AMD Ryzen 9 7950X AVX-512 benchmark comparison by Michael Larabel launch day embargo lift review. Stock/out-of-the-box build with AVX-512. For lack of any AVX-512 toggle from the ASUS BIOS, the AVX2 / non-AVX-512 run was carried out by booting kernel with "clearcpuid=304" to clear AVX-512 support from the kernel and for the binary programs that scan /proc/cpuinfo for avx512* extensions. Plus for the open-source benchmarks specifying CFLAGS/CXXFLAGS without AVX-512 extensions. See full launch day review @ https://www.phoronix.com/review/amd-zen4-avx512
Default, AVX-512 Enabled Kernel Notes: Transparent Huge Pages: madviseEnvironment Notes: CXXFLAGS="-O3 -march=native -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mprefer-vector-width=512" CFLAGS="-O3 -march=native -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mprefer-vector-width=512"Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate performance (Boost: Enabled) - CPU Microcode: 0xa601203Python Notes: Python 3.10.4Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Without AVX-512 Processor: AMD Ryzen 9 7950X 16-Core @ 5.88GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR X670E HERO (0604 BIOS), Chipset: AMD Device 14d8, Memory: 32GB, Disk: 2000GB Samsung SSD 980 PRO 2TB + 2000GB, Graphics: AMD Radeon RX 6800 XT 16GB (2575/1000MHz), Audio: AMD Navi 21 HDMI Audio, Monitor: ASUS VP28U, Network: Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411
OS: Ubuntu 22.04, Kernel: 6.0.0-060000rc1daily20220820-generic (x86_64), Desktop: GNOME Shell 42.2, Display Server: X Server + Wayland, OpenGL: 4.6 Mesa 22.3.0-devel (git-4685385 2022-08-23 jammy-oibaf-ppa) (LLVM 14.0.6 DRM 3.48), Vulkan: 1.3.224, Compiler: GCC 12.0.1 20220319, File-System: ext4, Screen Resolution: 3840x2160
Kernel Notes: Transparent Huge Pages: madviseEnvironment Notes: CXXFLAGS="-O3 -march=native -mno-avx512f" CFLAGS="-O3 -march=native -mno-avx512f"Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate performance (Boost: Enabled) - CPU Microcode: 0xa601203Python Notes: Python 3.10.4Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
AMD Ryzen 9 7950X AVX-512 OpenBenchmarking.org Phoronix Test Suite AMD Ryzen 9 7950X 16-Core @ 5.88GHz (16 Cores / 32 Threads) ASUS ROG CROSSHAIR X670E HERO (0604 BIOS) AMD Device 14d8 32GB 2000GB Samsung SSD 980 PRO 2TB + 2000GB AMD Radeon RX 6800 XT 16GB (2575/1000MHz) AMD Navi 21 HDMI Audio ASUS VP28U Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411 Ubuntu 22.04 6.0.0-060000rc1daily20220820-generic (x86_64) GNOME Shell 42.2 X Server + Wayland 4.6 Mesa 22.3.0-devel (git-4685385 2022-08-23 jammy-oibaf-ppa) (LLVM 14.0.6 DRM 3.48) 1.3.224 GCC 12.0.1 20220319 ext4 3840x2160 Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server OpenGL Vulkan Compiler File-System Screen Resolution AMD Ryzen 9 7950X AVX-512 Benchmarks System Logs - Transparent Huge Pages: madvise - Default, AVX-512 Enabled: CXXFLAGS="-O3 -march=native -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mprefer-vector-width=512" CFLAGS="-O3 -march=native -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mprefer-vector-width=512" - Without AVX-512: CXXFLAGS="-O3 -march=native -mno-avx512f" CFLAGS="-O3 -march=native -mno-avx512f" - --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v - Scaling Governor: amd-pstate performance (Boost: Enabled) - CPU Microcode: 0xa601203 - Python 3.10.4 - itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Default, AVX-512 Enabled vs. Without AVX-512 Comparison Phoronix Test Suite Baseline +66.4% +66.4% +132.8% +132.8% +199.2% +199.2% +265.6% +265.6% Myriad-Groestl 