AMD Ryzen Threadripper 3970X 32-Core testing with a ASUS ROG ZENITH II EXTREME (0702 BIOS) and llvmpipe 63GB on Ubuntu 19.10 via the Phoronix Test Suite.
Threadripper 3970X Processor: AMD Ryzen Threadripper 3970X 32-Core @ 3.70GHz (32 Cores / 64 Threads), Motherboard: ASUS ROG ZENITH II EXTREME (0702 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 1000GB Force MP600, Graphics: llvmpipe 63GB, Audio: AMD Navi 10 HDMI Audio, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200
OS: Ubuntu 19.10, Kernel: 5.3.0-18-generic (x86_64), Desktop: GNOME Shell 3.34.1, Display Server: X Server 1.20.5, Display Driver: modesetting 1.20.5, OpenGL: 3.3 Mesa 19.2.8 (LLVM 9.0 128 bits), Compiler: GCC 9.2.1 20191008 + Clang 9.0.0-2, File-System: ext4, Screen Resolution: 3840x2160
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-offload-targets=nvptx-none,hsa --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / errors=remount-ro,relatime,rwProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025Java Notes: OpenJDK Runtime Environment (build 11.0.6+10-post-Ubuntu-1ubuntu119.10.1)Python Notes: Python 2.7.17 + Python 3.7.5Security Notes: l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling
16 Cores Processor: AMD Ryzen Threadripper 3970X 32-Core @ 3.70GHz (16 Cores)Motherboard: ASUS ROG ZENITH II EXTREME (0702 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 1000GB Force MP600, Graphics: llvmpipe 63GB, Audio: AMD Navi 10 HDMI Audio, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200
OS: Ubuntu 19.10, Kernel: 5.3.0-40-generic (x86_64), Desktop: GNOME Shell 3.34.1, Display Server: X Server 1.20.5, Display Driver: modesetting 1.20.5, OpenGL: 3.3 Mesa 19.2.8 (LLVM 9.0 128 bits), Compiler: GCC 9.2.1 20191008 + Clang 9.0.0-2, File-System: ext4, Screen Resolution: 3840x2160
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-offload-targets=nvptx-none,hsa --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / errors=remount-ro,relatime,rwProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025Java Notes: OpenJDK Runtime Environment (build 11.0.6+10-post-Ubuntu-1ubuntu119.10.1)Python Notes: Python 2.7.17 + Python 3.7.5Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional RSB filling + tsx_async_abort: Not affected
OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: 3000 Fall 16 Cores Threadripper 3970X 0.7713 1.5426 2.3139 3.0852 3.8565 SE +/- 0.003898, N = 3 SE +/- 0.002468, N = 3 3.428156 3.394250 1. (CXX) g++ options: -O3 -rdynamic -lglut -lGL -lGLU
OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: 1000 Stack 16 Cores Threadripper 3970X 0.9297 1.8594 2.7891 3.7188 4.6485 SE +/- 0.002167, N = 3 SE +/- 0.005908, N = 3 4.132133 4.094888 1. (CXX) g++ options: -O3 -rdynamic -lglut -lGL -lGLU
OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: 1000 Convex 16 Cores Threadripper 3970X 0.8601 1.7202 2.5803 3.4404 4.3005 SE +/- 0.003706, N = 3 SE +/- 0.000553, N = 3 3.822767 3.787201 1. (CXX) g++ options: -O3 -rdynamic -lglut -lGL -lGLU
OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: 136 Ragdolls 16 Cores Threadripper 3970X 0.4939 0.9878 1.4817 1.9756 2.4695 SE +/- 0.001501, N = 3 SE +/- 0.001302, N = 3 2.195278 2.173580 1. (CXX) g++ options: -O3 -rdynamic -lglut -lGL -lGLU
OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: Convex Trimesh 16 Cores Threadripper 3970X 0.2163 0.4326 0.6489 0.8652 1.0815 SE +/- 0.000327, N = 3 SE +/- 0.000443, N = 3 0.961552 0.952042 1. (CXX) g++ options: -O3 -rdynamic -lglut -lGL -lGLU
Flexible IO Tester Fio is an advanced disk benchmark that depends upon the kernel's AIO access library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester Type: Sequential Write - IO Engine: POSIX AIO - Buffered: Yes - Direct: No - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 70K 140K 210K 280K 350K SE +/- 2185.81, N = 3 SE +/- 4041.45, N = 3 320333 307000
OpenBenchmarking.org MB/s, More Is Better Flexible IO Tester 3.12 Type: Sequential Read - IO Engine: POSIX AIO - Buffered: Yes - Direct: No - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 14.74, N = 3 SE +/- 7.64, N = 3 1424 1794
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester Type: Sequential Read - IO Engine: POSIX AIO - Buffered: Yes - Direct: No - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 100K 200K 300K 400K 500K SE +/- 3844.19, N = 3 SE +/- 1855.92, N = 3 364333 459333
TSCP This is a performance test of TSCP, Tom Kerrigan's Simple Chess Program, which has a built-in performance benchmark. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Nodes Per Second, More Is Better TSCP 1.81 AI Chess Performance 16 Cores Threadripper 3970X 300K 600K 900K 1200K 1500K SE +/- 12160.81, N = 5 SE +/- 3552.56, N = 5 1319372 1336877 1. (CC) gcc options: -O3 -march=native
SciMark This test runs the ANSI C version of SciMark 2.0, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This test is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Composite 16 Cores Threadripper 3970X 150 300 450 600 750 SE +/- 3.51, N = 3 SE +/- 7.72, N = 3 697.80 718.03 1. (CC) gcc options: -lm
OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Monte Carlo 16 Cores Threadripper 3970X 30 60 90 120 150 SE +/- 0.04, N = 3 SE +/- 1.57, N = 3 155.44 157.71 1. (CC) gcc options: -lm
OpenBenchmarking.org MB/s, More Is Better Flexible IO Tester 3.12 Type: Random Read - IO Engine: POSIX AIO - Buffered: Yes - Direct: No - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 13 26 39 52 65 SE +/- 0.15, N = 3 SE +/- 0.20, N = 3 45.2 56.2
SciMark This test runs the ANSI C version of SciMark 2.0, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This test is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Fast Fourier Transform 16 Cores Threadripper 3970X 80 160 240 320 400 SE +/- 0.60, N = 3 SE +/- 3.52, N = 3 328.58 356.71 1. (CC) gcc options: -lm
OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Sparse Matrix Multiply 16 Cores Threadripper 3970X 200 400 600 800 1000 SE +/- 0.56, N = 3 SE +/- 7.44, N = 3 782.85 793.80 1. (CC) gcc options: -lm
OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Dense LU Matrix Factorization 16 Cores Threadripper 3970X 200 400 600 800 1000 SE +/- 17.46, N = 3 SE +/- 13.80, N = 3 946.17 983.93 1. (CC) gcc options: -lm
OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Jacobi Successive Over-Relaxation 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 0.75, N = 3 SE +/- 12.34, N = 3 1275.97 1298.02 1. (CC) gcc options: -lm
OpenBenchmarking.org Nanoseconds/Operation, Fewer Is Better Go Benchmarks Test: json 16 Cores Threadripper 3970X 800K 1600K 2400K 3200K 4000K SE +/- 16615.41, N = 3 SE +/- 7331.16, N = 3 3523585 1564168
OpenBenchmarking.org Nanoseconds/Operation, Fewer Is Better Go Benchmarks Test: build 16 Cores Threadripper 3970X 4000M 8000M 12000M 16000M 20000M SE +/- 162720428.03, N = 3 SE +/- 170009930.60, N = 3 17244462905 20067677101
OpenBenchmarking.org Nanoseconds/Operation, Fewer Is Better Go Benchmarks Test: garbage 16 Cores Threadripper 3970X 200K 400K 600K 800K 1000K SE +/- 2753.72, N = 3 SE +/- 1243.04, N = 3 897936 589712
CLOMP CLOMP is the C version of the Livermore OpenMP benchmark developed to measure OpenMP overheads and other performance impacts due to threading in order to influence future system designs. This particular test profile configuration is currently set to look at the OpenMP static schedule speed-up across all available CPU cores using the recommended test configuration. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Speedup, More Is Better CLOMP 3.3 Static OMP Speedup 16 Cores Threadripper 3970X 0.684 1.368 2.052 2.736 3.42 SE +/- 0.05, N = 25 SE +/- 0.01, N = 25 3.04 0.71 1. (CC) gcc options: --openmp -O3 -lm
Compile Bench Compilebench tries to age a filesystem by simulating some of the disk IO common in creating, compiling, patching, stating and reading kernel trees. It indirectly measures how well filesystems can maintain directory locality as the disk fills up and directories age. This current test is setup to use the makej mode with 10 initial directories Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Compile Bench 0.6 Test: Initial Create 16 Cores Threadripper 3970X 150 300 450 600 750 SE +/- 3.20, N = 3 SE +/- 1.98, N = 3 548.75 716.22
OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy, Fixed Block Size - Array Size: 4096 MiB 16 Cores Threadripper 3970X 3K 6K 9K 12K 15K SE +/- 2.58, N = 3 SE +/- 6.90, N = 3 12748.61 12649.31 1. (CC) gcc options: -O3 -march=native
t-test1 This is a test of t-test1 for basic memory allocator benchmarks. Note this test profile is currently very basic and the overall time does include the warmup time of the custom t-test1 compilation. Improvements welcome. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better t-test1 2017-01-13 Threads: 1 16 Cores Threadripper 3970X 6 12 18 24 30 SE +/- 0.04, N = 3 SE +/- 0.11, N = 3 16.06 25.60 1. (CC) gcc options: -pthread
