Threadripper 3970X AMD July 77

AMD Ryzen Threadripper 3970X 32-Core testing with a ASUS ROG ZENITH II EXTREME (0702 BIOS) and NVIDIA TITAN RTX 24GB on Ubuntu 20.04 via the Phoronix Test Suite.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 2007090-PTS-THREADRI49
Jump To Table - Results

Statistics

Remove Outliers Before Calculating Averages

Graph Settings

Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Result
Identifier
Performance Per
Dollar
Date
Run
  Test
  Duration
Threadripper 3970X
July 08 2020
  12 Hours, 1 Minute
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):

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: NVIDIA TITAN RTX 24GB (1350/7000MHz), Audio: NVIDIA TU102 HD Audio, Monitor: ASUS MG28U, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200

OS: Ubuntu 20.04, Kernel: 5.4.0-37-generic (x86_64), Desktop: GNOME Shell 3.36.2, Display Server: X Server 1.20.8, Display Driver: NVIDIA 450.36.06, OpenGL: 4.6.0, OpenCL: OpenCL 1.2 CUDA 11.0.185, Compiler: GCC 9.3.0 + CUDA 11.0, File-System: ext4, Screen Resolution: 3840x2160

Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --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-objc-gc=auto --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 -v
Processor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025
OpenCL Notes: GPU Compute Cores: 4608
Java Notes: OpenJDK Runtime Environment (build 11.0.7+10-post-Ubuntu-3ubuntu1)
Python Notes: Python 3.8.2
Security 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 STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected

Threadripper 3970X AMD July 77OpenBenchmarking.orgPhoronix Test SuiteAMD Ryzen Threadripper 3970X 32-Core @ 3.70GHz (32 Cores / 64 Threads)ASUS ROG ZENITH II EXTREME (0702 BIOS)AMD Starship/Matisse64GB1000GB Force MP600NVIDIA TITAN RTX 24GB (1350/7000MHz)NVIDIA TU102 HD AudioASUS MG28UAquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200Ubuntu 20.045.4.0-37-generic (x86_64)GNOME Shell 3.36.2X Server 1.20.8NVIDIA 450.36.064.6.0OpenCL 1.2 CUDA 11.0.185GCC 9.3.0 + CUDA 11.0ext43840x2160ProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLOpenCLCompilerFile-SystemScreen ResolutionThreadripper 3970X AMD July 77 BenchmarksSystem Logs- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --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-objc-gc=auto --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 -v - Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025- GPU Compute Cores: 4608- OpenJDK Runtime Environment (build 11.0.7+10-post-Ubuntu-3ubuntu1)- Python 3.8.2- 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 STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected

Threadripper 3970X AMD July 77lczero: Eigenlczero: BLASblender: Barbershop - NVIDIA OptiXblender: Pabellon Barcelona - OpenCLblender: Fishy Cat - OpenCLjava-gradle-perf: Reactorbasis: UASTC Level 2 + RDO Post-Processingblender: Barbershop - CUDArenaissance: Akka Unbalanced Cobwebbed Treeblender: Barbershop - OpenCLlczero: OpenCLlczero: Randblender: BMW27 - OpenCLblender: Pabellon Barcelona - CUDAwireguard: blender: Classroom - OpenCLblender: Barbershop - CPU-Onlygmic: 2D Function Plotting, 1000 Timesblender: BMW27 - NVIDIA OptiXsvt-av1: Enc Mode 0 - 1080pblender: Pabellon Barcelona - CPU-Onlystress-ng: Forkingstress-ng: MMAPblender: Classroom - CUDAblender: Classroom - CPU-Onlystress-ng: CPU Cacheblender: Pabellon Barcelona - NVIDIA OptiXpyperformance: python_startupblosc: blosclzdaphne: OpenMP - Points2Imagedav1d: Chimera 1080p 10-bitrodinia: OpenMP HotSpot3Ddaphne: OpenCL - Points2Imagegmic: 3D Elevated Function In Rand Colors, 100 Timesblender: Classroom - NVIDIA OptiXdaphne: NVIDIA CUDA - Points2Imagepyperformance: raytraceblender: Fishy Cat - CPU-Onlymontage: Mosaic of M17, K band, 1.5 deg x 1.5 degblender: Fishy Cat - CUDArodinia: OpenMP LavaMDluxcorerender: DLSCluxcorerender: Rainbow Colors and Prismonednn: IP Batch All - u8s8f32 - CPUonednn: IP Batch All - f32 - CPUbuild2: Time To Compilepyperformance: 2to3avifenc: 0blender: BMW27 - CPU-Onlypyperformance: gorodinia: OpenMP Leukocyterodinia: NVIDIA CUDA GPU Myocytehugin: Panorama Photo Assistant + Stitching Timeblender: Fishy Cat - NVIDIA OptiXbasis: ETC1Sblender: BMW27 - CUDArodinia: OpenCL Myocytecompress-7zip: Compress Speed Testbuild-linux-kernel: Time To Compilepyperformance: django_templateonednn: Recurrent Neural Network Training - f32 - CPUcompress-zstd: 19onednn: Recurrent Neural Network Inference - f32 - CPUdaphne: OpenCL - NDT Mappingpyperformance: regex_compileavifenc: 2aom-av1: Speed 6 Realtimeaom-av1: Speed 0 Two-Passpyperformance: crypto_pyaesstress-ng: CPU Stressstress-ng: NUMAstress-ng: System V Message Passingstress-ng: Glibc C String Functionsstress-ng: Memory Copyingstress-ng: Mallocstress-ng: Cryptostress-ng: MEMFDstress-ng: Glibc Qsort Data Sortingstress-ng: Socket Activitystress-ng: Matrix Mathstress-ng: Context Switchingstress-ng: Vector Mathstress-ng: Semaphoresstress-ng: SENDFILEstress-ng: Atomicdaphne: NVIDIA CUDA - NDT Mappingpyperformance: pathlibrodinia: OpenMP Myocytecompress-zstd: 3aom-av1: Speed 6 Two-Passpyperformance: pickle_pure_pythondaphne: OpenMP - NDT Mappingpyperformance: nbodypyperformance: json_loadsonednn: Deconvolution Batch deconv_1d - u8s8f32 - CPUonednn: Deconvolution Batch deconv_1d - f32 - CPUpyperformance: chaosbuild-apache: Time To Compiledaphne: NVIDIA CUDA - Euclidean Clusterpyperformance: floatbasis: UASTC Level 3gmic: Plotting Isosurface Of A 3D Volume, 1000 Timesaom-av1: Speed 4 Two-Passdaphne: OpenMP - Euclidean Clusteraom-av1: Speed 8 Realtimeonednn: IP Batch 1D - f32 - CPUonednn: IP Batch 1D - u8s8f32 - CPUrodinia: OpenMP Streamclusterdav1d: Chimera 1080pdaphne: OpenCL - Euclidean Clusterbasis: UASTC Level 2dacapobench: Tradebeansonednn: Matrix Multiply Batch Shapes Transformer - f32 - CPUonednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPUsvt-av1: Enc Mode 4 - 1080poctave-benchmark: dacapobench: Tradesoapdacapobench: H2dav1d: Summer Nature 4Krodinia: OpenMP CFD Solverdacapobench: Jythonrodinia: OpenCL Particle Filtersvt-av1: Enc Mode 8 - 1080pbasis: UASTC Level 0onednn: Convolution Batch Shapes Auto - u8s8f32 - CPUonednn: Convolution Batch Shapes Auto - f32 - CPUavifenc: 8dav1d: Summer Nature 1080pavifenc: 10onednn: Deconvolution Batch deconv_3d - u8s8f32 - CPUonednn: Deconvolution Batch deconv_3d - f32 - CPUsvt-vp9: Visual Quality Optimized - Bosphorus 1080psvt-vp9: VMAF Optimized - Bosphorus 1080psvt-vp9: PSNR/SSIM Optimized - Bosphorus 1080pThreadripper 3970X14231423915.80852.94779.55261.104585.823532.7110832.966394.688458171896299.43279.62214.334202.95194.08162.77628.900.128153.4581712.04370.54146.77118.8358.04109.951211831.924838.949385548145.1383.44112883.53099706382.69777.4925147.84531778443870.7673.33169.6666.1876.006.919.9774320.664655.50629152.65347.2323346.37844.03943.07538.9042.25538.1439.59619945328.84446.4194.438103.157.4845357.0416031.79619.240.3499.212774.70912.6525455291.892266108.409799.16706855819.899123.591335.57521.9719673.22146810.3919752856.50290374.426323046.94574963.77513384.23581.5016.327.2468489.04.114341079.9511822.51.676191.4423110320.6321261.1610519.55119.1902.671244.2739.941.155830.93083314.732807.432236.7612.49745470.3892740.8068368.5386.57431413789364.477.52039776.95870.5456.2076.090545.421444.589799.264.4531.553002.68142371.30457.84461.95OpenBenchmarking.org

LeelaChessZero

LeelaChessZero (lc0 / lczero) is a chess engine automated vian neural networks. This test profile can be used for OpenCL, CUDA + cuDNN, and BLAS (CPU-based) benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.25Backend: EigenThreadripper 3970X30060090012001500SE +/- 18.00, N = 914231. (CXX) g++ options: -pthread