265.7% W.P.D.F.I - CPU 117.1% W.P.D.F.I - CPU 117% F.D.F - CPU 115.1% F.D.F - CPU 114.4% W.P.D.F - CPU 113.7% W.P.D.F - CPU 113.6% F.D.F.I - CPU 109.1% F.D.F.I - CPU 108.6% Blake-2 S 104.5% M.T.E.T.D.F - CPU 103.8% M.T.E.T.D.F - CPU 103.8% A.G.R.R.0.F - CPU 100% A.G.R.R.0.F - CPU 100% V.D.F.I - CPU 98.2% V.D.F.I - CPU 98.1% LBC, LBRY Credits 97.1% D.B.s - u8s8f32 - CPU 92.4% D.B.s - u8s8f32 - CPU 88% P.V.B.D.F - CPU 87% P.V.B.D.F - CPU 86.8% gravity_spheres_volume/dim_512/scivis/real_time 83.1% P.D.F - CPU 80.2% P.D.F - CPU 79.9% P.D.F - CPU 79.8% P.D.F - CPU 78.9% gravity_spheres_volume/dim_512/ao/real_time 77.6% A.G.R.R.0.F.I - CPU 70.8% A.G.R.R.0.F.I - CPU 68.2% Q.S.2.P 66.9% CPU - vision_transformer 58.6% R.N.N.I - u8s8f32 - CPU 56.1% R.N.N.I - bf16bf16bf16 - CPU 54.8% Garlicoin 53.1% IP Shapes 1D - u8s8f32 - CPU 50% V.D.F - CPU 48% V.D.F - CPU 47.9% C.B.S.A - u8s8f32 - CPU 40% Skeincoin 39.3% R.N.N.T - u8s8f32 - CPU 37.7% gravity_spheres_volume/dim_512/pathtracer/real_time 37.4% R.N.N.T - bf16bf16bf16 - CPU 35.7% DistinctUserID 32.6% vklBenchmark ISPC 30.1% TopTweet 26.1% PartialTweets 25.4% 1 - 4K - 1 - Path Tracer 23% 3 - 4K - 1 - Path Tracer 22.4% Kostya 21.4% 1 - 4K - 32 - Path Tracer 21.1% Pathtracer ISPC - Crown 21% 1 - 4K - 16 - Path Tracer 20.7% 3 - 4K - 32 - Path Tracer 20.1% 3 - 4K - 16 - Path Tracer 20% Pathtracer ISPC - Asian Dragon 15.6% LargeRand 14% SqueezeNetV1.0 12.8% Eigen 9.9% particle_volume/pathtracer/real_time 8.3% Summer Nature 4K 3.4% S.N.1 2.8% Cpuminer-Opt OpenVINO OpenVINO OpenVINO OpenVINO OpenVINO OpenVINO OpenVINO OpenVINO Cpuminer-Opt OpenVINO OpenVINO OpenVINO OpenVINO OpenVINO OpenVINO Cpuminer-Opt oneDNN oneDNN OpenVINO OpenVINO OSPRay OpenVINO OpenVINO OpenVINO OpenVINO OSPRay OpenVINO OpenVINO Cpuminer-Opt NCNN oneDNN oneDNN Cpuminer-Opt oneDNN OpenVINO OpenVINO oneDNN Cpuminer-Opt oneDNN OSPRay oneDNN simdjson OpenVKL simdjson simdjson OSPRay Studio OSPRay Studio simdjson OSPRay Studio Embree OSPRay Studio OSPRay Studio OSPRay Studio Embree simdjson Mobile Neural Network LeelaChessZero OSPRay dav1d dav1d Default, AVX-512 Enabled Without AVX-512
AMD Ryzen 9 7950X AVX-512 ospray: particle_volume/pathtracer/real_time lczero: Eigen openvkl: vklBenchmark ISPC mnn: SqueezeNetV1.0 ospray-studio: 3 - 4K - 32 - Path Tracer ospray-studio: 3 - 4K - 1 - Path Tracer ospray-studio: 1 - 4K - 1 - Path Tracer ospray-studio: 1 - 4K - 32 - Path Tracer ospray: gravity_spheres_volume/dim_512/scivis/real_time ospray: gravity_spheres_volume/dim_512/ao/real_time ospray: gravity_spheres_volume/dim_512/pathtracer/real_time cpuminer-opt: Myriad-Groestl simdjson: Kostya ospray-studio: 3 - 4K - 16 - Path Tracer simdjson: LargeRand cpuminer-opt: Garlicoin cpuminer-opt: Blake-2 S onednn: Recurrent Neural Network Training - u8s8f32 - CPU onednn: Recurrent Neural Network Training - bf16bf16bf16 - CPU onednn: Recurrent Neural Network Inference - u8s8f32 - CPU onednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPU ospray-studio: 1 - 4K - 16 - Path Tracer openvino: Person Detection FP16 - CPU openvino: Person Detection FP16 - CPU openvino: Person Detection FP32 - CPU openvino: Person Detection FP32 - CPU simdjson: DistinctUserID openvino: Face Detection FP16 - CPU openvino: Face Detection FP16 - CPU simdjson: PartialTweets simdjson: TopTweet openvino: Face Detection FP16-INT8 - CPU openvino: Face Detection FP16-INT8 - CPU openvino: Machine Translation EN To DE FP16 - CPU openvino: Machine Translation EN To DE FP16 - CPU openvino: Person Vehicle Bike Detection FP16 - CPU openvino: Person Vehicle