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Bsearch 16 Cores Threadripper 3970X 4K 8K 12K 16K 20K SE +/- 66.23, N = 3 SE +/- 115.68, N = 3 7239.75 20222.03 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Forking 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 11.00, N = 3 SE +/- 1093.43, N = 3 10970.25 97017.28 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Hsearch 16 Cores Threadripper 3970X 50K 100K 150K 200K 250K SE +/- 141.77, N = 3 SE +/- 134.40, N = 3 106584.80 253747.42 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Lsearch 16 Cores Threadripper 3970X 90 180 270 360 450 SE +/- 2.93, N = 15 SE +/- 3.61, N = 3 138.27 434.81 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Tsearch 16 Cores Threadripper 3970X 200 400 600 800 1000 SE +/- 2.00, N = 3 SE +/- 0.72, N = 3 383.18 1111.04 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: CPU Stress 16 Cores Threadripper 3970X 3K 6K 9K 12K 15K SE +/- 4.68, N = 3 SE +/- 35.07, N = 3 5019.09 12151.31 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Semaphores 16 Cores Threadripper 3970X 900K 1800K 2700K 3600K 4500K SE +/- 35197.46, N = 15 SE +/- 11364.25, N = 3 3998333.22 3379188.34 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Matrix Math 16 Cores Threadripper 3970X 50K 100K 150K 200K 250K SE +/- 114.46, N = 3 SE +/- 2518.16, N = 7 76991.04 234235.76 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Vector Math 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 10.48, N = 3 SE +/- 30.84, N = 3 35109.17 88021.72 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Memory Copying 16 Cores Threadripper 3970X 2K 4K 6K 8K 10K SE +/- 6.58, N = 3 SE +/- 46.71, N = 3 3963.81 8324.57 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Socket Activity 16 Cores Threadripper 3970X 5K 10K 15K 20K 25K SE +/- 30.56, N = 3 SE +/- 93.30, N = 3 10510.35 22217.76 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Context Switching 16 Cores Threadripper 3970X 5M 10M 15M 20M 25M SE +/- 9547.22, N = 3 SE +/- 434072.66, N = 15 2258922.93 21267958.31 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Glibc C String Functions 16 Cores Threadripper 3970X 700K 1400K 2100K 2800K 3500K SE +/- 21196.38, N = 3 SE +/- 19841.27, N = 3 1252570.26 3406391.46 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: Glibc Qsort Data Sorting 16 Cores Threadripper 3970X 110 220 330 440 550 SE +/- 2.46, N = 4 SE +/- 1.89, N = 3 181.55 489.81 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.07.26 Test: System V Message Passing 16 Cores Threadripper 3970X 6M 12M 18M 24M 30M SE +/- 23221.11, N = 3 SE +/- 271951.30, N = 15 16301159.73 26168429.69 1. (CC) gcc options: -O2 -std=gnu99 -lm -lz -lcrypt -lrt -lpthread -laio -lc
OpenBenchmarking.org usec, Fewer Is Better Sockperf 3.4 Test: Latency Ping Pong 16 Cores Threadripper 3970X 0.6595 1.319 1.9785 2.638 3.2975 SE +/- 0.009, N = 5 SE +/- 0.026, N = 25 2.210 2.931 1. (CXX) g++ options: --param -O3 -rdynamic -ldl -lpthread
OpenBenchmarking.org usec, Fewer Is Better Sockperf 3.4 Test: Latency Under Load 16 Cores Threadripper 3970X 3 6 9 12 15 SE +/- 0.41, N = 25 SE +/- 0.66, N = 25 10.78 13.23 1. (CXX) g++ options: --param -O3 -rdynamic -ldl -lpthread
OpenBenchmarking.org MB/s, More Is Better Stream 2013-01-17 Type: Scale 16 Cores Threadripper 3970X 11K 22K 33K 44K 55K SE +/- 11.54, N = 5 SE +/- 29.94, N = 5 49498.8 51667.1 1. (CC) gcc options: -O3 -march=native -fopenmp
OpenBenchmarking.org MB/s, More Is Better Stream 2013-01-17 Type: Triad 16 Cores Threadripper 3970X 12K 24K 36K 48K 60K SE +/- 13.13, N = 5 SE +/- 39.30, N = 5 54847.9 57321.8 1. (CC) gcc options: -O3 -march=native -fopenmp
OpenBenchmarking.org MB/s, More Is Better Stream 2013-01-17 Type: Add 16 Cores Threadripper 3970X 12K 24K 36K 48K 60K SE +/- 11.11, N = 5 SE +/- 47.58, N = 5 54720.5 56819.2 1. (CC) gcc options: -O3 -march=native -fopenmp
Minion Minion is an open-source constraint solver that is designed to be very scalable. This test profile uses Minion's integrated benchmarking problems to solve. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Graceful 16 Cores Threadripper 3970X 10 20 30 40 50 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 41.67 41.34 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic
OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Solitaire 16 Cores Threadripper 3970X 14 28 42 56 70 SE +/- 0.21, N = 3 SE +/- 0.14, N = 3 61.09 60.65 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic
OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Quasigroup 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.16, N = 3 SE +/- 0.37, N = 3 108.52 107.53 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: XZ 0 - Process: Decompression 16 Cores Threadripper 3970X 30 60 90 120 150 130 131 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Zstd 1 - Process: Compression 16 Cores Threadripper 3970X 120 240 360 480 600 SE +/- 1.53, N = 3 SE +/- 0.67, N = 3 543 543 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Zstd 1 - Process: Decompression 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 4.18, N = 3 1493 1503 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Zstd 8 - Process: Compression 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.67, N = 3 109 107 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Zstd 8 - Process: Decompression 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 4.18, N = 3 SE +/- 4.06, N = 3 1650 1656 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Crush 0 - Process: Compression 16 Cores Threadripper 3970X 30 60 90 120 150 SE +/- 0.58, N = 3 SE +/- 0.67, N = 3 120 121 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Crush 0 - Process: Decompression 16 Cores Threadripper 3970X 110 220 330 440 550 SE +/- 1.00, N = 3 SE +/- 1.15, N = 3 498 500 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Brotli 0 - Process: Compression 16 Cores Threadripper 3970X 120 240 360 480 600 SE +/- 0.33, N = 3 SE +/- 3.67, N = 3 539 547 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Brotli 0 - Process: Decompression 16 Cores Threadripper 3970X 140 280 420 560 700 SE +/- 2.33, N = 3 SE +/- 4.26, N = 3 628 633 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Brotli 2 - Process: Compression 16 Cores Threadripper 3970X 50 100 150 200 250 SE +/- 0.67, N = 3 215 216 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Brotli 2 - Process: Decompression 16 Cores Threadripper 3970X 160 320 480 640 800 SE +/- 0.88, N = 3 SE +/- 1.20, N = 3 736 739 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Libdeflate 1 - Process: Compression 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 1.45, N = 3 267 266 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MB/s, More Is Better lzbench 1.8 Test: Libdeflate 1 - Process: Decompression 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 0.67, N = 3 SE +/- 1.20, N = 3 1244 1245 1. (CXX) g++ options: -pthread -fomit-frame-pointer -fstrict-aliasing -ffast-math -O3
OpenBenchmarking.org MiB/s, More Is Better Botan 2.13.0 Test: AES-256 16 Cores Threadripper 3970X 1300 2600 3900 5200 6500 SE +/- 14.95, N = 3 SE +/- 7.95, N = 3 5858.03 5902.14 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
OpenBenchmarking.org MiB/s, More Is Better Botan 2.13.0 Test: Twofish 16 Cores Threadripper 3970X 80 160 240 320 400 SE +/- 0.02, N = 3 SE +/- 3.16, N = 3 385.68 391.43 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
OpenBenchmarking.org MiB/s, More Is Better Botan 2.13.0 Test: Blowfish 16 Cores Threadripper 3970X 100 200 300 400 500 SE +/- 0.01, N = 3 SE +/- 0.50, N = 3 471.10 475.10 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
OpenBenchmarking.org MiB/s, More Is Better Botan 2.13.0 Test: CAST-256 16 Cores Threadripper 3970X 30 60 90 120 150 SE +/- 0.27, N = 3 SE +/- 1.54, N = 3 154.99 154.15 1. (CXX) g++ options: -fstack-protector -m64 -pthread -lbotan-2 -ldl -lrt
Java SciMark This test runs the Java version of SciMark 2.0, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This benchmark is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.0 Computational Test: Composite 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 SE +/- 7.84, N = 4 SE +/- 1.94, N = 4 3005.07 3009.12
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.0 Computational Test: Monte Carlo 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 2.89, N = 4 SE +/- 2.62, N = 4 1679.21 1686.79
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.0 Computational Test: Fast Fourier Transform 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 14.00, N = 4 SE +/- 9.49, N = 4 2040.93 2043.32
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.0 Computational Test: Sparse Matrix Multiply 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 SE +/- 8.43, N = 4 SE +/- 4.74, N = 4 2716.60 2683.18
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.0 Computational Test: Dense LU Matrix Factorization 16 Cores Threadripper 3970X 1400 2800 4200 5600 7000 SE +/- 22.75, N = 4 SE +/- 2.88, N = 4 6637.13 6674.41
OpenBenchmarking.org Mflops, More Is Better Java SciMark 2.0 Computational Test: Jacobi Successive Over-Relaxation 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 1.78, N = 4 SE +/- 2.45, N = 4 1951.48 1957.92
CacheBench This is a performance test of CacheBench, which is part of LLCbench. CacheBench is designed to test the memory and cache bandwidth performance Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better CacheBench Test: Read 16 Cores Threadripper 3970X 700 1400 2100 2800 3500 SE +/- 0.65, N = 3 SE +/- 4.12, N = 3 3024.68 3035.03 MIN: 3022.65 / MAX: 3028.8 MIN: 3020.86 / MAX: 3044.34 1. (CC) gcc options: -lrt
OpenBenchmarking.org MB/s, More Is Better CacheBench Test: Write 16 Cores Threadripper 3970X 7K 14K 21K 28K 35K SE +/- 29.01, N = 3 SE +/- 14.14, N = 3 31631.79 31832.50 MIN: 27804.64 / MAX: 32902.03 MIN: 27271.39 / MAX: 33182.75 1. (CC) gcc options: -lrt
OpenBenchmarking.org MB/s, More Is Better CacheBench Test: Read / Modify / Write 16 Cores Threadripper 3970X 13K 26K 39K 52K 65K SE +/- 147.71, N = 3 SE +/- 167.50, N = 3 61009.39 61162.94 MIN: 55206.66 / MAX: 65416.9 MIN: 53743.85 / MAX: 65830.07 1. (CC) gcc options: -lrt
glibc bench The GNU C Library project provides the core libraries for the GNU system and GNU/Linux systems, as well as many other systems that use Linux as the kernel. These libraries provide critical APIs including ISO C11, POSIX.1-2008, BSD, OS-specific APIs and more.
Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: cos 16 Cores Threadripper 3970X 10 20 30 40 50 SE +/- 0.02, N = 3 SE +/- 0.27, N = 3 42.81 42.25
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: exp 16 Cores Threadripper 3970X 1.1235 2.247 3.3705 4.494 5.6175 SE +/- 0.00851, N = 3 SE +/- 0.01485, N = 3 4.95189 4.99317
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: ffs 16 Cores Threadripper 3970X 0.4022 0.8044 1.2066 1.6088 2.011 SE +/- 0.00045, N = 3 SE +/- 0.01753, N = 3 1.78754 1.75870
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: log2 16 Cores Threadripper 3970X 1.327 2.654 3.981 5.308 6.635 SE +/- 0.01183, N = 3 SE +/- 0.00811, N = 3 5.89799 5.88218
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: modf 16 Cores Threadripper 3970X 0.5072 1.0144 1.5216 2.0288 2.536 SE +/- 0.00154, N = 3 SE +/- 0.00267, N = 3 2.25415 2.23442
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: sqrt 16 Cores Threadripper 3970X 0.504 1.008 1.512 2.016 2.52 SE +/- 0.00107, N = 3 SE +/- 0.00274, N = 3 2.23979 2.22027
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: ffsll 16 Cores Threadripper 3970X 0.4001 0.8002 1.2003 1.6004 2.0005 SE +/- 0.00227, N = 3 SE +/- 0.01102, N = 3 1.77800 1.75897
OpenBenchmarking.org nanoseconds, Fewer Is Better glibc bench 1.0 Benchmark: pthread_once 16 Cores Threadripper 3970X 0.3996 0.7992 1.1988 1.5984 1.998 SE +/- 0.00019, N = 3 SE +/- 0.00270, N = 3 1.77594 1.76133
LAME MP3 Encoding LAME is an MP3 encoder licensed under the LGPL. This test measures the time required to encode a WAV file to MP3 format. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better LAME MP3 Encoding 3.100 WAV To MP3 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.007, N = 3 SE +/- 0.020, N = 3 7.182 7.188 1. (CC) gcc options: -O3 -ffast-math -funroll-loops -fschedule-insns2 -fbranch-count-reg -fforce-addr -pipe -lncurses -lm
Darktable Darktable is an open-source photography / workflow application this will use any system-installed Darktable program or on Windows will automatically download the pre-built binary from the project. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Darktable 2.6.0 Test: Boat - Acceleration: CPU-only 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.007, N = 3 SE +/- 0.097, N = 15 5.060 6.010
OpenBenchmarking.org Seconds, Fewer Is Better Darktable 2.6.0 Test: Masskrug - Acceleration: CPU-only 16 Cores Threadripper 3970X 0.8429 1.6858 2.5287 3.3716 4.2145 SE +/- 0.009, N = 3 SE +/- 0.017, N = 3 3.746 3.088
OpenBenchmarking.org Seconds, Fewer Is Better Darktable 2.6.0 Test: Server Rack - Acceleration: CPU-only 16 Cores Threadripper 3970X 0.027 0.054 0.081 0.108 0.135 SE +/- 0.000, N = 3 SE +/- 0.004, N = 12 0.097 0.120
OpenBenchmarking.org Seconds, Fewer Is Better Darktable 2.6.0 Test: Server Room - Acceleration: CPU-only 16 Cores Threadripper 3970X 0.4628 0.9256 1.3884 1.8512 2.314 SE +/- 0.010, N = 3 SE +/- 0.009, N = 3 2.057 1.991
BRL-CAD BRL-CAD 7.28.0 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 7.28.0 VGR Performance Metric 16 Cores Threadripper 3970X 110K 220K 330K 440K 550K 210066 507345 1. (CXX) g++ options: -std=c++98 -pipe -fno-strict-aliasing -fno-common -fexceptions -ftemplate-depth-128 -m64 -ggdb3 -O3 -fipa-pta -fstrength-reduce -finline-functions -flto -pedantic -rdynamic -lSM -lICE -lXi -lGLU -lGL -lX11 -lXext -lXrender -lpthread -ldl -luuid -lm
GEGL GEGL is the Generic Graphics Library and is the library/framework used by GIMP and other applications like GNOME Photos. This test profile times how long it takes to complete various GEGL operations on a static set of sample JPEG images. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Crop 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.043, N = 3 SE +/- 0.062, N = 3 7.491 8.003
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Scale 16 Cores Threadripper 3970X 1.2737 2.5474 3.8211 5.0948 6.3685 SE +/- 0.031, N = 3 SE +/- 0.040, N = 3 4.940 5.661
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Cartoon 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.08, N = 3 SE +/- 0.05, N = 3 88.68 90.95
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Reflect 16 Cores Threadripper 3970X 7 14 21 28 35 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 29.41 29.98
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Antialias 16 Cores Threadripper 3970X 9 18 27 36 45 SE +/- 0.01, N = 3 SE +/- 0.07, N = 3 36.36 39.00
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Tile Glass 16 Cores Threadripper 3970X 7 14 21 28 35 SE +/- 0.07, N = 3 SE +/- 0.02, N = 3 28.69 31.03
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Wavelet Blur 16 Cores Threadripper 3970X 14 28 42 56 70 SE +/- 0.06, N = 3 SE +/- 0.18, N = 3 60.11 62.40
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Color Enhance 16 Cores Threadripper 3970X 13 26 39 52 65 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 55.43 56.07
OpenBenchmarking.org Seconds, Fewer Is Better GEGL Operation: Rotate 90 Degrees 16 Cores Threadripper 3970X 9 18 27 36 45 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 37.69 38.17
librsvg RSVG/librsvg is an SVG vector graphics library. This test profile times how long it takes to complete various operations by rsvg-convert. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better librsvg Operation: SVG Files To PNG 16 Cores Threadripper 3970X 6 12 18 24 30 SE +/- 0.12, N = 3 SE +/- 0.12, N = 3 18.68 23.87 1. rsvg-convert version 2.44.14
LuaJIT This test profile is a collection of Lua scripts/benchmarks run against a locally-built copy of LuaJIT upstream. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better LuaJIT 2.1-git Test: Composite 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 1.55, N = 3 SE +/- 12.94, N = 3 1624.14 1659.51 1. (CC) gcc options: -lm -ldl -O2 -fomit-frame-pointer -U_FORTIFY_SOURCE -fno-stack-protector
OpenBenchmarking.org Mflops, More Is Better LuaJIT 2.1-git Test: Monte Carlo 16 Cores Threadripper 3970X 110 220 330 440 550 SE +/- 0.46, N = 3 SE +/- 4.85, N = 3 515.37 526.75 1. (CC) gcc options: -lm -ldl -O2 -fomit-frame-pointer -U_FORTIFY_SOURCE -fno-stack-protector
OpenBenchmarking.org Mflops, More Is Better LuaJIT 2.1-git Test: Fast Fourier Transform 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 1.67, N = 3 SE +/- 2.18, N = 3 278.18 287.97 1. (CC) gcc options: -lm -ldl -O2 -fomit-frame-pointer -U_FORTIFY_SOURCE -fno-stack-protector
OpenBenchmarking.org Mflops, More Is Better LuaJIT 2.1-git Test: Sparse Matrix Multiply 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 1.25, N = 3 SE +/- 11.74, N = 3 1319.00 1352.77 1. (CC) gcc options: -lm -ldl -O2 -fomit-frame-pointer -U_FORTIFY_SOURCE -fno-stack-protector
OpenBenchmarking.org Mflops, More Is Better LuaJIT 2.1-git Test: Dense LU Matrix Factorization 16 Cores Threadripper 3970X 900 1800 2700 3600 4500 SE +/- 7.05, N = 3 SE +/- 27.57, N = 3 3910.40 3996.69 1. (CC) gcc options: -lm -ldl -O2 -fomit-frame-pointer -U_FORTIFY_SOURCE -fno-stack-protector
OpenBenchmarking.org Mflops, More Is Better LuaJIT 2.1-git Test: Jacobi Successive Over-Relaxation 16 Cores Threadripper 3970X 500 1000 1500 2000 2500 SE +/- 0.94, N = 3 SE +/- 18.57, N = 3 2097.74 2133.37 1. (CC) gcc options: -lm -ldl -O2 -fomit-frame-pointer -U_FORTIFY_SOURCE -fno-stack-protector
Inkscape Inkscape is an open-source vector graphics editor. This test profile times how long it takes to complete various operations by Inkscape. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Inkscape Operation: SVG Files To PNG 16 Cores Threadripper 3970X 6 12 18 24 30 SE +/- 0.05, N = 3 SE +/- 0.09, N = 3 25.26 26.27 1. Inkscape 0.92.4 (5da689c313, 2019-01-14)
Dbench Dbench is a benchmark designed by the Samba project as a free alternative to netbench, but dbench contains only file-system calls for testing the disk performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Dbench 4.0 Client Count: 12 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 14.76, N = 8 SE +/- 13.11, N = 9 1466.07 595.26 1. (CC) gcc options: -lpopt -O2
Flexible IO Tester Fio is an advanced disk benchmark that depends upon the kernel's AIO access library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Flexible IO Tester 3.18 Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 10.95, N = 15 SE +/- 2.91, N = 3 1234 1151 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester 3.18 Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 70K 140K 210K 280K 350K SE +/- 2820.78, N = 15 SE +/- 881.92, N = 3 315733 294333 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
OpenBenchmarking.org MB/s, More Is Better Flexible IO Tester 3.18 Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 102.96, N = 12 1309 706 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
MKL-DNN DNNL This is a test of the Intel MKL-DNN (DNNL / Deep Neural Network Library) 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. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Deconvolution Batch deconv_1d - Data Type: f32 16 Cores Threadripper 3970X 0.7255 1.451 2.1765 2.902 3.6275 SE +/- 0.00642, N = 3 SE +/- 0.02372, N = 14 3.22435 2.78688 MIN: 3.13 MIN: 1.78 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
Flexible IO Tester Fio is an advanced disk benchmark that depends upon the kernel's AIO access library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester 3.18 Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 70K 140K 210K 280K 350K SE +/- 26330.47, N = 12 335000 180633 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
MKL-DNN DNNL This is a test of the Intel MKL-DNN (DNNL / Deep Neural Network Library) 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. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Recurrent Neural Network Training - Data Type: f32 16 Cores Threadripper 3970X 70 140 210 280 350 SE +/- 0.28, N = 3 SE +/- 2.80, N = 10 142.35 299.86 MIN: 140.83 MIN: 188.07 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
Flexible IO Tester Fio is an advanced disk benchmark that depends upon the kernel's AIO access library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Flexible IO Tester 3.18 Type: Sequential Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 27.78, N = 15 SE +/- 15.41, N = 15 1553 1346 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
