TR 3970X Intel Open Source

AMD Ryzen Threadripper 3970X 32-Core testing with a ASUS ROG ZENITH II EXTREME (0702 BIOS) and AMD Radeon VII 16GB 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 2002127-PTS-TR3970XI48
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
February 11 2020
  3 Hours, 8 Minutes
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: AMD Radeon VII 16GB (1801/1000MHz), Audio: AMD Vega 20 HDMI Audio, Monitor: ASUS MG28U, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200

OS: Ubuntu 20.04, Kernel: 5.6.0-050600rc1-generic (x86_64) 20200209, Desktop: GNOME Shell 3.34.3, Display Server: X Server 1.20.7, Display Driver: amdgpu 19.1.0, OpenGL: 4.5 Mesa 19.2.4 (LLVM 9.0.0), Compiler: GCC 9.2.1 20200123, 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-link-mutex --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-build-config=bootstrap-lto-lean --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
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 + tsx_async_abort: Not affected

TR 3970X Intel Open SourceOpenBenchmarking.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 MP600AMD Radeon VII 16GB (1801/1000MHz)AMD Vega 20 HDMI AudioASUS MG28UAquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200Ubuntu 20.045.6.0-050600rc1-generic (x86_64) 20200209GNOME Shell 3.34.3X Server 1.20.7amdgpu 19.1.04.5 Mesa 19.2.4 (LLVM 9.0.0)GCC 9.2.1 20200123ext43840x2160ProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLCompilerFile-SystemScreen ResolutionTR 3970X Intel Open Source BenchmarksSystem Logs- --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-link-mutex --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-build-config=bootstrap-lto-lean --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- 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 + tsx_async_abort: Not affected

TR 3970X Intel Open Sourcefftw: Stock - 1D FFT Size 32fftw: Stock - 1D FFT Size 64fftw: Stock - 2D FFT Size 32fftw: Stock - 2D FFT Size 64fftw: Stock - 1D FFT Size 128fftw: Stock - 1D FFT Size 256fftw: Stock - 1D FFT Size 512fftw: Stock - 2D FFT Size 128fftw: Stock - 2D FFT Size 256fftw: Stock - 2D FFT Size 512fftw: Stock - 1D FFT Size 1024fftw: Stock - 1D FFT Size 2048fftw: Stock - 1D FFT Size 4096fftw: Stock - 2D FFT Size 1024fftw: Stock - 2D FFT Size 2048fftw: Stock - 2D FFT Size 4096fftw: Float + SSE - 1D FFT Size 32fftw: Float + SSE - 1D FFT Size 64fftw: Float + SSE - 2D FFT Size 32fftw: Float + SSE - 2D FFT Size 64fftw: Float + SSE - 1D FFT Size 128fftw: Float + SSE - 1D FFT Size 256fftw: Float + SSE - 1D FFT Size 512fftw: Float + SSE - 2D FFT Size 128fftw: Float + SSE - 2D FFT Size 256fftw: Float + SSE - 2D FFT Size 512fftw: Float + SSE - 1D FFT Size 1024fftw: Float + SSE - 1D FFT Size 2048fftw: Float + SSE - 1D FFT Size 4096fftw: Float + SSE - 2D FFT Size 1024fftw: Float + SSE - 2D FFT Size 2048fftw: Float + SSE - 2D FFT Size 4096mkl-dnn: IP Batch 1D - f32mkl-dnn: IP Batch All - f32mkl-dnn: IP Batch 1D - u8s8f32mkl-dnn: IP Batch All - u8s8f32mkl-dnn: Convolution Batch conv_3d - f32mkl-dnn: Convolution Batch conv_all - f32mkl-dnn: Convolution Batch conv_3d - u8s8f32mkl-dnn: Deconvolution Batch deconv_1d - f32mkl-dnn: Deconvolution Batch deconv_3d - f32mkl-dnn: Convolution Batch conv_alexnet - f32mkl-dnn: Convolution Batch conv_all - u8s8f32mkl-dnn: Deconvolution Batch deconv_all - f32mkl-dnn: Deconvolution Batch deconv_1d - u8s8f32mkl-dnn: Deconvolution Batch deconv_3d - u8s8f32mkl-dnn: Recurrent Neural Network Training - f32mkl-dnn: Convolution Batch conv_alexnet - u8s8f32mkl-dnn: Convolution Batch conv_googlenet_v3 - f32mkl-dnn: Convolution Batch conv_googlenet_v3 - u8s8f32embree: Pathtracer - Crownembree: Pathtracer ISPC - Crownembree: Pathtracer - Asian Dragonembree: Pathtracer - Asian Dragon Objembree: Pathtracer ISPC - Asian Dragonembree: Pathtracer ISPC - Asian Dragon Objoidn: Memorialluxcorerender: DLSCluxcorerender: Rainbow Colors and Prismtungsten: Hairtungsten: Water Caustictungsten: Non-Exponentialtungsten: Volumetric CausticThreadripper 3970X1127310428110609556.68550.29086.19137.09178.58740.18704.09318.38989.78800.98173.97184.26858.1154892078845811482122790037548500804835944156450475805460512572254557432027219932.0180911.369119.9828127.1444.78292776.2563724.971.832112.7027397.969312308.91270.161395.602327.45190.0371356.5542.3252696.08038.663336.865141.848337.478841.716836.004824.702.992.418.1503718.52282.664393.53144OpenBenchmarking.org