Bike Detection FP16 - CPU ncnn: CPU - vision_transformer openvino: Age Gender Recognition Retail 0013 FP16-INT8 - CPU openvino: Age Gender Recognition Retail 0013 FP16-INT8 - CPU openvino: Vehicle Detection FP16-INT8 - CPU openvino: Vehicle Detection FP16-INT8 - CPU openvino: Age Gender Recognition Retail 0013 FP16 - CPU openvino: Age Gender Recognition Retail 0013 FP16 - CPU openvino: Vehicle Detection FP16 - CPU openvino: Vehicle Detection FP16 - CPU openvino: Weld Porosity Detection FP16 - CPU openvino: Weld Porosity Detection FP16 - CPU openvino: Weld Porosity Detection FP16-INT8 - CPU openvino: Weld Porosity Detection FP16-INT8 - CPU onednn: IP Shapes 1D - u8s8f32 - CPU cpuminer-opt: LBC, LBRY Credits cpuminer-opt: Quad SHA-256, Pyrite cpuminer-opt: Skeincoin onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPU embree: Pathtracer ISPC - Crown embree: Pathtracer ISPC - Asian Dragon dav1d: Summer Nature 4K onednn: Convolution Batch Shapes Auto - u8s8f32 - CPU onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPU dav1d: Summer Nature 1080p Default, AVX-512 Enabled Without AVX-512 250.655 1735 199 3.543 143981 4368 3659 120727 7.93432 8.05397 9.50906 59957 6.01 73617 1.79 3977.80 2113434 1137.29 1137.96 583.316 582.901 62029 1050.37 7.57 1054.92 7.55 10.26 549.94 14.50 9.64 9.96 280.86 28.44 59.39 134.59 4.71 1696.82 39.57 0.24 64463.94 4.23 1890.25 0.35 45555.67 10.76 742.90 5.49 1455.45 5.44 2938.13 0.417196 151720 321503 284667 0.415940 35.2293 35.1217 393.40 5.32127 0.576432 1451.94 231.378 1578 153 3.997 172872 5348 4502 146156 4.33398 4.53529 6.92227 16393 4.95 88344 1.57 2597.65 1033700 1566.09 1544.39 910.460 902.102 74844 1889.40 4.20 1887.13 4.20 7.74 1178.81 6.74 7.69 7.90 585.81 13.60 121.06 66.03 8.80 907.36 62.74 0.41 38316.87 8.38 953.79 0.7 22781.70 15.92 502.28 11.73 681.37 11.81 1353.84 0.625942 76960 192667 204320 0.800116 29.1170 30.3870 380.32 7.44936 1.08396 1412.27 OpenBenchmarking.org
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.10 Benchmark: particle_volume/pathtracer/real_time Default, AVX-512 Enabled Without AVX-512 50 100 150 200 250 SE +/- 0.91, N = 3 SE +/- 0.59, N = 3 250.66 231.38 Items Per Second Per Watt
OpenBenchmarking.org Items Per Second Per Watt, More Is Better OSPRay 2.10 Benchmark: particle_volume/pathtracer/real_time Default, AVX-512 Enabled Without AVX-512 0.3137 0.6274 0.9411 1.2548 1.5685 1.394 1.237 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5127 5270 5881 Without AVX-512 5070 5185 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay 2.10 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.4 179.8 215.5 Without AVX-512 22.9 187.0 222.6 OpenBenchmarking.org Watts, Fewer Is Better OSPRay 2.10 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 43.1 86.3 93.0 Without AVX-512 45.0 90.8 95.8 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay 2.10 CPU Temperature Monitor 20 40 60 80 100
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.
Result
OpenBenchmarking.org Nodes Per Second, More Is Better LeelaChessZero 0.28 Backend: Eigen Default, AVX-512 Enabled Without AVX-512 400 800 1200 1600 2000 SE +/- 7.31, N = 3 SE +/- 16.33, N = 3 1735 1578 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -flto -O3 -march=native -pthread Nodes Per Second Per Watt