MKL-DNN DNNL This is a test of the Intel MKL-DNN (DNNL / Deep Neural Network Library) 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. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Convolution Batch conv_googlenet_v3 - Data Type: f32 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.24, N = 3 SE +/- 0.89, N = 3 76.35 56.58 MIN: 75.02 MIN: 41.14 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Convolution Batch conv_alexnet - Data Type: f32 16 Cores Threadripper 3970X 40 80 120 160 200 SE +/- 0.48, N = 3 SE +/- 1.57, N = 6 178.67 137.79 MIN: 176.74 MIN: 97.04 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Deconvolution Batch deconv_all - Data Type: f32 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 0.54, N = 3 SE +/- 2.16, N = 3 1182.61 1713.90 MIN: 1163.99 MIN: 1362.26 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Convolution Batch conv_3d - Data Type: f32 16 Cores Threadripper 3970X 3 6 9 12 15 SE +/- 0.09666, N = 3 SE +/- 0.08221, N = 3 9.50372 6.82846 MIN: 9.16 MIN: 4.63 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Deconvolution Batch deconv_3d - Data Type: f32 16 Cores Threadripper 3970X 1.1053 2.2106 3.3159 4.4212 5.5265 SE +/- 0.02100, N = 3 SE +/- 0.26685, N = 12 3.69166 4.91239 MIN: 3.53 MIN: 2.56 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: Convolution Batch conv_all - Data Type: f32 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 2.68, N = 3 SE +/- 4.86, N = 3 1369.92 968.72 MIN: 1354.97 MIN: 791.36 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: IP Batch All - Data Type: f32 16 Cores Threadripper 3970X 4 8 12 16 20 SE +/- 0.04, N = 3 SE +/- 0.09, N = 3 11.57 17.33 MIN: 11.24 MIN: 11.67 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
OpenBenchmarking.org ms, Fewer Is Better MKL-DNN DNNL 1.1 Harness: IP Batch 1D - Data Type: f32 16 Cores Threadripper 3970X 0.7988 1.5976 2.3964 3.1952 3.994 SE +/- 0.02584, N = 3 SE +/- 0.05025, N = 15 3.55042 3.41655 MIN: 3.4 MIN: 2.08 1. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl
Flexible IO Tester Fio is an advanced disk benchmark that depends upon the kernel's AIO access library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester 3.18 Type: Sequential Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 90K 180K 270K 360K 450K SE +/- 7133.29, N = 15 SE +/- 3928.97, N = 15 397600 344533 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
OpenBenchmarking.org MB/s, More Is Better Flexible IO Tester 3.18 Type: Sequential Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 200 400 600 800 1000 SE +/- 12.94, N = 4 961 1052 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester 3.18 Type: Sequential Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory 16 Cores Threadripper 3970X 60K 120K 180K 240K 300K SE +/- 3424.79, N = 4 SE +/- 10149.45, N = 15 245750 269200 1. (CC) gcc options: -rdynamic -std=gnu99 -ffast-math -include -O3 -U_FORTIFY_SOURCE -march=native -ll -lcurl -lssl -lcrypto -lnuma -libverbs -lrt -laio -lz -lpthread -lm -ldl
OpenBenchmarking.org Files/s, More Is Better FS-Mark 3.3 Test: 5000 Files, 1MB Size, 4 Threads 16 Cores Threadripper 3970X 130 260 390 520 650 SE +/- 81.95, N = 12 SE +/- 28.74, N = 12 602.3 485.9 1. (CC) gcc options: -static
OpenBenchmarking.org Files/s, More Is Better FS-Mark 3.3 Test: 4000 Files, 32 Sub Dirs, 1MB Size 16 Cores Threadripper 3970X 140 280 420 560 700 SE +/- 4.34, N = 3 SE +/- 3.35, N = 3 633.2 560.7 1. (CC) gcc options: -static
OpenBenchmarking.org Files/s, More Is Better FS-Mark 3.3 Test: 1000 Files, 1MB Size, No Sync/FSync 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 SE +/- 6.33, N = 3 SE +/- 6.09, N = 3 2012.9 2724.8 1. (CC) gcc options: -static
OpenBenchmarking.org MB/s, More Is Better IOzone 3.465 Record Size: 4Kb - File Size: 512MB - Disk Test: Write Performance 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 4.10, N = 3 SE +/- 20.79, N = 3 1283.87 1594.31 1. (CC) gcc options: -O3
OpenBenchmarking.org MB/s, More Is Better IOzone 3.465 Record Size: 64Kb - File Size: 512MB - Disk Test: Write Performance 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 3.66, N = 3 SE +/- 6.54, N = 3 1361.61 1677.55 1. (CC) gcc options: -O3
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 2D FFT Size 4096 16 Cores Threadripper 3970X 1500 3000 4500 6000 7500 SE +/- 16.83, N = 3 SE +/- 2.20, N = 3 6841.8 6875.5 1. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm
OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 2D FFT Size 4096 16 Cores Threadripper 3970X 5K 10K 15K 20K 25K SE +/- 48.64, N = 3 SE +/- 102.34, N = 3 21534 21610 1. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm
Timed MrBayes Analysis This test performs a bayesian analysis of a set of primate genome sequences in order to estimate their phylogeny. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Timed MrBayes Analysis 3.2.7 Primate Phylogeny Analysis 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.16, N = 3 SE +/- 0.44, N = 3 69.52 75.41 1. (CC) gcc options: -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4a -msha -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -madx -mabm -O3 -std=c99 -pedantic -lm -lreadline
Numenta Anomaly Benchmark Numenta Anomaly Benchmark (NAB) is a benchmark for evaluating algorithms for anomaly detection in streaming, real-time applications. It is comprised of over 50 labeled real-world and artificial timeseries data files plus a novel scoring mechanism designed for real-time applications. This test profile currently measures the time to run various detectors. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Numenta Anomaly Benchmark 1.1 Detector: Relative Entropy 16 Cores Threadripper 3970X 3 6 9 12 15 SE +/- 0.10, N = 3 SE +/- 0.18, N = 3 11.59 12.42
HPC Challenge HPC Challenge (HPCC) is a cluster-focused benchmark consisting of the HPL Linpack TPP benchmark, DGEMM, STREAM, PTRANS, RandomAccess, FFT, and communication bandwidth and latency. This HPC Challenge test profile attempts to ship with standard yet versatile configuration/input files though they can be modified. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org GFLOPS, More Is Better HPC Challenge 1.5.0 Test / Class: G-HPL 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.19, N = 3 SE +/- 0.17, N = 3 89.99 73.20 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org GFLOPS, More Is Better HPC Challenge 1.5.0 Test / Class: G-Ffte 16 Cores Threadripper 3970X 4 8 12 16 20 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 11.17 16.12 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org GFLOPS, More Is Better HPC Challenge 1.5.0 Test / Class: EP-DGEMM 16 Cores Threadripper 3970X 11 22 33 44 55 SE +/- 0.43, N = 3 SE +/- 0.44, N = 3 49.58 29.80 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org GB/s, More Is Better HPC Challenge 1.5.0 Test / Class: G-Ptrans 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.00225, N = 3 SE +/- 0.01053, N = 3 2.12629 7.44119 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org GB/s, More Is Better HPC Challenge 1.5.0 Test / Class: EP-STREAM Triad 16 Cores Threadripper 3970X 0.7821 1.5642 2.3463 3.1284 3.9105 SE +/- 0.00112, N = 3 SE +/- 0.00672, N = 3 3.47614 1.90348 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org GUP/s, More Is Better HPC Challenge 1.5.0 Test / Class: G-Random Access 16 Cores Threadripper 3970X 0.0386 0.0772 0.1158 0.1544 0.193 SE +/- 0.00040, N = 3 SE +/- 0.00021, N = 3 0.11362 0.17138 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org usecs, Fewer Is Better HPC Challenge 1.5.0 Test / Class: Random Ring Latency 16 Cores Threadripper 3970X 0.1057 0.2114 0.3171 0.4228 0.5285 SE +/- 0.00064, N = 3 SE +/- 0.00770, N = 3 0.40788 0.46962 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org GB/s, More Is Better HPC Challenge 1.5.0 Test / Class: Random Ring Bandwidth 16 Cores Threadripper 3970X 0.8948 1.7896 2.6844 3.5792 4.474 SE +/- 0.00338, N = 3 SE +/- 0.02486, N = 3 3.97686 2.20814 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
OpenBenchmarking.org MB/s, More Is Better HPC Challenge 1.5.0 Test / Class: Max Ping Pong Bandwidth 16 Cores Threadripper 3970X 5K 10K 15K 20K 25K SE +/- 607.12, N = 3 SE +/- 271.12, N = 3 22271.54 21719.65 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 3.1.3
CloverLeaf CloverLeaf is a Lagrangian-Eulerian hydrodynamics benchmark. This test profile currently makes use of CloverLeaf's OpenMP version and benchmarked with the clover_bm8192.in input file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better CloverLeaf Lagrangian-Eulerian Hydrodynamics 16 Cores Threadripper 3970X 0.2835 0.567 0.8505 1.134 1.4175 SE +/- 0.00, N = 3 SE +/- 0.06, N = 12 1.26 1.05 1. (F9X) gfortran options: -O3 -march=native -funroll-loops -fopenmp
DeepSpeech Mozilla DeepSpeech is a speech-to-text engine powered by TensorFlow for machine learning and derived from Baidu's Deep Speech research paper. This test profile times the speech-to-text process for a roughly three minute audio recording. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better DeepSpeech 0.6 Acceleration: CPU 16 Cores Threadripper 3970X 13 26 39 52 65 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 49.10 56.68
OpenBenchmarking.org Seconds, Fewer Is Better Mlpack Benchmark Benchmark: scikit_linearridgeregression 16 Cores Threadripper 3970X 0.3308 0.6616 0.9924 1.3232 1.654 SE +/- 0.02, N = 12 SE +/- 0.02, N = 3 1.42 1.47
GROMACS The GROMACS (GROningen MAchine for Chemical Simulations) molecular dynamics package testing on the CPU with the water_GMX50 data. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2020 Water Benchmark 16 Cores Threadripper 3970X 0.7398 1.4796 2.2194 2.9592 3.699 SE +/- 0.002, N = 3 SE +/- 0.004, N = 3 2.048 3.288 1. (CXX) g++ options: -O3 -pthread -lrt -lpthread -lm
Parboil The Parboil Benchmarks from the IMPACT Research Group at University of Illinois are a set of throughput computing applications for looking at computing architecture and compilers. Parboil test-cases support OpenMP, OpenCL, and CUDA multi-processing environments. However, at this time the test profile is just making use of the OpenMP and OpenCL test workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Parboil 2.5 Test: OpenMP LBM 16 Cores Threadripper 3970X 12 24 36 48 60 SE +/- 0.08, N = 3 SE +/- 0.75, N = 4 36.35 52.97 1. (CXX) g++ options: -lm -lpthread -lgomp -O3 -ffast-math -fopenmp
OpenBenchmarking.org Seconds, Fewer Is Better Parboil 2.5 Test: OpenMP CUTCP 16 Cores Threadripper 3970X 0.4775 0.955 1.4325 1.91 2.3875 SE +/- 0.019941, N = 3 SE +/- 0.005586, N = 3 2.122203 0.993342 1. (CXX) g++ options: -lm -lpthread -lgomp -O3 -ffast-math -fopenmp
Parboil The Parboil Benchmarks from the IMPACT Research Group at University of Illinois are a set of throughput computing applications for looking at computing architecture and compilers. Parboil test-cases support OpenMP, OpenCL, and CUDA multi-processing environments. However, at this time the test profile is just making use of the OpenMP and OpenCL test workloads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Parboil 2.5 Test: OpenMP Stencil 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.124137, N = 3 SE +/- 0.009377, N = 3 7.782015 5.282141 1. (CXX) g++ options: -lm -lpthread -lgomp -O3 -ffast-math -fopenmp
OpenBenchmarking.org Seconds, Fewer Is Better Parboil 2.5 Test: OpenMP MRI Gridding 16 Cores Threadripper 3970X 13 26 39 52 65 SE +/- 0.64, N = 3 SE +/- 0.26, N = 3 37.09 60.18 1. (CXX) g++ options: -lm -lpthread -lgomp -O3 -ffast-math -fopenmp
