7808 1.5616 2.3424 3.1232 3.904 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 SE +/- 0.00, N = 6 3.46 3.47 3.07 3.08 3.09 3.10 3.09 3.10 1. (CC) gcc options: -fvisibility=hidden -O2 -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: IP Shapes 3D - Engine: CPU Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.973 1.946 2.919 3.892 4.865 SE +/- 0.004380, N = 5 SE +/- 0.001432, N = 5 SE +/- 0.005935, N = 5 SE +/- 0.003025, N = 5 SE +/- 0.005278, N = 5 SE +/- 0.010419, N = 5 SE +/- 0.004988, N = 5 SE +/- 0.003490, N = 5 0.738442 0.841608 0.714699 0.771506 4.175530 4.274810 4.283990 4.324650 MIN: 0.81 MIN: 4.06 MIN: 4.18 MIN: 4.18 MIN: 4.21 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1500 3000 4500 6000 7500 SE +/- 20.74, N = 7 SE +/- 30.70, N = 7 SE +/- 18.12, N = 6 SE +/- 14.74, N = 6 SE +/- 22.13, N = 6 SE +/- 9.53, N = 6 SE +/- 7.99, N = 6 SE +/- 7.18, N = 6 4693 4565 5672 5670 6777 6693 6821 6707 1. (CXX) g++ options: -O3
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.001169, N = 7 SE +/- 0.002339, N = 7 SE +/- 0.001522, N = 7 SE +/- 0.001388, N = 7 SE +/- 0.015670, N = 7 SE +/- 0.017741, N = 7 SE +/- 0.039215, N = 7 SE +/- 0.005800, N = 7 0.306795 0.585804 0.438154 0.650618 7.060430 8.751050 7.754310 9.194340 MIN: 6.83 MIN: 7.59 MIN: 6.74 MIN: 6.88 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 13 26 39 52 65 SE +/- 0.11, N = 8 SE +/- 0.08, N = 6 SE +/- 0.28, N = 7 SE +/- 0.09, N = 5 SE +/- 0.04, N = 7 SE +/- 0.09, N = 5 SE +/- 0.06, N = 7 SE +/- 0.04, N = 4 58.25 29.14 45.97 23.54 44.67 22.54 39.35 19.77 1. (CC) gcc options: -O3 -march=native -fopenmp
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 120 240 360 480 600 SE +/- 1.56, N = 7 SE +/- 1.17, N = 6 SE +/- 2.79, N = 6 SE +/- 0.16, N = 6 SE +/- 0.77, N = 6 SE +/- 0.08, N = 6 SE +/- 0.52, N = 6 SE +/- 0.03, N = 6 550.02 327.78 484.55 333.99 483.10 284.33 441.65 247.30 1. (CXX) g++ options: -O3 -flto -pthread
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, Lossless Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.031, N = 7 SE +/- 0.025, N = 7 SE +/- 0.022, N = 7 SE +/- 0.013, N = 7 SE +/- 0.035, N = 6 SE +/- 0.033, N = 6 SE +/- 0.027, N = 6 SE +/- 0.045, N = 6 5.613 5.784 6.172 6.278 7.038 6.605 7.121 6.836 1. (CXX) g++ options: -O3 -fPIC -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 12 - Input: Bosphorus 4K Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 40 80 120 160 200 SE +/- 0.42, N = 7 SE +/- 0.20, N = 7 SE +/- 0.25, N = 7 SE +/- 0.41, N = 7 SE +/- 0.97, N = 6 SE +/- 0.40, N = 7 SE +/- 0.83, N = 5 SE +/- 0.61, N = 7 165.50 177.86 156.52 164.25 127.55 151.72 122.09 146.10 1. (CXX) g++ options: -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1100 2200 3300 4400 5500 SE +/- 21.09, N = 7 SE +/- 22.92, N = 8 SE +/- 25.25, N = 7 SE +/- 23.47, N = 7 SE +/- 25.38, N = 6 SE +/- 15.47, N = 6 SE +/- 18.24, N = 6 SE +/- 21.44, N = 6 3894 3861 4733 4731 5322 5188 5342 5217 1. (CXX) g++ options: -O3
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 12 24 36 48 60 SE +/- 0.97, N = 15 SE +/- 0.29, N = 15 SE +/- 0.33, N = 8 SE +/- 0.36, N = 15 SE +/- 0.09, N = 8 SE +/- 0.04, N = 11 SE +/- 0.11, N = 9 SE +/- 0.05, N = 11 55.50 34.30 42.28 28.72 52.24 30.84 47.51 27.63 1. (CXX) g++ options: -O3 -lm -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 1.1963 2.3926 3.5889 4.7852 5.9815 SE +/- 0.015, N = 8 SE +/- 0.008, N = 8 SE +/- 0.014, N = 7 SE +/- 0.006, N = 7 SE +/- 0.028, N = 7 SE +/- 0.015, N = 7 SE +/- 0.019, N = 7 SE +/- 0.005, N = 7 4.579 4.456 5.074 4.923 5.311 4.867 5.317 4.916 1. (CXX) g++ options: -O3 -fPIC -lm