OpenBenchmarking.org Nodes Per Second Per Watt, More Is Better LeelaChessZero 0.28 Backend: Eigen Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 10.156 9.107 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5191 5321 5881 Without AVX-512 5152 5307 5881 OpenBenchmarking.org Megahertz, More Is Better LeelaChessZero 0.28 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.5 170.8 192.1 Without AVX-512 23.3 173.3 195.1 OpenBenchmarking.org Watts, Fewer Is Better LeelaChessZero 0.28 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 33.6 84.4 87.4 Without AVX-512 36.1 88.3 91.0 OpenBenchmarking.org Celsius, Fewer Is Better LeelaChessZero 0.28 CPU Temperature Monitor 20 40 60 80 100
OpenVKL OpenVKL is the Intel Open Volume Kernel Library that offers high-performance volume computation kernels and part of the Intel oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org Items / Sec, More Is Better OpenVKL 1.0 Benchmark: vklBenchmark ISPC Default, AVX-512 Enabled Without AVX-512 40 80 120 160 200 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 199 153 MIN: 17 / MAX: 2254 MIN: 14 / MAX: 1843 Items / Sec Per Watt
OpenBenchmarking.org Items / Sec Per Watt, More Is Better OpenVKL 1.0 Benchmark: vklBenchmark ISPC Default, AVX-512 Enabled Without AVX-512 0.2873 0.5746 0.8619 1.1492 1.4365 1.277 0.912 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5168 5403 5881 Without AVX-512 5072 5346 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVKL 1.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.4 155.8 211.8 Without AVX-512 23.6 167.7 228.1 OpenBenchmarking.org Watts, Fewer Is Better OpenVKL 1.0 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 43.8 79.1 89.5 Without AVX-512 43.9 84.9 95.4 OpenBenchmarking.org Celsius, Fewer Is Better OpenVKL 1.0 CPU Temperature Monitor 20 40 60 80 100
Mobile Neural Network MNN is the Mobile Neural Network as a highly efficient, lightweight deep learning framework developed by Alibaba. This MNN test profile is building the OpenMP / CPU threaded version for processor benchmarking and not any GPU-accelerated test. MNN does allow making use of AVX-512 extensions. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Mobile Neural Network 2.1 Model: SqueezeNetV1.0 Default, AVX-512 Enabled Without AVX-512 0.8993 1.7986 2.6979 3.5972 4.4965 SE +/- 0.015, N = 6 SE +/- 0.121, N = 3 3.543 3.997 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 3.43 / MAX: 10.23 -mno-avx512f - MIN: 3.72 / MAX: 5.69 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fvisibility=hidden -fomit-frame-pointer -fstrict-aliasing -ffunction-sections -fdata-sections -ffast-math -fno-rtti -fno-exceptions -rdynamic -pthread -ldl
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 0.11 Camera: 3 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer Default, AVX-512 Enabled Without AVX-512 40K 80K 120K 160K 200K SE +/- 116.66, N = 3 SE +/- 155.59, N = 3 143981 172872 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -ldl CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5131 5220 5881 Without AVX-512 5118 5194 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay Studio 0.11 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.9 199.7 222.9 Without AVX-512 23.1 201.2 218.3 OpenBenchmarking.org Watts, Fewer Is Better OSPRay Studio 0.11 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.4 90.0 93.4 Without AVX-512 48.0 92.3 95.1 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay Studio 0.11 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 0.11 Camera: 3 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer Default, AVX-512 Enabled Without AVX-512 1100 2200 3300 4400 5500 SE +/- 5.55, N = 3 SE +/- 3.28, N = 3 4368 5348 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -ldl CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5134 5227 5881 Without AVX-512 5117 5207 