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.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 32Threadripper 3970X2K4K6K8K10K112731. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 64Threadripper 3970X2K4K6K8K10KSE +/- 2.73, N = 3104281. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 32Threadripper 3970X2K4K6K8K10KSE +/- 12.12, N = 3110601. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 64Threadripper 3970X2K4K6K8K10KSE +/- 130.25, N = 39556.61. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 128Threadripper 3970X2K4K6K8K10KSE +/- 9.65, N = 38550.21. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 256Threadripper 3970X2K4K6K8K10KSE +/- 24.48, N = 39086.11. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 512Threadripper 3970X2K4K6K8K10KSE +/- 18.81, N = 39137.01. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 128Threadripper 3970X2K4K6K8K10KSE +/- 11.46, N = 39178.51. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 256Threadripper 3970X2K4K6K8K10KSE +/- 76.80, N = 38740.11. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 512Threadripper 3970X2K4K6K8K10KSE +/- 89.15, N = 38704.01. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 1024Threadripper 3970X2K4K6K8K10KSE +/- 24.99, N = 39318.31. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 2048Threadripper 3970X2K4K6K8K10KSE +/- 5.17, N = 38989.71. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 1D FFT Size 4096Threadripper 3970X2K4K6K8K10KSE +/- 5.26, N = 38800.91. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 1024Threadripper 3970X2K4K6K8K10KSE +/- 12.92, N = 38173.91. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 2048Threadripper 3970X15003000450060007500SE +/- 37.20, N = 37184.21. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Stock - Size: 2D FFT Size 4096Threadripper 3970X15003000450060007500SE +/- 8.90, N = 36858.11. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 32Threadripper 3970X3K6K9K12K15KSE +/- 31.01, N = 3154891. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 64Threadripper 3970X4K8K12K16K20KSE +/- 75.35, N = 3207881. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 32Threadripper 3970X10K20K30K40K50KSE +/- 112.90, N = 3458111. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 64Threadripper 3970X10K20K30K40K50KSE +/- 200.21, N = 3482121. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 128Threadripper 3970X6K12K18K24K30KSE +/- 226.32, N = 3279001. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 256Threadripper 3970X8K16K24K32K40KSE +/- 528.05, N = 15375481. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 512Threadripper 3970X11K22K33K44K55KSE +/- 78.30, N = 3500801. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 128Threadripper 3970X10K20K30K40K50KSE +/- 437.77, N = 3483591. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 256Threadripper 3970X9K18K27K36K45KSE +/- 508.53, N = 3441561. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 512Threadripper 3970X10K20K30K40K50KSE +/- 87.88, N = 3450471. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 1024Threadripper 3970X12K24K36K48K60KSE +/- 302.03, N = 3580541. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 2048Threadripper 3970X13K26K39K52K65KSE +/- 548.16, N = 3605121. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 4096Threadripper 3970X12K24K36K48K60KSE +/- 549.89, N = 3572251. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 1024Threadripper 3970X10K20K30K40K50KSE +/- 116.74, N = 3455741. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 2048Threadripper 3970X7K14K21K28K35KSE +/- 333.39, N = 3320271. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 4096Threadripper 3970X5K10K15K20K25KSE +/- 7.81, N = 3219931. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