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 160 320 480 640 800 SE +/- 1.89, N = 8 SE +/- 1.34, N = 8 SE +/- 2.57, N = 8 SE +/- 1.17, N = 8 SE +/- 5.02, N = 8 SE +/- 1.27, N = 8 SE +/- 3.86, N = 8 SE +/- 1.64, N = 8 755.5 742.9 675.5 680.2 663.8 705.5 651.9 702.6 1. (CXX) g++ options: -O3 -march=native -mavx2 -mavx -msse4.1 -mfma -mavx512f -mavx512cd -mavx512bw -mavx512dq -fno-trapping-math -fno-math-errno -ldl
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 2 4 6 8 10 SE +/- 0.008, N = 11 SE +/- 0.007, N = 8 SE +/- 0.006, N = 10 SE +/- 0.019, N = 8 SE +/- 0.005, N = 10 SE +/- 0.010, N = 7 SE +/- 0.004, N = 9 SE +/- 0.006, N = 7 2.187 4.250 2.325 4.576 2.691 5.332 3.121 6.160 1. (CXX) g++ options: -O3
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.8496 1.6992 2.5488 3.3984 4.248 SE +/- 0.009, N = 10 SE +/- 0.016, N = 9 SE +/- 0.008, N = 9 SE +/- 0.022, N = 9 SE +/- 0.022, N = 9 SE +/- 0.011, N = 9 SE +/- 0.029, N = 8 SE +/- 0.017, N = 9 2.805 2.971 3.044 3.169 3.738 3.550 3.776 3.730 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: MG.C Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 50K 100K 150K 200K 250K SE +/- 1422.07, N = 7 SE +/- 406.29, N = 9 SE +/- 1262.04, N = 9 SE +/- 152.92, N = 9 SE +/- 1298.67, N = 9 SE +/- 382.59, N = 10 SE +/- 1415.96, N = 9 SE +/- 436.97, N = 10 220893.78 107321.89 163333.21 83448.70 171436.47 86407.46 172614.10 85910.95 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
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.5922 1.1844 1.7766 2.3688 2.961 SE +/- 0.024402, N = 15 SE +/- 0.023842, N = 15 SE +/- 0.017240, N = 15 SE +/- 0.007749, N = 10 SE +/- 0.028863, N = 15 SE +/- 0.020941, N = 15 SE +/- 0.030907, N = 15 SE +/- 0.023295, N = 15 1.432362 1.847537 1.469210 2.144857 1.711296 2.287766 1.982932 2.632077 1. (CXX) g++ options: -O2 -pthread -lbz2 -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 0.2844 0.5688 0.8532 1.1376 1.422 SE +/- 0.000902, N = 9 SE +/- 0.000392, N = 9 SE +/- 0.001274, N = 9 SE +/- 0.000805, N = 9 SE +/- 0.001240, N = 9 SE +/- 0.000927, N = 9 SE +/- 0.000806, N = 9 SE +/- 0.000508, N = 9 0.624250 0.907161 0.736170 1.077430 0.864358 1.196430 0.945075 1.264070 MIN: 0.58 MIN: 0.87 MIN: 0.66 MIN: 1.06 MIN: 0.81 MIN: 1.16 MIN: 0.9 MIN: 1.21 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread
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 Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 3 6 9 12 15 SE +/- 0.01, N = 10 SE +/- 0.00, N = 10 SE +/- 0.01, N = 10 SE +/- 0.01, N = 10 SE +/- 0.00, N = 10 SE +/- 0.00, N = 10 SE +/- 0.00, N = 10 SE +/- 0.00, N = 10 11.31 11.36 10.05 10.11 10.03 10.13 10.01 10.13 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
Timed CPython Compilation This test times how long it takes to build the reference Python implementation, CPython, with optimizations and LTO enabled for a release build. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed CPython Compilation 3.10.6 Build Configuration: Default Xeon Platinum 8592+ 2P Xeon Platinum 8592+ Xeon Platinum 8490H 2P Xeon Platinum 8490H Xeon Max 9480 2P Xeon Max 9480 Xeon Max 9468 2P Xeon Max 9468 5 10 15 20 25 16.93 17.08 18.93 19.20 18.82 18.85 18.81 19.12