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.25Backend: BLASThreadripper 3970X30060090012001500SE +/- 20.75, N = 914231. (CXX) g++ options: -pthread

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Barbershop - Compute: NVIDIA OptiXThreadripper 3970X2004006008001000SE +/- 1.51, N = 3915.80

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Pabellon Barcelona - Compute: OpenCLThreadripper 3970X2004006008001000SE +/- 4.28, N = 3852.94

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Fishy Cat - Compute: OpenCLThreadripper 3970X2004006008001000SE +/- 1.11, N = 3779.55

Java Gradle Build

This test runs Java software project builds using the Gradle build system. It is intended to give developers an idea as to the build performance for development activities and build servers. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterJava Gradle BuildGradle Build: ReactorThreadripper 3970X60120180240300SE +/- 3.76, N = 9261.10

Basis Universal

Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.12Settings: UASTC Level 2 + RDO Post-ProcessingThreadripper 3970X130260390520650SE +/- 0.02, N = 3585.821. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Barbershop - Compute: CUDAThreadripper 3970X120240360480600SE +/- 0.35, N = 3532.71

Renaissance

Renaissance is a suite of benchmarks designed to test the Java JVM from Apache Spark to a Twitter-like service to Scala and other features. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetterRenaissance 0.10.0Test: Akka Unbalanced Cobwebbed TreeThreadripper 3970X2K4K6K8K10KSE +/- 98.53, N = 510832.97

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Barbershop - Compute: OpenCLThreadripper 3970X90180270360450SE +/- 0.27, N = 3394.68

LeelaChessZero

LeelaChessZero (lc0 / lczero) is a chess engine automated vian neural networks. This test profile can be used for OpenCL, CUDA + cuDNN, and BLAS (CPU-based) benchmarking. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.25Backend: OpenCLThreadripper 3970X2K4K6K8K10KSE +/- 140.23, N = 384581. (CXX) g++ options: -pthread

OpenBenchmarking.orgNodes Per Second, More Is BetterLeelaChessZero 0.25Backend: RandomThreadripper 3970X40K80K120K160K200KSE +/- 448.09, N = 31718961. (CXX) g++ options: -pthread

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: BMW27 - Compute: OpenCLThreadripper 3970X70140210280350SE +/- 0.17, N = 3299.43

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Pabellon Barcelona - Compute: CUDAThreadripper 3970X60120180240300SE +/- 0.14, N = 3279.62

WireGuard + Linux Networking Stack Stress Test

This is a benchmark of the WireGuard secure VPN tunnel and Linux networking stack stress test. The test runs on the local host but does require root permissions to run. The way it works is it creates three namespaces. ns0 has a loopback device. ns1 and ns2 each have wireguard devices. Those two wireguard devices send traffic through the loopback device of ns0. The end result of this is that tests wind up testing encryption and decryption at the same time -- a pretty CPU and scheduler-heavy workflow. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterWireGuard + Linux Networking Stack Stress TestThreadripper 3970X50100150200250SE +/- 0.04, N = 3214.33

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Classroom - Compute: OpenCLThreadripper 3970X4080120160200SE +/- 1.04, N = 3202.95

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Barbershop - Compute: CPU-OnlyThreadripper 3970X4080120160200SE +/- 0.13, N = 3194.08

G'MIC

G'MIC is an open-source framework for image processing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterG'MICTest: 2D Function Plotting, 1000 TimesThreadripper 3970X4080120160200SE +/- 0.52, N = 3162.781. Version 2.4.5, Copyright (c) 2008-2019, David Tschumperle.

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: BMW27 - Compute: NVIDIA OptiXThreadripper 3970X714212835SE +/- 3.75, N = 1528.90

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.orgFrames Per Second, More Is BetterSVT-AV1 0.8Encoder Mode: Enc Mode 0 - Input: 1080pThreadripper 3970X0.02880.05760.08640.11520.144SE +/- 0.000, N = 30.1281. (CXX) g++ options: -O3 -fcommon -fPIE -fPIC -pie

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Pabellon Barcelona - Compute: CPU-OnlyThreadripper 3970X306090120150SE +/- 0.96, N = 3153.45

Stress-NG

Stress-NG is a Linux stress tool developed by Colin King of Canonical. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: ForkingThreadripper 3970X20K40K60K80K100KSE +/- 1288.31, N = 1581712.041. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: MMAPThreadripper 3970X80160240320400SE +/- 11.04, N = 15370.541. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Classroom - Compute: CUDAThreadripper 3970X306090120150SE +/- 0.95, N = 3146.77

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Classroom - Compute: CPU-OnlyThreadripper 3970X306090120150SE +/- 0.52, N = 3118.83