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.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: IP Batch 1D - Data Type: f32Threadripper 3970X0.45410.90821.36231.81642.2705SE +/- 0.00613, N = 32.01809MIN: 1.951. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: IP Batch All - Data Type: f32Threadripper 3970X3691215SE +/- 0.02, N = 311.37MIN: 10.871. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: IP Batch 1D - Data Type: u8s8f32Threadripper 3970X510152025SE +/- 0.05, N = 319.98MIN: 19.061. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: IP Batch All - Data Type: u8s8f32Threadripper 3970X306090120150SE +/- 0.18, N = 3127.14MIN: 122.581. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_3d - Data Type: f32Threadripper 3970X1.07622.15243.22864.30485.381SE +/- 0.01968, N = 34.78292MIN: 4.591. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_all - Data Type: f32Threadripper 3970X2004006008001000SE +/- 3.90, N = 3776.26MIN: 763.611. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_3d - Data Type: u8s8f32Threadripper 3970X8001600240032004000SE +/- 1.09, N = 33724.97MIN: 3712.291. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Deconvolution Batch deconv_1d - Data Type: f32Threadripper 3970X0.41220.82441.23661.64882.061SE +/- 0.00100, N = 31.83211MIN: 1.781. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Deconvolution Batch deconv_3d - Data Type: f32Threadripper 3970X0.60811.21621.82432.43243.0405SE +/- 0.00427, N = 32.70273MIN: 2.631. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_alexnet - Data Type: f32Threadripper 3970X20406080100SE +/- 0.30, N = 397.97MIN: 96.761. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_all - Data Type: u8s8f32Threadripper 3970X3K6K9K12K15KSE +/- 16.64, N = 312308.9MIN: 12070.91. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Deconvolution Batch deconv_all - Data Type: f32Threadripper 3970X30060090012001500SE +/- 2.36, N = 31270.16MIN: 1251.441. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Deconvolution Batch deconv_1d - Data Type: u8s8f32Threadripper 3970X30060090012001500SE +/- 4.74, N = 31395.60MIN: 1385.991. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Deconvolution Batch deconv_3d - Data Type: u8s8f32Threadripper 3970X5001000150020002500SE +/- 12.87, N = 32327.45MIN: 2305.281. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Recurrent Neural Network Training - Data Type: f32Threadripper 3970X4080120160200SE +/- 0.30, N = 3190.04MIN: 187.611. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_alexnet - Data Type: u8s8f32Threadripper 3970X30060090012001500SE +/- 1.28, N = 31356.55MIN: 1346.131. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_googlenet_v3 - Data Type: f32Threadripper 3970X1020304050SE +/- 0.27, N = 342.33MIN: 41.21. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

OpenBenchmarking.orgms, Fewer Is BetterMKL-DNN DNNL 1.1Harness: Convolution Batch conv_googlenet_v3 - Data Type: u8s8f32Threadripper 3970X150300450600750SE +/- 1.31, N = 3696.08MIN: 676.411. (CXX) g++ options: -O3 -march=native -std=c++11 -msse4.1 -fPIC -fopenmp -pie -lpthread -ldl

Embree

Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer - Model: CrownThreadripper 3970X918273645SE +/- 0.10, N = 338.66MIN: 38.14 / MAX: 39.58

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer ISPC - Model: CrownThreadripper 3970X816243240SE +/- 0.06, N = 336.87MIN: 36.37 / MAX: 37.82

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer - Model: Asian DragonThreadripper 3970X1020304050SE +/- 0.06, N = 341.85MIN: 41.47 / MAX: 42.6

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer - Model: Asian Dragon ObjThreadripper 3970X918273645SE +/- 0.06, N = 337.48MIN: 37.1 / MAX: 38.17

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer ISPC - Model: Asian DragonThreadripper 3970X1020304050SE +/- 0.03, N = 341.72MIN: 41.42 / MAX: 42.36

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer ISPC - Model: Asian Dragon ObjThreadripper 3970X816243240SE +/- 0.01, N = 336.00MIN: 35.74 / MAX: 36.59

Intel Open Image Denoise

Open Image Denoise is a denoising library for ray-tracing and part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgImages / Sec, More Is BetterIntel Open Image Denoise 1.0.0Scene: MemorialThreadripper 3970X612182430SE +/- 0.03, N = 324.70

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 2.2Scene: DLSCThreadripper 3970X0.67281.34562.01842.69123.364SE +/- 0.02, N = 32.99MIN: 2.9 / MAX: 3.07

OpenBenchmarking.orgM samples/sec, More Is BetterLuxCoreRender 2.2Scene: Rainbow Colors and PrismThreadripper 3970X0.54231.08461.62692.16922.7115SE +/- 0.15, N = 152.41MIN: 1.17 / MAX: 3.12

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.orgSeconds, Fewer Is BetterTungsten Renderer 0.2.2Scene: HairThreadripper 3970X246810SE +/- 0.00414, N = 38.150371. (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 -ljpeg -lpthread -ldl