NAS Parallel Benchmarks NPB, NAS Parallel Benchmarks, is a benchmark developed by NASA for high-end computer systems. This test profile currently uses the MPI version of NPB. This test profile offers selecting the different NPB tests/problems and varying problem sizes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: BT.C 16 Cores Threadripper 3970X 16K 32K 48K 64K 80K SE +/- 12.23, N = 3 SE +/- 105.14, N = 3 54825.13 74815.99 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: CG.C 16 Cores Threadripper 3970X 2K 4K 6K 8K 10K SE +/- 13.07, N = 3 SE +/- 35.93, N = 3 10888.19 10815.59 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: EP.C 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 SE +/- 2.47, N = 3 SE +/- 48.12, N = 3 1486.44 2831.86 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: EP.D 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 SE +/- 0.41, N = 3 SE +/- 2.52, N = 3 1480.10 2907.58 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: FT.C 16 Cores Threadripper 3970X 7K 14K 21K 28K 35K SE +/- 46.20, N = 3 SE +/- 17.98, N = 3 31324.20 34599.49 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: IS.D 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 1.71, N = 3 SE +/- 17.87, N = 3 1521.34 1464.45 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: LU.C 16 Cores Threadripper 3970X 14K 28K 42K 56K 70K SE +/- 21.03, N = 3 SE +/- 279.99, N = 3 54335.46 65854.97 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: MG.C 16 Cores Threadripper 3970X 6K 12K 18K 24K 30K SE +/- 43.00, N = 3 SE +/- 110.51, N = 3 28788.06 30218.91 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
OpenBenchmarking.org Total Mop/s, More Is Better NAS Parallel Benchmarks 3.4 Test / Class: SP.B 16 Cores Threadripper 3970X 7K 14K 21K 28K 35K SE +/- 37.51, N = 3 SE +/- 446.30, N = 3 24225.64 33710.70 1. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi 2. Open MPI 3.1.3
Rodinia Rodinia is a suite focused upon accelerating compute-intensive applications with accelerators. CUDA, OpenMP, and OpenCL parallel models are supported by the included applications. This profile utilizes the OpenCL and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Rodinia 2.4 Test: OpenMP LavaMD 16 Cores Threadripper 3970X 4 8 12 16 20 SE +/- 0.025, N = 3 SE +/- 0.047, N = 3 16.534 8.337 1. (CXX) g++ options: -O2 -lOpenCL
Rodinia Rodinia is a suite focused upon accelerating compute-intensive applications with accelerators. CUDA, OpenMP, and OpenCL parallel models are supported by the included applications. This profile utilizes the OpenCL and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Rodinia 2.4 Test: OpenMP CFD Solver 16 Cores Threadripper 3970X 3 6 9 12 15 SE +/- 0.011, N = 3 SE +/- 0.095, N = 3 12.077 9.784 1. (CXX) g++ options: -O2 -lOpenCL
OpenBenchmarking.org Seconds, Fewer Is Better Rodinia 2.4 Test: OpenMP Streamcluster 16 Cores Threadripper 3970X 5 10 15 20 25 SE +/- 0.08, N = 3 SE +/- 0.29, N = 3 11.27 20.40 1. (CXX) g++ options: -O2 -lOpenCL
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 days/ns, Fewer Is Better NAMD 2.13 ATPase Simulation - 327,506 Atoms 16 Cores Threadripper 3970X 0.2788 0.5576 0.8364 1.1152 1.394 SE +/- 0.00167, N = 3 SE +/- 0.00773, N = 15 1.23901 0.62357
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 2018-11-10 Test: tConvolve MT - Degridding 16 Cores Threadripper 3970X 800 1600 2400 3200 4000 SE +/- 2.81, N = 3 SE +/- 12.65, N = 3 3463.50 3738.11 1. (CXX) g++ options: -lpthread
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 2018-11-10 Test: tConvolve MPI - Gridding 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 SE +/- 1.16, N = 3 SE +/- 2.78, N = 3 2368.48 2682.40 1. (CXX) g++ options: -lpthread
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 2018-11-10 Test: tConvolve MPI - Degridding 16 Cores Threadripper 3970X 800 1600 2400 3200 4000 SE +/- 0.00, N = 3 SE +/- 6.40, N = 3 3544.17 3795.21 1. (CXX) g++ options: -lpthread
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 2018-11-10 Test: tConvolve OpenMP - Gridding 16 Cores Threadripper 3970X 1000 2000 3000 4000 5000 SE +/- 29.54, N = 11 SE +/- 42.14, N = 15 3330.88 4771.16 1. (CXX) g++ options: -lpthread
OpenBenchmarking.org Million Grid Points Per Second, More Is Better ASKAP 2018-11-10 Test: tConvolve OpenMP - Degridding 16 Cores Threadripper 3970X 800 1600 2400 3200 4000 SE +/- 7.16, N = 11 SE +/- 8.15, N = 15 3482.69 3764.47 1. (CXX) g++ options: -lpthread
Intel MPI Benchmarks Intel MPI Benchmarks for stressing MPI implementations. At this point the test profile aggregates results for some common MPI functionality. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Average Msg/sec, More Is Better Intel MPI Benchmarks 2019.3 Test: IMB-P2P PingPong 16 Cores Threadripper 3970X 4M 8M 12M 16M 20M SE +/- 15568.26, N = 3 SE +/- 175648.50, N = 3 9362495.74 17789641.38 MIN: 3568 / MAX: 21761153 MIN: 4183 / MAX: 42591960 1. (CXX) g++ options: -O0 -pedantic -fopenmp -pthread -lmpi_cxx -lmpi
OpenBenchmarking.org Average Mbytes/sec, More Is Better Intel MPI Benchmarks 2019.3 Test: IMB-MPI1 Exchange 16 Cores Threadripper 3970X 1500 3000 4500 6000 7500 SE +/- 11.57, N = 3 SE +/- 53.06, N = 3 7217.49 6258.07 MAX: 27690.75 MAX: 25730.7 1. (CXX) g++ options: -O0 -pedantic -fopenmp -pthread -lmpi_cxx -lmpi
OpenBenchmarking.org Average usec, More Is Better Intel MPI Benchmarks 2019.3 Test: IMB-MPI1 Exchange 16 Cores Threadripper 3970X 30 60 90 120 150 SE +/- 0.54, N = 3 SE +/- 3.02, N = 3 104.95 149.28 MIN: 0.58 / MAX: 3114.62 MIN: 0.65 / MAX: 5739.97 1. (CXX) g++ options: -O0 -pedantic -fopenmp -pthread -lmpi_cxx -lmpi
OpenBenchmarking.org Average Mbytes/sec, More Is Better Intel MPI Benchmarks 2019.3 Test: IMB-MPI1 PingPong 16 Cores Threadripper 3970X 1200 2400 3600 4800 6000 SE +/- 201.66, N = 15 SE +/- 139.42, N = 15 5828.12 5106.63 MIN: 5.29 / MAX: 21984.97 MIN: 5.04 / MAX: 18618.89 1. (CXX) g++ options: -O0 -pedantic -fopenmp -pthread -lmpi_cxx -lmpi
OpenBenchmarking.org Average Mbytes/sec, More Is Better Intel MPI Benchmarks 2019.3 Test: IMB-MPI1 Sendrecv 16 Cores Threadripper 3970X 1300 2600 3900 5200 6500 SE +/- 114.95, N = 15 SE +/- 64.08, N = 3 6263.73 4955.00 MAX: 41562.16 MAX: 30545.39 1. (CXX) g++ options: -O0 -pedantic -fopenmp -pthread -lmpi_cxx -lmpi
OpenBenchmarking.org Average usec, More Is Better Intel MPI Benchmarks 2019.3 Test: IMB-MPI1 Sendrecv 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.45, N = 15 SE +/- 0.48, N = 3 63.49 82.61 MIN: 0.19 / MAX: 1638.64 MIN: 0.38 / MAX: 2576.08 1. (CXX) g++ options: -O0 -pedantic -fopenmp -pthread -lmpi_cxx -lmpi
OpenBenchmarking.org ms, Fewer Is Better ArrayFire 3.7 Test: Conjugate Gradient CPU 16 Cores Threadripper 3970X 6 12 18 24 30 SE +/- 0.23, N = 3 SE +/- 0.25, N = 3 23.52 25.18 1. (CXX) g++ options: -rdynamic
OpenBenchmarking.org Hydro Cycle Time - Seconds, Fewer Is Better Pennant 1.0.1 Test: leblancbig 16 Cores Threadripper 3970X 2K 4K 6K 8K 10K SE +/- 1.62, N = 3 SE +/- 2.82, N = 3 4088.41 8623.65 1. (CXX) g++ options: -fopenmp -pthread -lmpi_cxx -lmpi
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 16 Cores Threadripper 3970X 6K 12K 18K 24K 30K SE +/- 24.90, N = 3 SE +/- 14.75, N = 3 26671.34 24786.28 1. (CC) gcc options: -lparcsr_ls -lparcsr_mv -lseq_mv -lIJ_mv -lkrylov -lHYPRE_utilities -lm -fopenmp -pthread -lmpi
FFTE FFTE is a package by Daisuke Takahashi to compute Discrete Fourier Transforms of 1-, 2- and 3- dimensional sequences of length (2^p)*(3^q)*(5^r). Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MFLOPS, More Is Better FFTE 6.0 Test: N=256, 1D Complex FFT Routine Threadripper 3970X 2K 4K 6K 8K 10K SE +/- 15.75, N = 3 9483.92 1. (F9X) gfortran options: -O3 -fomit-frame-pointer -fopenmp -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi
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. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Total Execution Time - Seconds, Fewer Is Better QMCPACK 3.8 16 Cores Threadripper 3970X 600 1200 1800 2400 3000 1039.4 2581.7 1. (CXX) g++ options: -fopenmp -fomit-frame-pointer -finline-limit=1000 -fstrict-aliasing -funroll-all-loops -march=native -O3 -ffast-math -lm
Zstd Compression This test measures the time needed to compress a sample file (an Ubuntu file-system image) using Zstd compression supplied by the system or otherwise externally of the test profile. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Zstd Compression Compressing ubuntu-16.04.3-server-i386.img, Compression Level 19 16 Cores Threadripper 3970X 40 80 120 160 200 SE +/- 0.45, N = 3 SE +/- 1.03, N = 3 163.05 170.78
CP2K Molecular Dynamics CP2K is an open-source molecular dynamics software package focused on quantum chemistry and solid-state physics. This test profile currently makes use of the OpenMP implementation and using the Fayalite-FIST molecular dynamics run and measures the total time to complete. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better CP2K Molecular Dynamics 6.1 Fayalite-FIST Data 16 Cores Threadripper 3970X 110 220 330 440 550 423.75 494.65
OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 128 16 Cores Threadripper 3970X 800K 1600K 2400K 3200K 4000K SE +/- 43879.72, N = 6 SE +/- 36545.27, N = 8 3701949 3558893
OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 128 16 Cores Threadripper 3970X 700K 1400K 2100K 2800K 3500K SE +/- 30085.47, N = 15 SE +/- 32378.59, N = 15 3353946 3388324
OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 128 16 Cores Threadripper 3970X 500K 1000K 1500K 2000K 2500K SE +/- 26516.90, N = 3 SE +/- 10877.55, N = 3 2555431 2416213
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Files 16 Cores Threadripper 3970X 4 8 12 16 20 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 13.88 11.45 1. (CC) gcc options: -lm
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Threads 16 Cores Threadripper 3970X 10 20 30 40 50 SE +/- 0.22, N = 3 SE +/- 0.05, N = 3 46.12 11.33 1. (CC) gcc options: -lm
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Launch Programs 16 Cores Threadripper 3970X 50 100 150 200 250 SE +/- 0.65, N = 3 SE +/- 0.36, N = 3 228.58 49.30 1. (CC) gcc options: -lm
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Processes 16 Cores Threadripper 3970X 12 24 36 48 60 SE +/- 0.76, N = 3 SE +/- 0.58, N = 12 54.98 27.96 1. (CC) gcc options: -lm
OpenBenchmarking.org Ns Per Event, Fewer Is Better OSBench Test: Memory Allocations 16 Cores Threadripper 3970X 16 32 48 64 80 SE +/- 0.22, N = 3 SE +/- 0.29, N = 3 67.16 70.32 1. (CC) gcc options: -lm
Schbench This is a benchmark of Schbench, a Linux kernel scheduler benchmark developed by Facebook. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org usec, 99.9th Latency Percentile, Fewer Is Better Schbench Message Threads: 8 - Workers Per Message Thread: 8 16 Cores Threadripper 3970X 10K 20K 30K 40K 50K SE +/- 740.24, N = 3 SE +/- 468.91, N = 12 46315 3587 1. (CC) gcc options: -O2 -lpthread
PostMark This is a test of NetApp's PostMark benchmark designed to simulate small-file testing similar to the tasks endured by web and mail servers. This test profile will set PostMark to perform 25,000 transactions with 500 files simultaneously with the file sizes ranging between 5 and 512 kilobytes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org TPS, More Is Better PostMark 1.51 Disk Transaction Performance 16 Cores Threadripper 3970X 2K 4K 6K 8K 10K SE +/- 53.33, N = 3 6357 7812 1. (CC) gcc options: -O3