Xeon Platinum 8490H Processor: Intel Xeon Platinum 8490H @ 3.50GHz (60 Cores / 120 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 512GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x2b000590Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 16 April 2024 00:42 by user phoronix.
Xeon Platinum 8490H 2P Processor: 2 x Intel Xeon Platinum 8490H @ 3.50GHz (120 Cores / 240 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 1008GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED, Network: 2 x Intel X710 for 10GBASE-T
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x2b000590Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 17 April 2024 09:38 by user phoronix.
Xeon Platinum 8592+ Processor: INTEL XEON PLATINUM 8592+ @ 3.90GHz (64 Cores / 128 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 512GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x21000200Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 11 April 2024 18:15 by user phoronix.
Xeon Platinum 8592+ 2P Processor: 2 x INTEL XEON PLATINUM 8592+ @ 3.90GHz (128 Cores / 256 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 1008GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED, Network: 2 x Intel X710 for 10GBASE-T
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x21000200Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 8 April 2024 10:49 by user phoronix.
Xeon Max 9468 Processor: Intel Xeon Max 9468 @ 3.50GHz (48 Cores / 96 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 576GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x2c000290Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 21 April 2024 09:58 by user phoronix.
Xeon Max 9468 2P Processor: 2 x Intel Xeon Max 9468 @ 3.50GHz (96 Cores / 192 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 1136GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED, Network: 2 x Intel X710 for 10GBASE-T
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x2c000290Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 22 April 2024 15:30 by user phoronix.
Xeon Max 9480 Processor: Intel Xeon Max 9480 @ 3.50GHz (56 Cores / 112 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 576GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x2c000290Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 26 April 2024 09:38 by user phoronix.
Xeon Max 9480 2P Processor: 2 x Intel Xeon Max 9480 @ 3.50GHz (112 Cores / 224 Threads), Motherboard: Quanta Cloud QuantaGrid D54Q-2U S6Q-MB-MPS (3B05.TEL4P1 BIOS), Chipset: Intel Device 1bce, Memory: 1136GB, Disk: 2 x 1920GB KIOXIA KCD8XPUG1T92, Graphics: ASPEED, Network: 2 x Intel X710 for 10GBASE-T
OS: Ubuntu 24.04, Kernel: 6.9.0-060900rc3-generic (x86_64), Compiler: GCC 13.2.0, File-System: ext4, Screen Resolution: 1920x1200
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-OiuXZC/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-OiuXZC/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: intel_pstate performance (EPP: performance) - CPU Microcode: 0x2c000290Java Notes: OpenJDK Runtime Environment (build 21.0.3-ea+7-Ubuntu-1build1)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 + reg_file_data_sampling: 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 27 April 2024 16:18 by user phoronix.