OpenBenchmarking.orgSeconds, Fewer Is BetterTungsten Renderer 0.2.2Scene: Water CausticThreadripper 3970X510152025SE +/- 0.01, N = 318.521. (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 -ljpeg -lpthread -ldl

OpenBenchmarking.orgSeconds, Fewer Is BetterTungsten Renderer 0.2.2Scene: Non-ExponentialThreadripper 3970X0.59951.1991.79852.3982.9975SE +/- 0.01210, N = 32.664391. (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 -ljpeg -lpthread -ldl

OpenBenchmarking.orgSeconds, Fewer Is BetterTungsten Renderer 0.2.2Scene: Volumetric CausticThreadripper 3970X0.79461.58922.38383.17843.973SE +/- 0.00240, N = 33.531441. (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 -ljpeg -lpthread -ldl

63 Results Shown

FFTW:
  Stock - 1D FFT Size 32
  Stock - 1D FFT Size 64
  Stock - 2D FFT Size 32
  Stock - 2D FFT Size 64
  Stock - 1D FFT Size 128
  Stock - 1D FFT Size 256
  Stock - 1D FFT Size 512
  Stock - 2D FFT Size 128
  Stock - 2D FFT Size 256
  Stock - 2D FFT Size 512
  Stock - 1D FFT Size 1024
  Stock - 1D FFT Size 2048
  Stock - 1D FFT Size 4096
  Stock - 2D FFT Size 1024
  Stock - 2D FFT Size 2048
  Stock - 2D FFT Size 4096
  Float + SSE - 1D FFT Size 32
  Float + SSE - 1D FFT Size 64
  Float + SSE - 2D FFT Size 32
  Float + SSE - 2D FFT Size 64
  Float + SSE - 1D FFT Size 128
  Float + SSE - 1D FFT Size 256
  Float + SSE - 1D FFT Size 512
  Float + SSE - 2D FFT Size 128
  Float + SSE - 2D FFT Size 256
  Float + SSE - 2D FFT Size 512
  Float + SSE - 1D FFT Size 1024
  Float + SSE - 1D FFT Size 2048
  Float + SSE - 1D FFT Size 4096
  Float + SSE - 2D FFT Size 1024
  Float + SSE - 2D FFT Size 2048
  Float + SSE - 2D FFT Size 4096
MKL-DNN DNNL:
  IP Batch 1D - f32
  IP Batch All - f32
  IP Batch 1D - u8s8f32
  IP Batch All - u8s8f32
  Convolution Batch conv_3d - f32
  Convolution Batch conv_all - f32
  Convolution Batch conv_3d - u8s8f32
  Deconvolution Batch deconv_1d - f32
  Deconvolution Batch deconv_3d - f32
  Convolution Batch conv_alexnet - f32
  Convolution Batch conv_all - u8s8f32
  Deconvolution Batch deconv_all - f32
  Deconvolution Batch deconv_1d - u8s8f32
  Deconvolution Batch deconv_3d - u8s8f32
  Recurrent Neural Network Training - f32
  Convolution Batch conv_alexnet - u8s8f32
  Convolution Batch conv_googlenet_v3 - f32
  Convolution Batch conv_googlenet_v3 - u8s8f32
Embree:
  Pathtracer - Crown
  Pathtracer ISPC - Crown
  Pathtracer - Asian Dragon
  Pathtracer - Asian Dragon Obj
  Pathtracer ISPC - Asian Dragon
  Pathtracer ISPC - Asian Dragon Obj
Intel Open Image Denoise
LuxCoreRender:
  DLSC
  Rainbow Colors and Prism
Tungsten Renderer:
  Hair
  Water Caustic
  Non-Exponential
  Volumetric Caustic

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: AMD Radeon VII 16GB (1801/1000MHz), Audio: AMD Vega 20 HDMI Audio, Monitor: ASUS MG28U, Network: Aquantia AQC107 NBase-T/IEEE + Intel I211 + Intel Wi-Fi 6 AX200

OS: Ubuntu 20.04, Kernel: 5.6.0-050600rc1-generic (x86_64) 20200209, Desktop: GNOME Shell 3.34.3, Display Server: X Server 1.20.7, Display Driver: amdgpu 19.1.0, OpenGL: 4.5 Mesa 19.2.4 (LLVM 9.0.0), Compiler: GCC 9.2.1 20200123, 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-link-mutex --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-build-config=bootstrap-lto-lean --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
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 + tsx_async_abort: Not affected

Testing initiated at 11 February 2020 21:07 by user pts.