Ethr Ethr is a cross-platform Golang-written network performance measurement tool developed by Microsoft that is capable of testing multiple protocols and different measurements. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbits/sec, More Is Better Ethr 2019-01-02 Server Address: localhost - Protocol: HTTP - Test: Bandwidth - Threads: 2 16 Cores Threadripper 3970X 700 1400 2100 2800 3500 SE +/- 6.36, N = 3 SE +/- 3.00, N = 3 3314.79 1844.58 MIN: 3240 / MAX: 3400 MIN: 1810 / MAX: 1910
iPerf iPerf is a network bandwidth throughput testing software. This test profile is used for automated testing of an iperf client and requires you have access to an iperf server. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 64 16 Cores Threadripper 3970X 12K 24K 36K 48K 60K SE +/- 1271.94, N = 12 SE +/- 423.19, N = 3 47216 54840 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 64 16 Cores Threadripper 3970X 15 30 45 60 75 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 67.1 67.1 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Semaphore Release And Acquire 16 Cores Threadripper 3970X 3 6 9 12 15 11 11 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
Rust Prime Benchmark Based on petehunt/rust-benchmark, this is a prime number benchmark that is multi-threaded and written in Rustlang. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Rust Prime Benchmark Prime Number Test To 200,000,000 16 Cores Threadripper 3970X 5 10 15 20 25 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 19.96 10.27 1. (CC) gcc options: -m64 -pie -nodefaultlibs -ldl -lrt -lpthread -lgcc_s -lc -lm -lutil
Stockfish This is a test of Stockfish, an advanced C++11 chess benchmark that can scale up to 128 CPU cores. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Nodes Per Second, More Is Better Stockfish 9 Total Time 16 Cores Threadripper 3970X 20M 40M 60M 80M 100M SE +/- 303886.69, N = 3 SE +/- 914534.22, N = 3 36932742 95728787 1. (CXX) g++ options: -m64 -lpthread -fno-exceptions -std=c++11 -pedantic -O3 -msse -msse3 -mpopcnt -flto
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 1.9.0-jumbo-1 Test: MD5 16 Cores Threadripper 3970X 500K 1000K 1500K 2000K 2500K SE +/- 3055.05, N = 3 SE +/- 33415.00, N = 4 1626000 2520750 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -pthread -lm -lz -ldl -lcrypt -lbz2
Rust Mandelbrot This test profile is of the combined time for the serial and parallel Mandelbrot sets written in Rustlang via willi-kappler/mandel-rust. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Rust Mandelbrot Time To Complete Serial/Parallel Mandelbrot 16 Cores Threadripper 3970X 9 18 27 36 45 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 40.17 32.09 1. (CC) gcc options: -m64 -pie -nodefaultlibs -lutil -ldl -lrt -lpthread -lgcc_s -lc -lm
OpenBenchmarking.org kH/s - Hash Speed, More Is Better Cpuminer-Opt 3.8.8.1 Algorithm: deep 16 Cores Threadripper 3970X 6K 12K 18K 24K 30K SE +/- 1488.63, N = 15 SE +/- 31.80, N = 3 6879.91 29323.00 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s - Hash Speed, More Is Better Cpuminer-Opt 3.8.8.1 Algorithm: lbry 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 330.05, N = 15 SE +/- 464.63, N = 3 41063 89853 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s - Hash Speed, More Is Better Cpuminer-Opt 3.8.8.1 Algorithm: skein 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 200.78, N = 3 SE +/- 194.28, N = 3 49073 102703 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s - Hash Speed, More Is Better Cpuminer-Opt 3.8.8.1 Algorithm: myr-gr 16 Cores Threadripper 3970X 3K 6K 9K 12K 15K SE +/- 1707.65, N = 15 SE +/- 71.26, N = 3 15369.09 13923.00 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org kH/s - Hash Speed, More Is Better Cpuminer-Opt 3.8.8.1 Algorithm: sha256t 16 Cores Threadripper 3970X 50K 100K 150K 200K 250K SE +/- 826.56, N = 3 SE +/- 26.03, N = 3 106950 223623 1. (CXX) g++ options: -O2 -lcurl -lz -lpthread -lssl -lcrypto -lgmp
OpenBenchmarking.org Events Per Second, More Is Better Sysbench 2018-07-28 Test: CPU 16 Cores Threadripper 3970X 15K 30K 45K 60K 75K SE +/- 26.51, N = 3 SE +/- 71.18, N = 3 33077.17 68765.38 1. (CC) gcc options: -pthread -O3 -funroll-loops -ggdb3 -march=amdfam10 -rdynamic -ldl -laio -lm
Tungsten Renderer Tungsten is a C++ physically based renderer that makes use of Intel's Embree ray tracing library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Tungsten Renderer 0.2.2 Scene: Hair 16 Cores Threadripper 3970X 5 10 15 20 25 SE +/- 0.01435, N = 3 SE +/- 0.03105, N = 3 19.25580 8.18217 1. (CXX) g++ options: -std=c++0x -march=znver1 -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4a -mfma -mbmi2 -mno-avx -mno-avx2 -mno-xop -mno-fma4 -mno-avx512f -mno-avx512vl -mno-avx512pf -mno-avx512er -mno-avx512cd -mno-avx512dq -mno-avx512bw -mno-avx512ifma -mno-avx512vbmi -fstrict-aliasing -O3 -rdynamic -lIlmImf -lIlmThread -lImath -lHalf -lIex -lz -ljpeg -lGL -lGLU -lpthread -ldl
OpenBenchmarking.org Seconds, Fewer Is Better Tungsten Renderer 0.2.2 Scene: Water Caustic 16 Cores Threadripper 3970X 6 12 18 24 30 SE +/- 0.06, N = 3 SE +/- 0.03, N = 3 24.13 18.68 1. (CXX) g++ options: -std=c++0x -march=znver1 -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4a -mfma -mbmi2 -mno-avx -mno-avx2 -mno-xop -mno-fma4 -mno-avx512f -mno-avx512vl -mno-avx512pf -mno-avx512er -mno-avx512cd -mno-avx512dq -mno-avx512bw -mno-avx512ifma -mno-avx512vbmi -fstrict-aliasing -O3 -rdynamic -lIlmImf -lIlmThread -lImath -lHalf -lIex -lz -ljpeg -lGL -lGLU -lpthread -ldl
OpenBenchmarking.org Seconds, Fewer Is Better Tungsten Renderer 0.2.2 Scene: Non-Exponential 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.01806, N = 3 SE +/- 0.00612, N = 3 8.92215 2.64289 1. (CXX) g++ options: -std=c++0x -march=znver1 -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4a -mfma -mbmi2 -mno-avx -mno-avx2 -mno-xop -mno-fma4 -mno-avx512f -mno-avx512vl -mno-avx512pf -mno-avx512er -mno-avx512cd -mno-avx512dq -mno-avx512bw -mno-avx512ifma -mno-avx512vbmi -fstrict-aliasing -O3 -rdynamic -lIlmImf -lIlmThread -lImath -lHalf -lIex -lz -ljpeg -lGL -lGLU -lpthread -ldl
OpenBenchmarking.org Seconds, Fewer Is Better Tungsten Renderer 0.2.2 Scene: Volumetric Caustic 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.00838, N = 3 SE +/- 0.00339, N = 3 8.56420 3.56382 1. (CXX) g++ options: -std=c++0x -march=znver1 -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -msse4a -mfma -mbmi2 -mno-avx -mno-avx2 -mno-xop -mno-fma4 -mno-avx512f -mno-avx512vl -mno-avx512pf -mno-avx512er -mno-avx512cd -mno-avx512dq -mno-avx512bw -mno-avx512ifma -mno-avx512vbmi -fstrict-aliasing -O3 -rdynamic -lIlmImf -lIlmThread -lImath -lHalf -lIex -lz -ljpeg -lGL -lGLU -lpthread -ldl
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample 6000x4000 pixel JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Swirl 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 1.86, N = 3 784 1590 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Rotate 16 Cores Threadripper 3970X 160 320 480 640 800 SE +/- 1.53, N = 3 SE +/- 3.46, N = 3 732 693 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Sharpen 16 Cores Threadripper 3970X 100 200 300 400 500 SE +/- 0.33, N = 3 SE +/- 3.38, N = 3 252 468 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Enhanced 16 Cores Threadripper 3970X 140 280 420 560 700 SE +/- 2.91, N = 3 357 669 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Resizing 16 Cores Threadripper 3970X 500 1000 1500 2000 2500 SE +/- 1.73, N = 3 SE +/- 8.01, N = 3 1847 2243 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Noise-Gaussian 16 Cores Threadripper 3970X 140 280 420 560 700 SE +/- 2.00, N = 3 351 631 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: HWB Color Space 16 Cores Threadripper 3970X 300 600 900 1200 1500 SE +/- 1.33, N = 3 SE +/- 1.53, N = 3 1417 1499 1. (CC) gcc options: -fopenmp -O2 -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lm -lpthread
SVT-VP9 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-VP9 CPU-based multi-threaded video encoder for the VP9 video format with a sample 1080p YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-VP9 0.1 Tuning: Visual Quality Optimized - Input: Bosphorus 1080p 16 Cores Threadripper 3970X 80 160 240 320 400 SE +/- 2.55, N = 3 SE +/- 4.19, N = 13 172.13 368.41 1. (CC) gcc options: -fPIE -fPIC -fvisibility=hidden -O3 -pie -rdynamic -lpthread -lrt -lm
x264 This is a simple test of the x264 encoder run on the CPU (OpenCL support disabled) with a sample video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better x264 2019-12-17 H.264 Video Encoding 16 Cores Threadripper 3970X 50 100 150 200 250 SE +/- 2.11, N = 3 SE +/- 3.52, N = 12 131.88 205.31 1. (CC) gcc options: -ldl -lavformat -lavcodec -lavutil -lswscale -m64 -lm -lpthread -O3 -ffast-math -std=gnu99 -fPIC -fomit-frame-pointer -fno-tree-vectorize
dav1d Dav1d is an open-source, speedy AV1 video decoder. This test profile times how long it takes to decode sample AV1 video content. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better dav1d 0.5.0 Video Input: Summer Nature 4K 16 Cores Threadripper 3970X 70 140 210 280 350 SE +/- 0.51, N = 3 SE +/- 0.26, N = 3 205.50 324.66 MIN: 190.74 / MAX: 228.56 MIN: 214.04 / MAX: 344.18 1. (CC) gcc options: -pthread
OpenBenchmarking.org FPS, More Is Better dav1d 0.5.0 Video Input: Summer Nature 1080p 16 Cores Threadripper 3970X 160 320 480 640 800 SE +/- 1.76, N = 3 SE +/- 3.50, N = 3 556.99 726.09 MIN: 471.9 / MAX: 604.99 MIN: 460.2 / MAX: 805.89 1. (CC) gcc options: -pthread
SVT-AV1 This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-AV1 CPU-based multi-threaded video encoder for the AV1 video format with a sample 1080p YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-AV1 0.8 Encoder Mode: Enc Mode 8 - Input: 1080p 16 Cores Threadripper 3970X 16 32 48 64 80 SE +/- 0.18, N = 3 SE +/- 0.37, N = 3 35.89 69.74 1. (CXX) g++ options: -fPIE -fPIC -pie
C-Ray This is a test of C-Ray, a simple raytracer designed to test the floating-point CPU performance. This test is multi-threaded (16 threads per core), will shoot 8 rays per pixel for anti-aliasing, and will generate a 1600 x 1200 image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 1.1 Total Time - 4K, 16 Rays Per Pixel 16 Cores Threadripper 3970X 8 16 24 32 40 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 34.16 16.38 1. (CC) gcc options: -lm -lpthread -O3
TTSIOD 3D Renderer A portable GPL 3D software renderer that supports OpenMP and Intel Threading Building Blocks with many different rendering modes. This version does not use OpenGL but is entirely CPU/software based. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org FPS, More Is Better TTSIOD 3D Renderer 2.3b Phong Rendering With Soft-Shadow Mapping 16 Cores Threadripper 3970X 140 280 420 560 700 SE +/- 0.91, N = 3 SE +/- 9.67, N = 13 617.85 667.34 1. (CXX) g++ options: -O3 -fomit-frame-pointer -ffast-math -mtune=native -flto -msse -mrecip -mfpmath=sse -msse2 -mssse3 -lSDL -fopenmp -fwhole-program -lstdc++