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay Studio 0.11 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.9 198.4 217.9 Without AVX-512 22.8 197.3 224.0 OpenBenchmarking.org Watts, Fewer Is Better OSPRay Studio 0.11 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.0 89.7 92.5 Without AVX-512 47.6 91.8 94.5 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay Studio 0.11 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 0.11 Camera: 1 - Resolution: 4K - Samples Per Pixel: 1 - Renderer: Path Tracer Default, AVX-512 Enabled Without AVX-512 1000 2000 3000 4000 5000 SE +/- 5.21, N = 3 SE +/- 6.51, N = 3 3659 4502 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -ldl CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5136 5230 5881 Without AVX-512 5123 5212 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay Studio 0.11 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 17.2 198.2 221.8 Without AVX-512 17.7 198.8 223.6 OpenBenchmarking.org Watts, Fewer Is Better OSPRay Studio 0.11 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 44.9 89.5 92.5 Without AVX-512 41.9 91.0 94.1 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay Studio 0.11 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 0.11 Camera: 1 - Resolution: 4K - Samples Per Pixel: 32 - Renderer: Path Tracer Default, AVX-512 Enabled Without AVX-512 30K 60K 90K 120K 150K SE +/- 140.99, N = 3 SE +/- 168.25, N = 3 120727 146156 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -ldl CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5118 5229 5881 Without AVX-512 5106 5202 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay Studio 0.11 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 14.7 198.1 226.1 Without AVX-512 16.1 199.9 221.5 OpenBenchmarking.org Watts, Fewer Is Better OSPRay Studio 0.11 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 46.6 90.0 93.5 Without AVX-512 48.1 92.3 95.3 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay Studio 0.11 CPU Temperature Monitor 20 40 60 80 100
OSPRay Intel OSPRay is a portable ray-tracing engine for high-performance, high-fidelity scientific visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.10 Benchmark: gravity_spheres_volume/dim_512/scivis/real_time Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.03212, N = 3 SE +/- 0.00159, N = 3 7.93432 4.33398 Items Per Second Per Watt
OpenBenchmarking.org Items Per Second Per Watt, More Is Better OSPRay 2.10 Benchmark: gravity_spheres_volume/dim_512/scivis/real_time Default, AVX-512 Enabled Without AVX-512 0.0095 0.019 0.0285 0.038 0.0475 0.042 0.022 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5166 5235 5881 Without AVX-512 5121 5201 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay 2.10 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.4 188.0 204.2 Without AVX-512 23.1 193.1 207.2 OpenBenchmarking.org Watts, Fewer Is Better OSPRay 2.10 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 46.4 87.3 89.4 Without AVX-512 47.4 92.0 93.9 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay 2.10 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.10 Benchmark: gravity_spheres_volume/dim_512/ao/real_time Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.00437, N = 3 SE +/- 0.00574, N = 3 8.05397 4.53529 Items Per Second Per Watt