Stress-NG

Stress-NG is a Linux stress tool developed by Colin King of Canonical. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: CPU CacheThreadripper 3970X1326395265SE +/- 0.99, N = 1258.041. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Pabellon Barcelona - Compute: NVIDIA OptiXThreadripper 3970X20406080100SE +/- 0.35, N = 3109.95

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: python_startupThreadripper 3970X369121512

C-Blosc

A simple, compressed, fast and persistent data store library for C. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterC-Blosc 2.0 Beta 5Compressor: blosclzThreadripper 3970X3K6K9K12K15KSE +/- 107.49, N = 311831.91. (CXX) g++ options: -rdynamic

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: OpenMP - Kernel: Points2ImageThreadripper 3970X5K10K15K20K25KSE +/- 192.44, N = 324838.951. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

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.orgFPS, More Is Betterdav1d 0.7.0Video Input: Chimera 1080p 10-bitThreadripper 3970X306090120150SE +/- 0.24, N = 3145.13MIN: 99.72 / MAX: 242.081. (CC) gcc options: -pthread

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenMP HotSpot3DThreadripper 3970X20406080100SE +/- 0.49, N = 383.441. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: OpenCL - Kernel: Points2ImageThreadripper 3970X3K6K9K12K15KSE +/- 100.13, N = 312883.531. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

G'MIC

G'MIC is an open-source framework for image processing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterG'MICTest: 3D Elevated Function In Random Colors, 100 TimesThreadripper 3970X20406080100SE +/- 0.25, N = 382.701. Version 2.4.5, Copyright (c) 2008-2019, David Tschumperle.

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Classroom - Compute: NVIDIA OptiXThreadripper 3970X20406080100SE +/- 0.28, N = 377.49

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: NVIDIA CUDA - Kernel: Points2ImageThreadripper 3970X5K10K15K20K25KSE +/- 29.08, N = 325147.851. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: raytraceThreadripper 3970X90180270360450438

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Fishy Cat - Compute: CPU-OnlyThreadripper 3970X1632486480SE +/- 0.19, N = 370.76

Montage Astronomical Image Mosaic Engine

Montage is an open-source astronomical image mosaic engine. This BSD-licensed astronomy software is developed by the California Institute of Technology, Pasadena. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterMontage Astronomical Image Mosaic Engine 6.0Mosaic of M17, K band, 1.5 deg x 1.5 degThreadripper 3970X1632486480SE +/- 0.24, N = 373.331. (CC) gcc options: -std=gnu99 -lcfitsio -lm -O2

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Fishy Cat - Compute: CUDAThreadripper 3970X1632486480SE +/- 0.10, N = 369.66

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenMP LavaMDThreadripper 3970X1530456075SE +/- 0.10, N = 366.191. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

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.orgM samples/sec, More Is BetterLuxCoreRender 2.3Scene: DLSCThreadripper 3970X246810SE +/- 0.02, N = 36.00MIN: 5.76 / MAX: 6.37

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.3Scene: Rainbow Colors and PrismThreadripper 3970X246810SE +/- 0.01, N = 36.91MIN: 6.77 / MAX: 6.98

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: IP Batch All - Data Type: u8s8f32 - Engine: CPUThreadripper 3970X3691215SE +/- 0.01272, N = 39.97743MIN: 9.581. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: IP Batch All - Data Type: f32 - Engine: CPUThreadripper 3970X510152025SE +/- 0.07, N = 320.66MIN: 20.121. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Build2

This test profile measures the time to bootstrap/install the build2 C++ build toolchain from source. Build2 is a cross-platform build toolchain for C/C++ code and features Cargo-like features. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBuild2 0.12Time To CompileThreadripper 3970X1224364860SE +/- 0.13, N = 355.51

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: 2to3Threadripper 3970X60120180240300291

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.7.3Encoder Speed: 0Threadripper 3970X1224364860SE +/- 0.27, N = 352.651. (CXX) g++ options: -O3 -fPIC

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: BMW27 - Compute: CPU-OnlyThreadripper 3970X1122334455SE +/- 0.02, N = 347.23

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: goThreadripper 3970X50100150200250233

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenMP LeukocyteThreadripper 3970X1122334455SE +/- 0.28, N = 346.381. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: NVIDIA CUDA GPU MyocyteThreadripper 3970X1020304050SE +/- 0.03, N = 344.041. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

Hugin

Hugin is an open-source, cross-platform panorama photo stitcher software package. This test profile times how long it takes to run the assistant and panorama photo stitching on a set of images. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterHuginPanorama Photo Assistant + Stitching TimeThreadripper 3970X1020304050SE +/- 0.13, N = 343.08