SVT-HEVC This is a test of the Intel Open Visual Cloud Scalable Video Technology SVT-HEVC CPU-based multi-threaded video encoder for the HEVC / H.265 video format with a sample 1080p YUV video file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Frames Per Second, More Is Better SVT-HEVC 1.4.1 1080p 8-bit YUV To HEVC Video Encode 16 Cores Threadripper 3970X 40 80 120 160 200 SE +/- 0.38, N = 3 SE +/- 0.72, N = 3 82.13 174.75 1. (CC) gcc options: -fPIE -fPIC -O3 -O2 -pie -rdynamic -lpthread -lrt
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.82 Blend File: Classroom - Compute: CPU-Only 16 Cores Threadripper 3970X 70 140 210 280 350 SE +/- 0.36, N = 3 SE +/- 0.31, N = 3 314.82 119.47
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.82 Blend File: Fishy Cat - Compute: CPU-Only 16 Cores Threadripper 3970X 40 80 120 160 200 SE +/- 0.09, N = 3 SE +/- 0.08, N = 3 178.46 71.28
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.82 Blend File: Barbershop - Compute: CPU-Only 16 Cores Threadripper 3970X 110 220 330 440 550 SE +/- 0.54, N = 3 SE +/- 0.21, N = 3 501.07 194.17
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.82 Blend File: Pabellon Barcelona - Compute: CPU-Only 16 Cores Threadripper 3970X 90 180 270 360 450 SE +/- 0.44, N = 3 SE +/- 0.59, N = 3 414.42 153.88
POV-Ray This is a test of POV-Ray, the Persistence of Vision Raytracer. POV-Ray is used to create 3D graphics using ray-tracing. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better POV-Ray 3.7.0.7 Trace Time 16 Cores Threadripper 3970X 7 14 21 28 35 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 32.25 13.24 1. (CXX) g++ options: -pipe -O3 -ffast-math -march=native -pthread -lSDL -lSM -lICE -lX11 -lIlmImf -lImath -lHalf -lIex -lIexMath -lIlmThread -lpthread -ltiff -ljpeg -lpng -lz -lrt -lm -lboost_thread -lboost_system
FFmpeg This test uses FFmpeg for testing the system's audio/video encoding performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better FFmpeg 4.0.2 H.264 HD To NTSC DV 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.039, N = 3 SE +/- 0.056, N = 3 5.041 7.331 1. (CC) gcc options: -lavdevice -lavfilter -lavformat -lavcodec -lswresample -lswscale -lavutil -lXv -lX11 -lXext -lm -lxcb -lxcb-shm -lxcb-shape -lxcb-xfixes -lasound -lSDL2 -lsndio -pthread -lbz2 -llzma -std=c11 -fomit-frame-pointer -fPIC -O3 -fno-math-errno -fno-signed-zeros -fno-tree-vectorize -MMD -MF -MT
OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.6.1 Binary: Pathtracer ISPC - Model: Crown 16 Cores Threadripper 3970X 8 16 24 32 40 SE +/- 0.02, N = 3 SE +/- 0.09, N = 3 12.83 35.55 MIN: 12.72 / MAX: 13.03 MIN: 32.78 / MAX: 37.31
Smallpt Smallpt is a C++ global illumination renderer written in less than 100 lines of code. Global illumination is done via unbiased Monte Carlo path tracing and there is multi-threading support via the OpenMP library. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Smallpt 1.0 Global Illumination Renderer; 128 Samples 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.008, N = 3 SE +/- 0.010, N = 3 7.407 2.894 1. (CXX) g++ options: -fopenmp -O3
LuxCoreRender LuxCoreRender is an open-source physically based renderer. This test profile is focused on running LuxCoreRender on the CPU as opposed to the OpenCL version. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.2 Scene: DLSC 16 Cores Threadripper 3970X 0.5648 1.1296 1.6944 2.2592 2.824 SE +/- 0.02, N = 3 SE +/- 0.19, N = 15 2.30 2.51 MIN: 2.18 / MAX: 2.41 MIN: 1.3 / MAX: 3
OpenBenchmarking.org M samples/sec, More Is Better LuxCoreRender 2.2 Scene: Rainbow Colors and Prism 16 Cores Threadripper 3970X 0.54 1.08 1.62 2.16 2.7 SE +/- 0.02, N = 3 SE +/- 0.17, N = 15 2.40 2.07 MIN: 2.33 / MAX: 2.49 MIN: 1.13 / MAX: 3.1
OSPray Intel OSPray is a portable ray-tracing engine for high-performance, high-fidenlity 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 FPS, More Is Better OSPray 1.8.5 Demo: San Miguel - Renderer: SciVis 16 Cores Threadripper 3970X 11 22 33 44 55 SE +/- 0.00, N = 12 18.52 50.00 MIN: 17.86 / MAX: 19.61 MIN: 43.48 / MAX: 52.63
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: XFrog Forest - Renderer: SciVis 16 Cores Threadripper 3970X 3 6 9 12 15 SE +/- 0.00, N = 4 SE +/- 0.02, N = 5 3.32 8.99 MIN: 3.27 / MAX: 3.38 MIN: 8.77 / MAX: 9.17
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: San Miguel - Renderer: Path Tracer 16 Cores Threadripper 3970X 0.8258 1.6516 2.4774 3.3032 4.129 SE +/- 0.00, N = 3 SE +/- 0.00, N = 4 1.37 3.67 MIN: 1.36 / MAX: 1.38 MIN: 3.65 / MAX: 3.72
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: NASA Streamlines - Renderer: SciVis 16 Cores Threadripper 3970X 15 30 45 60 75 SE +/- 0.00, N = 12 25.00 66.67 MIN: 23.81 / MAX: 25.64 MIN: 52.63 / MAX: 71.43
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: XFrog Forest - Renderer: Path Tracer 16 Cores Threadripper 3970X 1.0553 2.1106 3.1659 4.2212 5.2765 SE +/- 0.00, N = 3 SE +/- 0.00, N = 12 1.75 4.69 MIN: 1.73 / MAX: 1.77 MIN: 4.65 / MAX: 4.76
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: Magnetic Reconnection - Renderer: SciVis 16 Cores Threadripper 3970X 8 16 24 32 40 SE +/- 0.00, N = 12 SE +/- 0.36, N = 3 13.16 32.62 MIN: 12.5 / MAX: 13.51 MIN: 31.25 / MAX: 33.33
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: NASA Streamlines - Renderer: Path Tracer 16 Cores Threadripper 3970X 3 6 9 12 15 SE +/- 0.00, N = 5 SE +/- 0.06, N = 3 4.97 13.58 MIN: 4.9 / MAX: 5.05 MIN: 13.33 / MAX: 13.89
OpenBenchmarking.org FPS, More Is Better OSPray 1.8.5 Demo: Magnetic Reconnection - Renderer: Path Tracer 16 Cores Threadripper 3970X 100 200 300 400 500 SE +/- 21.00, N = 15 250.00 444.44 MIN: 200 / MAX: 333.33 MIN: 43.48 / MAX: 500
BlogBench BlogBench is designed to replicate the load of a real-world busy file server by stressing the file-system with multiple threads of random reads, writes, and rewrites. The behavior is mimicked of that of a blog by creating blogs with content and pictures, modifying blog posts, adding comments to these blogs, and then reading the content of the blogs. All of these blogs generated are created locally with fake content and pictures. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Final Score, More Is Better BlogBench 1.1 Test: Write 16 Cores Threadripper 3970X 11K 22K 33K 44K 55K SE +/- 102.18, N = 3 SE +/- 1160.31, N = 3 11700 49874 1. (CC) gcc options: -O2 -pthread
BlogBench BlogBench is designed to replicate the load of a real-world busy file server by stressing the file-system with multiple threads of random reads, writes, and rewrites. The behavior is mimicked of that of a blog by creating blogs with content and pictures, modifying blog posts, adding comments to these blogs, and then reading the content of the blogs. All of these blogs generated are created locally with fake content and pictures. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Final Score, More Is Better BlogBench 1.1 Test: Read 16 Cores Threadripper 3970X 300K 600K 900K 1200K 1500K SE +/- 3455.51, N = 3 SE +/- 12676.46, N = 8 1283483 1260198 1. (CC) gcc options: -O2 -pthread
OpenBenchmarking.org Requests Per Second, More Is Better Redis 5.0.5 Test: SADD 16 Cores Threadripper 3970X 400K 800K 1200K 1600K 2000K SE +/- 19315.80, N = 15 SE +/- 20821.87, N = 6 1874999.40 1851271.35
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.10 Method: Get 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 689.43, N = 15 SE +/- 422.66, N = 3 89370.8 77680.2 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.10 Method: Set 16 Cores Threadripper 3970X 16K 32K 48K 64K 80K SE +/- 2447.09, N = 12 SE +/- 556.85, N = 5 75424.0 45117.5 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.10 Method: Append 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 4614.05, N = 15 SE +/- 1452.18, N = 12 85868.9 49940.5 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.10 Method: Delete 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 702.49, N = 15 SE +/- 80.10, N = 3 88516.6 74147.4 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.10 Method: Prepend 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 5449.52, N = 12 SE +/- 36.76, N = 3 88961.4 46495.9 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.10 Method: Replace 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 5340.89, N = 12 SE +/- 677.13, N = 3 84273.8 47267.1 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.22 Method: Add - Connections: 64 16 Cores Threadripper 3970X 15K 30K 45K 60K 75K SE +/- 347.35, N = 3 SE +/- 803.43, N = 3 71799.0 48669.6 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.22 Method: Get - Connections: 64 16 Cores Threadripper 3970X 20K 40K 60K 80K 100K SE +/- 238.31, N = 3 SE +/- 810.66, N = 3 88109.7 78526.0 1. (CC) gcc options: -O2 -lm -rdynamic
OpenBenchmarking.org Operations Per Second, More Is Better Memcached mcperf 1.5.22 Method: Set - Connections: 64 16 Cores Threadripper 3970X 16K 32K 48K 64K 80K SE +/- 722.58, N = 3 SE +/- 524.58, N = 3 72538.5 48126.0 1. (CC) gcc options: -O2 -lm -rdynamic
NGINX Benchmark This is a test of ab, which is the Apache Benchmark program running against nginx. This test profile measures how many requests per second a given system can sustain when carrying out 2,000,000 requests with 500 requests being carried out concurrently. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better NGINX Benchmark 1.9.9 Static Web Page Serving 16 Cores Threadripper 3970X 8K 16K 24K 32K 40K SE +/- 50.77, N = 3 SE +/- 177.49, N = 3 36762.59 36707.66 1. (CC) gcc options: -lpthread -lcrypt -lcrypto -lz -O3 -march=native
Apache Benchmark This is a test of ab, which is the Apache benchmark program. This test profile measures how many requests per second a given system can sustain when carrying out 1,000,000 requests with 100 requests being carried out concurrently. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better Apache Benchmark 2.4.29 Static Web Page Serving 16 Cores Threadripper 3970X 9K 18K 27K 36K 45K SE +/- 50.68, N = 3 SE +/- 25.01, N = 3 41019.20 36286.78 1. (CC) gcc options: -shared -fPIC -O2 -pthread
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test measures the RSA 4096-bit performance of OpenSSL. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Signs Per Second, More Is Better OpenSSL 1.1.1 RSA 4096-bit Performance 16 Cores Threadripper 3970X 2K 4K 6K 8K 10K SE +/- 10.52, N = 3 SE +/- 32.86, N = 3 4550.9 9307.4 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
Blender Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing is supported. This system/blender test profile makes use of the system-supplied Blender. Use pts/blender if wishing to stick to a fixed version of Blender. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.80 Blend File: Pabellon Barcelona - Compute: CPU-Only 16 Cores Threadripper 3970X 80 160 240 320 400 SE +/- 0.17, N = 3 SE +/- 0.24, N = 3 382.22 146.02
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.80 Blend File: Fishy Cat - Compute: CPU-Only 16 Cores Threadripper 3970X 40 80 120 160 200 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 163.13 67.04