OpenBenchmarking.org Items Per Second Per Watt, More Is Better OSPRay 2.10 Benchmark: gravity_spheres_volume/dim_512/ao/real_time Default, AVX-512 Enabled Without AVX-512 0.0097 0.0194 0.0291 0.0388 0.0485 0.043 0.023 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5165 5243 5881 Without AVX-512 5131 5201 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay 2.10 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.5 189.2 205.4 Without AVX-512 23.1 193.5 209.6 OpenBenchmarking.org Watts, Fewer Is Better OSPRay 2.10 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 42.5 87.0 89.1 Without AVX-512 44.8 91.4 93.5 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay 2.10 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org Items Per Second, More Is Better OSPRay 2.10 Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 SE +/- 0.00541, N = 3 SE +/- 0.00119, N = 3 9.50906 6.92227 Items Per Second Per Watt
OpenBenchmarking.org Items Per Second Per Watt, More Is Better OSPRay 2.10 Benchmark: gravity_spheres_volume/dim_512/pathtracer/real_time Default, AVX-512 Enabled Without AVX-512 0.0113 0.0226 0.0339 0.0452 0.0565 0.050 0.035 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5155 5225 5881 Without AVX-512 5107 5187 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay 2.10 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.8 191.7 209.5 Without AVX-512 23.4 197.7 214.7 OpenBenchmarking.org Watts, Fewer Is Better OSPRay 2.10 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 46.3 88.9 91.0 Without AVX-512 48.0 93.3 95.1 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay 2.10 CPU Temperature Monitor 20 40 60 80 100
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 3.18 Algorithm: Myriad-Groestl Default, AVX-512 Enabled Without AVX-512 13K 26K 39K 52K 65K SE +/- 669.95, N = 15 SE +/- 105.88, N = 3 59957 16393 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -lcurl -lz -lpthread -lssl -lcrypto -lgmp kH/s Per Watt
OpenBenchmarking.org kH/s Per Watt, More Is Better Cpuminer-Opt 3.18 Algorithm: Myriad-Groestl Default, AVX-512 Enabled Without AVX-512 110 220 330 440 550 496.89 128.78 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5298 5503 5881 Without AVX-512 5344 5449 5881 OpenBenchmarking.org Megahertz, More Is Better Cpuminer-Opt 3.18 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 21.7 120.7 179.8 Without AVX-512 23.6 127.3 150.7 OpenBenchmarking.org Watts, Fewer Is Better Cpuminer-Opt 3.18 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 42.1 67.6 71.3 Without AVX-512 43.3 67.7 73.0 OpenBenchmarking.org Celsius, Fewer Is Better Cpuminer-Opt 3.18 CPU Temperature Monitor 20 40 60 80 100
simdjson This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org GB/s, More Is Better simdjson 2.0 Throughput Test: Kostya Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.04, N = 3 SE +/- 0.00, N = 3 6.01 4.95 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native GB/s Per Watt
OpenBenchmarking.org GB/s Per Watt, More Is Better simdjson 2.0 Throughput Test: Kostya Default, AVX-512 Enabled Without AVX-512 0.0212 0.0424 0.0636 0.0848 0.106 0.094 0.076 CPU Peak Freq (Highest CPU Core Frequency
OpenBenchmarking.org Megahertz, More Is Better simdjson 2.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor Without AVX-512 1300 2600 3900 5200 6500 5881 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 21.9 63.7 66.4 Without AVX-512 16.7 65.6 68.6 OpenBenchmarking.org Watts, Fewer Is Better simdjson 2.0 CPU Power Consumption Monitor 20 40 60 80 100 CPU Temp
Min Avg Max Default, AVX-512 Enabled 39.0 67.1 68.8 Without AVX-512 39.4 67.7 68.8 OpenBenchmarking.org Celsius, Fewer Is Better simdjson 2.0 CPU Temperature Monitor 20 40 60 80 100