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: Fishy Cat - Compute: NVIDIA OptiXThreadripper 3970X918273645SE +/- 0.04, N = 338.90

Basis Universal

Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.12Settings: ETC1SThreadripper 3970X1020304050SE +/- 0.09, N = 342.261. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

Blender

Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing via OpenCL or CUDA is supported. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 2.82Blend File: BMW27 - Compute: CUDAThreadripper 3970X918273645SE +/- 0.43, N = 338.14

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenCL MyocyteThreadripper 3970X918273645SE +/- 0.08, N = 339.601. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

7-Zip Compression

This is a test of 7-Zip using p7zip with its integrated benchmark feature or upstream 7-Zip for the Windows x64 build. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 16.02Compress Speed TestThreadripper 3970X40K80K120K160K200KSE +/- 29.79, N = 31994531. (CXX) g++ options: -pipe -lpthread

Timed Linux Kernel Compilation

This test times how long it takes to build the Linux kernel in a default configuration. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 5.4Time To CompileThreadripper 3970X714212835SE +/- 0.42, N = 428.84

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: django_templateThreadripper 3970X1122334455SE +/- 0.12, N = 346.4

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPUThreadripper 3970X4080120160200SE +/- 0.31, N = 3194.44MIN: 192.341. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Zstd Compression

This test measures the time needed to compress a sample file (an Ubuntu ISO) using Zstd compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.5Compression Level: 19Threadripper 3970X20406080100SE +/- 0.09, N = 3103.11. (CC) gcc options: -O3 -pthread -lz -llzma

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPUThreadripper 3970X1326395265SE +/- 0.39, N = 357.48MIN: 55.441. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: OpenCL - Kernel: NDT MappingThreadripper 3970X80160240320400SE +/- 1.03, N = 3357.041. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: regex_compileThreadripper 3970X4080120160200160

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.7.3Encoder Speed: 2Threadripper 3970X714212835SE +/- 0.13, N = 331.801. (CXX) g++ options: -O3 -fPIC

AOM AV1

This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 2.0Encoder Mode: Speed 6 RealtimeThreadripper 3970X510152025SE +/- 0.02, N = 319.241. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 2.0Encoder Mode: Speed 0 Two-PassThreadripper 3970X0.07650.1530.22950.3060.3825SE +/- 0.00, N = 30.341. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: crypto_pyaesThreadripper 3970X20406080100SE +/- 0.12, N = 399.2

Stress-NG

Stress-NG is a Linux stress tool developed by Colin King of Canonical. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: CPU StressThreadripper 3970X3K6K9K12K15KSE +/- 7.13, N = 312774.701. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: NUMAThreadripper 3970X2004006008001000SE +/- 5.04, N = 3912.651. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: System V Message PassingThreadripper 3970X5M10M15M20M25MSE +/- 98449.98, N = 325455291.891. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Glibc C String FunctionsThreadripper 3970X500K1000K1500K2000K2500KSE +/- 5825.25, N = 32266108.401. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Memory CopyingThreadripper 3970X2K4K6K8K10KSE +/- 36.24, N = 39799.161. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: MallocThreadripper 3970X150M300M450M600M750MSE +/- 580409.98, N = 3706855819.891. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: CryptoThreadripper 3970X2K4K6K8K10KSE +/- 16.71, N = 39123.591. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: MEMFDThreadripper 3970X30060090012001500SE +/- 1.44, N = 31335.571. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Glibc Qsort Data SortingThreadripper 3970X110220330440550SE +/- 1.81, N = 3521.971. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Socket ActivityThreadripper 3970X4K8K12K16K20KSE +/- 83.91, N = 319673.221. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Matrix MathThreadripper 3970X30K60K90K120K150KSE +/- 2167.56, N = 3146810.391. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Context SwitchingThreadripper 3970X4M8M12M16M20MSE +/- 97212.12, N = 319752856.501. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Vector MathThreadripper 3970X60K120K180K240K300KSE +/- 512.79, N = 3290374.421. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: SemaphoresThreadripper 3970X1.4M2.8M4.2M5.6M7MSE +/- 944.04, N = 36323046.941. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: SENDFILEThreadripper 3970X120K240K360K480K600KSE +/- 472.95, N = 3574963.771. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: AtomicThreadripper 3970X110K220K330K440K550KSE +/- 320.42, N = 3513384.231. (CC) gcc options: -O2 -std=gnu99 -lm -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: NVIDIA CUDA - Kernel: NDT MappingThreadripper 3970X130260390520650SE +/- 0.36, N = 3581.501. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: pathlibThreadripper 3970X48121620SE +/- 0.03, N = 316.3

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenMP MyocyteThreadripper 3970X612182430SE +/- 0.15, N = 327.251. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