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.80 Blend File: Barbershop - Compute: CPU-Only 16 Cores Threadripper 3970X 90 180 270 360 450 SE +/- 0.65, N = 3 SE +/- 0.33, N = 3 430.81 159.93
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.80 Blend File: BMW27 - Compute: CPU-Only 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 0.11, N = 3 SE +/- 0.05, N = 3 97.69 40.40
OpenBenchmarking.org Seconds, Fewer Is Better Blender 2.80 Blend File: Classroom - Compute: CPU-Only 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 0.30, N = 3 SE +/- 0.18, N = 3 283.82 109.39
PHPBench PHPBench is a benchmark suite for PHP. It performs a large number of simple tests in order to bench various aspects of the PHP interpreter. PHPBench can be used to compare hardware, operating systems, PHP versions, PHP accelerators and caches, compiler options, etc. The number of iterations used is 1,000,000. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Score, More Is Better PHPBench 0.8.1 PHP Benchmark Suite 16 Cores Threadripper 3970X 140K 280K 420K 560K 700K SE +/- 266.61, N = 3 SE +/- 808.97, N = 3 641380 641756
OpenBenchmarking.org Microseconds - Average Latency, Fewer Is Better Apache HBase 2.2.3 Test: Random Write - Clients: 16 16 Cores Threadripper 3970X 20 40 60 80 100 SE +/- 29.55, N = 15 SE +/- 20.61, N = 15 81 72
Memtier_benchmark Memtier_benchmark is a NoSQL Redis/Memcache traffic generation plus benchmarking tool. This current test profile currently just stresses the Redis protocol and basic options exposed wotj a 1:1 Set/Get ratio, 30 pipeline, 100 clients per thread, and thread count equal to the number of CPU cores/threads present. Patches to extend the test are welcome as always. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Memtier_benchmark 1.2.17 Protocol: Redis 16 Cores Threadripper 3970X 500K 1000K 1500K 2000K 2500K SE +/- 8570.47, N = 3 SE +/- 34708.07, N = 4 2272891.42 2351667.58 1. (CXX) g++ options: -O2 -levent -lpthread -lz -lpcre
OpenBenchmarking.org Requests Per Second, More Is Better Redis 5.0.5 Test: SET 16 Cores Threadripper 3970X 400K 800K 1200K 1600K 2000K SE +/- 20540.77, N = 15 SE +/- 10340.63, N = 3 1723424.59 1674234.13 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
Facebook RocksDB This is a benchmark of 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 Facebook RocksDB 6.3.6 Test: Random Fill 16 Cores Threadripper 3970X 200K 400K 600K 800K 1000K SE +/- 6931.83, N = 3 SE +/- 8852.93, N = 3 606976 1083225 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Read 16 Cores Threadripper 3970X 40M 80M 120M 160M 200M SE +/- 155873.46, N = 3 SE +/- 1599226.86, N = 3 72774397 180626444 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Sequential Fill 16 Cores Threadripper 3970X 200K 400K 600K 800K 1000K SE +/- 8372.80, N = 15 SE +/- 3258.34, N = 3 702809 1147975 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill Sync 16 Cores Threadripper 3970X 6K 12K 18K 24K 30K SE +/- 17.38, N = 3 SE +/- 999.84, N = 15 9106 29167 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Read While Writing 16 Cores Threadripper 3970X 1.3M 2.6M 3.9M 5.2M 6.5M SE +/- 32658.12, N = 4 SE +/- 71829.99, N = 6 2381360 6204076 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
LevelDB LevelDB is a key-value storage library developed by Google that supports making use of Snappy for data compression and has other modern features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Hot Read 16 Cores Threadripper 3970X 10 20 30 40 50 SE +/- 0.106, N = 15 SE +/- 0.383, N = 11 8.974 44.286 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Fill Sync 16 Cores Threadripper 3970X 0.81 1.62 2.43 3.24 4.05 SE +/- 0.00, N = 3 1.0 3.6 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Fill Sync 16 Cores Threadripper 3970X 400 800 1200 1600 2000 SE +/- 4.63, N = 3 SE +/- 2.37, N = 3 1677.71 1938.17 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Overwrite 16 Cores Threadripper 3970X 9 18 27 36 45 SE +/- 0.10, N = 3 SE +/- 0.37, N = 3 40.3 26.8 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Overwrite 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 0.12, N = 3 SE +/- 3.84, N = 3 43.85 263.94 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Random Fill 16 Cores Threadripper 3970X 9 18 27 36 45 SE +/- 0.13, N = 3 SE +/- 0.03, N = 3 40.3 27.3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Fill 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 0.15, N = 3 SE +/- 0.33, N = 3 43.87 259.84 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Read 16 Cores Threadripper 3970X 10 20 30 40 50 SE +/- 0.093, N = 3 SE +/- 0.404, N = 10 9.313 44.353 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Seek Random 16 Cores Threadripper 3970X 15 30 45 60 75 SE +/- 0.02, N = 3 SE +/- 0.08, N = 3 14.89 68.59 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Delete 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 0.14, N = 3 SE +/- 0.71, N = 3 39.90 256.44 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Sequential Fill 16 Cores Threadripper 3970X 10 20 30 40 50 SE +/- 0.18, N = 3 SE +/- 0.32, N = 3 41.8 27.6 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Sequential Fill 16 Cores Threadripper 3970X 60 120 180 240 300 SE +/- 0.17, N = 3 SE +/- 2.86, N = 3 42.30 256.33 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 12.0 Scaling: Buffer Test - Test: Normal Load - Mode: Read Write 16 Cores Threadripper 3970X 3K 6K 9K 12K 15K SE +/- 1319.04, N = 15 SE +/- 333.19, N = 15 12881.18 11196.80 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm
MariaDB This is a MariaDB MySQL database server benchmark making use of mysqlslap. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 10.3.8 Clients: 64 16 Cores Threadripper 3970X 200 400 600 800 1000 SE +/- 1.35, N = 3 SE +/- 2.66, N = 3 1080 207 1. (CXX) g++ options: -pie -fPIC -fstack-protector -fno-rtti -O2 -lpthread -llzma -lbz2 -laio -lnuma -lz -lm -lpcre -lcrypt -lssl -lcrypto -ldl
OpenBenchmarking.org Seconds, Fewer Is Better Perl Benchmarks Test: Interpreter 16 Cores Threadripper 3970X 0.0003 0.0006 0.0009 0.0012 0.0015 SE +/- 0.00000115, N = 3 SE +/- 0.00000312, N = 3 0.00091186 0.00136623
Fhourstones This integer benchmark solves positions in the game of Connect-4, as played on a vertical 7x6 board. By default, it uses a 64Mb transposition table with the twobig replacement strategy. Positions are represented as 64-bit bitboards, and the hash function is computed using a single 64-bit modulo operation, giving 64-bit machines a slight edge. The alpha-beta searcher sorts moves dynamically based on the history heuristic. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Kpos / sec, More Is Better Fhourstones 3.1 Complex Connect-4 Solving 16 Cores Threadripper 3970X 3K 6K 9K 12K 15K SE +/- 53.43, N = 3 SE +/- 35.91, N = 3 15067.4 14702.3 1. (CC) gcc options: -O3
OpenBenchmarking.org Seconds, Fewer Is Better PolyBench-C 4.2 Test: Correlation Computation 16 Cores Threadripper 3970X 0.477 0.954 1.431 1.908 2.385 SE +/- 0.003, N = 3 SE +/- 0.002, N = 3 2.114 2.120 1. (CC) gcc options: -O3 -march=native
OpenBenchmarking.org Seconds, Fewer Is Better PolyBench-C 4.2 Test: 3 Matrix Multiplications 16 Cores Threadripper 3970X 0.4813 0.9626 1.4439 1.9252 2.4065 SE +/- 0.023, N = 3 SE +/- 0.012, N = 3 2.135 2.139 1. (CC) gcc options: -O3 -march=native
Bork File Encrypter Bork is a small, cross-platform file encryption utility. It is written in Java and designed to be included along with the files it encrypts for long-term storage. This test measures the amount of time it takes to encrypt a sample file. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Bork File Encrypter 1.4 File Encryption Time 16 Cores Threadripper 3970X 2 4 6 8 10 SE +/- 0.011, N = 3 SE +/- 0.005, N = 3 8.556 8.121
Git This test measures the time needed to carry out some sample Git operations on an example, static repository that happens to be a copy of the GNOME GTK tool-kit repository. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Git Time To Complete Common Git Commands 16 Cores Threadripper 3970X 1.1291 2.2582 3.3873 4.5164 5.6455 SE +/- 0.006, N = 3 SE +/- 0.014, N = 3 5.018 4.914 1. git version 2.20.1
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 16 Cores Threadripper 3970X 200 400 600 800 1000 SE +/- 2.31, N = 3 SE +/- 0.88, N = 3 955 958
SQLite This is a simple benchmark of SQLite. At present this test profile just measures the time to perform a pre-defined number of insertions on an indexed database. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better SQLite 3.30.1 Threads / Copies: 1 16 Cores Threadripper 3970X 4 8 12 16 20 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 13.92 14.41 1. (CC) gcc options: -O2 -lreadline -ltermcap -lz -lm -ldl -lpthread
Threadripper 3970X Processor: AMD Ryzen Threadripper 3970X 32-Core @ 3.70GHz (32 Cores / 64 Threads), Motherboard: ASUS ROG ZENITH II EXTREME (0702 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 1000GB Force MP600, Graphics: llvmpipe 63GB, Audio: AMD Navi 10 HDMI Audio, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200
OS: Ubuntu 19.10, Kernel: 5.3.0-18-generic (x86_64), Desktop: GNOME Shell 3.34.1, Display Server: X Server 1.20.5, Display Driver: modesetting 1.20.5, OpenGL: 3.3 Mesa 19.2.8 (LLVM 9.0 128 bits), Compiler: GCC 9.2.1 20191008 + Clang 9.0.0-2, File-System: ext4, Screen Resolution: 3840x2160
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-offload-targets=nvptx-none,hsa --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / errors=remount-ro,relatime,rwProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025Java Notes: OpenJDK Runtime Environment (build 11.0.6+10-post-Ubuntu-1ubuntu119.10.1)Python Notes: Python 2.7.17 + Python 3.7.5Security Notes: l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional STIBP: conditional RSB filling
Testing initiated at 6 March 2020 07:46 by user pts.
16 Cores Processor: AMD Ryzen Threadripper 3970X 32-Core @ 3.70GHz (16 Cores), Motherboard: ASUS ROG ZENITH II EXTREME (0702 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 1000GB Force MP600, Graphics: llvmpipe 63GB, Audio: AMD Navi 10 HDMI Audio, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200
OS: Ubuntu 19.10, Kernel: 5.3.0-40-generic (x86_64), Desktop: GNOME Shell 3.34.1, Display Server: X Server 1.20.5, Display Driver: modesetting 1.20.5, OpenGL: 3.3 Mesa 19.2.8 (LLVM 9.0 128 bits), Compiler: GCC 9.2.1 20191008 + Clang 9.0.0-2, File-System: ext4, Screen Resolution: 3840x2160
Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-offload-targets=nvptx-none,hsa --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vDisk Notes: NONE / errors=remount-ro,relatime,rwProcessor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025Java Notes: OpenJDK Runtime Environment (build 11.0.6+10-post-Ubuntu-1ubuntu119.10.1)Python Notes: Python 2.7.17 + Python 3.7.5Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional RSB filling + tsx_async_abort: Not affected
Testing initiated at 8 March 2020 18:50 by user pts.