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 0.11 Camera: 3 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer Default, AVX-512 Enabled Without AVX-512 20K 40K 60K 80K 100K SE +/- 149.41, N = 3 SE +/- 7.57, N = 3 73617 88344 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -ldl CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5128 5264 5881 Without AVX-512 5122 5240 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay Studio 0.11 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.9 188.7 219.2 Without AVX-512 23.0 192.2 220.6 OpenBenchmarking.org Watts, Fewer Is Better OSPRay Studio 0.11 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.5 88.3 93.0 Without AVX-512 48.3 90.4 94.5 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay Studio 0.11 CPU Temperature Monitor 20 40 60 80 100
simdjson This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org GB/s, More Is Better simdjson 2.0 Throughput Test: LargeRandom Default, AVX-512 Enabled Without AVX-512 0.4028 0.8056 1.2084 1.6112 2.014 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 1.79 1.57 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native GB/s Per Watt
OpenBenchmarking.org GB/s Per Watt, More Is Better simdjson 2.0 Throughput Test: LargeRandom Default, AVX-512 Enabled Without AVX-512 0.0065 0.013 0.0195 0.026 0.0325 0.029 0.025 CPU Peak Freq (Highest CPU Core Frequency
OpenBenchmarking.org Megahertz, More Is Better simdjson 2.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor Without AVX-512 1300 2600 3900 5200 6500 5881 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.2 62.7 65.8 Without AVX-512 22.9 62.6 65.3 OpenBenchmarking.org Watts, Fewer Is Better simdjson 2.0 CPU Power Consumption Monitor 20 40 60 80 100 CPU Temp
Min Avg Max Default, AVX-512 Enabled 38.5 66.6 68.0 Without AVX-512 38.9 68.8 69.9 OpenBenchmarking.org Celsius, Fewer Is Better simdjson 2.0 CPU Temperature Monitor 20 40 60 80 100
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 3.18 Algorithm: Garlicoin Default, AVX-512 Enabled Without AVX-512 900 1800 2700 3600 4500 SE +/- 75.31, N = 12 SE +/- 25.96, N = 3 3977.80 2597.65 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -lcurl -lz -lpthread -lssl -lcrypto -lgmp kH/s Per Watt
OpenBenchmarking.org kH/s Per Watt, More Is Better Cpuminer-Opt 3.18 Algorithm: Garlicoin Default, AVX-512 Enabled Without AVX-512 6 12 18 24 30 24.63 15.45 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5283 5374 5881 Without AVX-512 5244 5339 5881 OpenBenchmarking.org Megahertz, More Is Better Cpuminer-Opt 3.18 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.9 161.5 185.7 Without AVX-512 23.1 168.2 190.7 OpenBenchmarking.org Watts, Fewer Is Better Cpuminer-Opt 3.18 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 40.8 72.9 76.3 Without AVX-512 44.3 75.9 79.9 OpenBenchmarking.org Celsius, Fewer Is Better Cpuminer-Opt 3.18 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 3.18 Algorithm: Blake-2 S Default, AVX-512 Enabled Without AVX-512 500K 1000K 1500K 2000K 2500K SE +/- 38034.65, N = 12 SE +/- 4192.40, N = 3 2113434 1033700 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -lcurl -lz -lpthread -lssl -lcrypto -lgmp kH/s Per Watt
OpenBenchmarking.org kH/s Per Watt, More Is Better Cpuminer-Opt 3.18 Algorithm: Blake-2 S Default, AVX-512 Enabled Without AVX-512 4K 8K 12K 16K 20K 19614.14 9800.37 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5260 5542 5881 Without AVX-512 5204 5521 5881 OpenBenchmarking.org Megahertz, More Is Better Cpuminer-Opt 3.18 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 19.2 107.8 200.8 Without AVX-512 21.0 105.5 175.6 OpenBenchmarking.org Watts, Fewer Is Better Cpuminer-Opt 3.18 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.1 66.4 79.6 Without AVX-512 48.6 69.2 85.9 OpenBenchmarking.org Celsius, Fewer Is Better Cpuminer-Opt 3.18 CPU Temperature Monitor 20 40 60 80 100