Zstd Compression

This test measures the time needed to compress a sample file (an Ubuntu ISO) using Zstd compression. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.4.5Compression Level: 3Threadripper 3970X2K4K6K8K10KSE +/- 5.10, N = 38489.01. (CC) gcc options: -O3 -pthread -lz -llzma

AOM AV1

This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 2.0Encoder Mode: Speed 6 Two-PassThreadripper 3970X0.92481.84962.77443.69924.624SE +/- 0.00, N = 34.111. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: pickle_pure_pythonThreadripper 3970X90180270360450434

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: OpenMP - Kernel: NDT MappingThreadripper 3970X2004006008001000SE +/- 0.77, N = 31079.951. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: nbodyThreadripper 3970X306090120150118

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: json_loadsThreadripper 3970X510152025SE +/- 0.00, N = 322.5

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Deconvolution Batch deconv_1d - Data Type: u8s8f32 - Engine: CPUThreadripper 3970X0.37710.75421.13131.50841.8855SE +/- 0.00080, N = 31.67619MIN: 1.61. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Deconvolution Batch deconv_1d - Data Type: f32 - Engine: CPUThreadripper 3970X0.32450.6490.97351.2981.6225SE +/- 0.00284, N = 31.44231MIN: 1.381. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: chaosThreadripper 3970X20406080100SE +/- 0.33, N = 3103

Timed Apache Compilation

This test times how long it takes to build the Apache HTTPD web server. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Apache Compilation 2.4.41Time To CompileThreadripper 3970X510152025SE +/- 0.02, N = 320.63

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: NVIDIA CUDA - Kernel: Euclidean ClusterThreadripper 3970X30060090012001500SE +/- 12.76, N = 31261.161. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

PyPerformance

PyPerformance is the reference Python performance benchmark suite. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.0.0Benchmark: floatThreadripper 3970X20406080100105

Basis Universal

Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.12Settings: UASTC Level 3Threadripper 3970X510152025SE +/- 0.01, N = 319.551. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

G'MIC

G'MIC is an open-source framework for image processing. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterG'MICTest: Plotting Isosurface Of A 3D Volume, 1000 TimesThreadripper 3970X510152025SE +/- 0.03, N = 319.191. Version 2.4.5, Copyright (c) 2008-2019, David Tschumperle.

AOM AV1

This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 2.0Encoder Mode: Speed 4 Two-PassThreadripper 3970X0.60081.20161.80242.40323.004SE +/- 0.00, N = 32.671. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: OpenMP - Kernel: Euclidean ClusterThreadripper 3970X30060090012001500SE +/- 3.29, N = 31244.271. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

AOM AV1

This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterAOM AV1 2.0Encoder Mode: Speed 8 RealtimeThreadripper 3970X918273645SE +/- 0.05, N = 339.941. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: IP Batch 1D - Data Type: f32 - Engine: CPUThreadripper 3970X0.26010.52020.78031.04041.3005SE +/- 0.00226, N = 31.15583MIN: 1.111. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: IP Batch 1D - Data Type: u8s8f32 - Engine: CPUThreadripper 3970X0.20940.41880.62820.83761.047SE +/- 0.000851, N = 30.930833MIN: 0.891. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenMP StreamclusterThreadripper 3970X48121620SE +/- 0.04, N = 314.731. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

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.orgFPS, More Is Betterdav1d 0.7.0Video Input: Chimera 1080pThreadripper 3970X2004006008001000SE +/- 1.22, N = 3807.43MIN: 562.76 / MAX: 1032.591. (CC) gcc options: -pthread

Darmstadt Automotive Parallel Heterogeneous Suite

DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgTest Cases Per Minute, More Is BetterDarmstadt Automotive Parallel Heterogeneous SuiteBackend: OpenCL - Kernel: Euclidean ClusterThreadripper 3970X5001000150020002500SE +/- 5.66, N = 32236.761. (CXX) g++ options: -O3 -m64 -std=c++11 -lcudadevrt -lcudart_static -lrt -lpthread -ldl

Basis Universal

Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.12Settings: UASTC Level 2Threadripper 3970X3691215SE +/- 0.01, N = 312.501. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

DaCapo Benchmark

This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: TradebeansThreadripper 3970X10002000300040005000SE +/- 27.27, N = 44547

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPUThreadripper 3970X0.08760.17520.26280.35040.438SE +/- 0.000621, N = 30.389274MIN: 0.371. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPUThreadripper 3970X0.18150.3630.54450.7260.9075SE +/- 0.001072, N = 30.806836MIN: 0.761. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

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.orgFrames Per Second, More Is BetterSVT-AV1 0.8Encoder Mode: Enc Mode 4 - Input: 1080pThreadripper 3970X246810SE +/- 0.029, N = 38.5381. (CXX) g++ options: -O3 -fcommon -fPIE -fPIC -pie

GNU Octave Benchmark

This test profile measures how long it takes to complete several reference GNU Octave files via octave-benchmark. GNU Octave is used for numerical computations and is an open-source alternative to MATLAB. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterGNU Octave Benchmark 5.2.0Threadripper 3970X246810SE +/- 0.067, N = 56.574

DaCapo Benchmark

This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: TradesoapThreadripper 3970X7001400210028003500SE +/- 38.78, N = 43141

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: H2Threadripper 3970X8001600240032004000SE +/- 48.23, N = 43789

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.orgFPS, More Is Betterdav1d 0.7.0Video Input: Summer Nature 4KThreadripper 3970X80160240320400SE +/- 1.17, N = 3364.47MIN: 215.65 / MAX: 390.681. (CC) gcc options: -pthread

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenMP CFD SolverThreadripper 3970X246810SE +/- 0.043, N = 37.5201. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

DaCapo Benchmark

This test runs the DaCapo Benchmarks written in Java and intended to test system/CPU performance. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: JythonThreadripper 3970X9001800270036004500SE +/- 6.99, N = 43977

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 select OpenCL, NVIDIA CUDA and OpenMP test binaries at the moment. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 3.1Test: OpenCL Particle FilterThreadripper 3970X246810SE +/- 0.012, N = 36.9581. (CXX) g++ options: -m64 -lm -lcuda -lcudart -lcudadevrt -lcudart_static -lrt -lpthread -ldl

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.orgFrames Per Second, More Is BetterSVT-AV1 0.8Encoder Mode: Enc Mode 8 - Input: 1080pThreadripper 3970X1632486480SE +/- 0.32, N = 370.551. (CXX) g++ options: -O3 -fcommon -fPIE -fPIC -pie

Basis Universal

Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterBasis Universal 1.12Settings: UASTC Level 0Threadripper 3970X246810SE +/- 0.007, N = 36.2071. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPUThreadripper 3970X246810SE +/- 0.00282, N = 36.09054MIN: 5.971. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPUThreadripper 3970X1.21982.43963.65944.87926.099SE +/- 0.00411, N = 35.42144MIN: 5.31. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.7.3Encoder Speed: 8Threadripper 3970X1.03252.0653.09754.135.1625SE +/- 0.017, N = 34.5891. (CXX) g++ options: -O3 -fPIC

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.orgFPS, More Is Betterdav1d 0.7.0Video Input: Summer Nature 1080pThreadripper 3970X2004006008001000SE +/- 1.51, N = 3799.26MIN: 513.62 / MAX: 883.671. (CC) gcc options: -pthread

libavif avifenc

This is a test of the AOMedia libavif library testing the encoding of a JPEG image to AV1 Image Format (AVIF). Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.7.3Encoder Speed: 10Threadripper 3970X1.00192.00383.00574.00765.0095SE +/- 0.010, N = 34.4531. (CXX) g++ options: -O3 -fPIC

oneDNN

This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Deconvolution Batch deconv_3d - Data Type: u8s8f32 - Engine: CPUThreadripper 3970X0.34940.69881.04821.39761.747SE +/- 0.00242, N = 31.55300MIN: 1.481. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 1.5Harness: Deconvolution Batch deconv_3d - Data Type: f32 - Engine: CPUThreadripper 3970X0.60331.20661.80992.41323.0165SE +/- 0.01564, N = 32.68142MIN: 2.61. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl

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.orgFrames Per Second, More Is BetterSVT-VP9 0.1Tuning: Visual Quality Optimized - Input: Bosphorus 1080pThreadripper 3970X80160240320400SE +/- 4.11, N = 3371.301. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.1Tuning: VMAF Optimized - Input: Bosphorus 1080pThreadripper 3970X100200300400500SE +/- 3.72, N = 3457.841. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-VP9 0.1Tuning: PSNR/SSIM Optimized - Input: Bosphorus 1080pThreadripper 3970X100200300400500SE +/- 3.37, N = 3461.951. (CC) gcc options: -O3 -fcommon -fPIE -fPIC -fvisibility=hidden -pie -rdynamic -lpthread -lrt -lm

134 Results Shown

LeelaChessZero:
  Eigen
  BLAS
Blender:
  Barbershop - NVIDIA OptiX
  Pabellon Barcelona - OpenCL
  Fishy Cat - OpenCL
Java Gradle Build
Basis Universal
Blender
Renaissance
Blender
LeelaChessZero:
  OpenCL
  Rand
Blender:
  BMW27 - OpenCL
  Pabellon Barcelona - CUDA
WireGuard + Linux Networking Stack Stress Test
Blender:
  Classroom - OpenCL
  Barbershop - CPU-Only
G'MIC
Blender
SVT-AV1
Blender
Stress-NG:
  Forking
  MMAP
Blender:
  Classroom - CUDA
  Classroom - CPU-Only
Stress-NG
Blender
PyPerformance
C-Blosc
Darmstadt Automotive Parallel Heterogeneous Suite
dav1d
Rodinia
Darmstadt Automotive Parallel Heterogeneous Suite
G'MIC
Blender
Darmstadt Automotive Parallel Heterogeneous Suite
PyPerformance
Blender
Montage Astronomical Image Mosaic Engine
Blender
Rodinia
LuxCoreRender:
  DLSC
  Rainbow Colors and Prism
oneDNN:
  IP Batch All - u8s8f32 - CPU
  IP Batch All - f32 - CPU
Build2
PyPerformance
libavif avifenc
Blender
PyPerformance
Rodinia:
  OpenMP Leukocyte
  NVIDIA CUDA GPU Myocyte
Hugin
Blender
Basis Universal
Blender
Rodinia
7-Zip Compression
Timed Linux Kernel Compilation
PyPerformance
oneDNN
Zstd Compression
oneDNN
Darmstadt Automotive Parallel Heterogeneous Suite
PyPerformance
libavif avifenc
AOM AV1:
  Speed 6 Realtime
  Speed 0 Two-Pass
PyPerformance
Stress-NG:
  CPU Stress
  NUMA
  System V Message Passing
  Glibc C String Functions
  Memory Copying
  Malloc
  Crypto
  MEMFD
  Glibc Qsort Data Sorting
  Socket Activity
  Matrix Math
  Context Switching
  Vector Math
  Semaphores
  SENDFILE
  Atomic
Darmstadt Automotive Parallel Heterogeneous Suite
PyPerformance
Rodinia
Zstd Compression
AOM AV1
PyPerformance
Darmstadt Automotive Parallel Heterogeneous Suite
PyPerformance:
  nbody
  json_loads
oneDNN:
  Deconvolution Batch deconv_1d - u8s8f32 - CPU
  Deconvolution Batch deconv_1d - f32 - CPU
PyPerformance
Timed Apache Compilation
Darmstadt Automotive Parallel Heterogeneous Suite
PyPerformance
Basis Universal
G'MIC
AOM AV1
Darmstadt Automotive Parallel Heterogeneous Suite
AOM AV1
oneDNN:
  IP Batch 1D - f32 - CPU
  IP Batch 1D - u8s8f32 - CPU
Rodinia
dav1d
Darmstadt Automotive Parallel Heterogeneous Suite
Basis Universal
DaCapo Benchmark
oneDNN:
  Matrix Multiply Batch Shapes Transformer - f32 - CPU
  Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU
SVT-AV1
GNU Octave Benchmark
DaCapo Benchmark:
  Tradesoap
  H2
dav1d
Rodinia
DaCapo Benchmark
Rodinia
SVT-AV1
Basis Universal
oneDNN:
  Convolution Batch Shapes Auto - u8s8f32 - CPU
  Convolution Batch Shapes Auto - f32 - CPU
libavif avifenc
dav1d
libavif avifenc
oneDNN:
  Deconvolution Batch deconv_3d - u8s8f32 - CPU
  Deconvolution Batch deconv_3d - f32 - CPU
SVT-VP9:
  Visual Quality Optimized - Bosphorus 1080p
  VMAF Optimized - Bosphorus 1080p
  PSNR/SSIM Optimized - Bosphorus 1080p

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: NVIDIA TITAN RTX 24GB (1350/7000MHz), Audio: NVIDIA TU102 HD Audio, Monitor: ASUS MG28U, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200

OS: Ubuntu 20.04, Kernel: 5.4.0-37-generic (x86_64), Desktop: GNOME Shell 3.36.2, Display Server: X Server 1.20.8, Display Driver: NVIDIA 450.36.06, OpenGL: 4.6.0, OpenCL: OpenCL 1.2 CUDA 11.0.185, Compiler: GCC 9.3.0 + CUDA 11.0, File-System: ext4, Screen Resolution: 3840x2160

Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --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-objc-gc=auto --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 -v
Processor Notes: Scaling Governor: acpi-cpufreq ondemand - CPU Microcode: 0x8301025
OpenCL Notes: GPU Compute Cores: 4608
Java Notes: OpenJDK Runtime Environment (build 11.0.7+10-post-Ubuntu-3ubuntu1)
Python Notes: Python 3.8.2
Security 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 STIBP: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected

Testing initiated at 8 July 2020 21:34 by user pts.