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 Intel oneAPI. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 300 600 900 1200 1500 SE +/- 1.54, N = 3 SE +/- 8.37, N = 3 1137.29 1566.09 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 1130.88 -mno-avx512f - MIN: 1541.57 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 4822 5166 5881 Without AVX-512 4772 5007 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.8 200.9 226.3 Without AVX-512 23.2 187.5 219.5 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 42.5 87.8 92.0 Without AVX-512 43.5 88.1 95.9 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 300 600 900 1200 1500 SE +/- 1.66, N = 3 SE +/- 19.60, N = 3 1137.96 1544.39 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 1131.23 -mno-avx512f - MIN: 1494.51 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 4560 5163 5881 Without AVX-512 4745 4985 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 20.3 200.1 223.4 Without AVX-512 23.4 189.5 220.2 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.5 88.8 92.0 Without AVX-512 47.3 89.8 96.1 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 200 400 600 800 1000 SE +/- 0.21, N = 3 SE +/- 3.36, N = 3 583.32 910.46 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 579.29 -mno-avx512f - MIN: 898.05 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 4802 5091 5881 Without AVX-512 4804 5051 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.8 200.3 227.2 Without AVX-512 23.3 181.1 224.4 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.0 89.2 92.6 Without AVX-512 47.5 86.4 95.8 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 200 400 600 800 1000 SE +/- 1.02, N = 3 SE +/- 5.88, N = 3 582.90 902.10 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 578.18 -mno-avx512f - MIN: 885.67 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 4421 5088 5881 Without AVX-512 4495 5058 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.4 200.3 226.4 Without AVX-512 23.1 180.3 220.1 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.5 88.9 92.8 Without AVX-512 47.6 86.8 96.0 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
OSPRay Studio Intel OSPRay Studio is an open-source, interactive visualization and ray-tracing software package. OSPRay Studio makes use of Intel OSPRay, a portable ray-tracing engine for high-performance, high-fidelity visualizations. OSPRay builds off Intel's Embree and Intel SPMD Program Compiler (ISPC) components as part of the oneAPI rendering toolkit. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OSPRay Studio 0.11 Camera: 1 - Resolution: 4K - Samples Per Pixel: 16 - Renderer: Path Tracer Default, AVX-512 Enabled Without AVX-512 16K 32K 48K 64K 80K SE +/- 138.54, N = 3 SE +/- 91.34, N = 3 62029 74844 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -ldl CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5142 5295 5881 Without AVX-512 5118 5254 5881 OpenBenchmarking.org Megahertz, More Is Better OSPRay Studio 0.11 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 19.2 185.5 223.2 Without AVX-512 12.7 189.2 223.7 OpenBenchmarking.org Watts, Fewer Is Better OSPRay Studio 0.11 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 46.9 87.8 92.4 Without AVX-512 48.0 90.2 95.0 OpenBenchmarking.org Celsius, Fewer Is Better OSPRay Studio 0.11 CPU Temperature Monitor 20 40 60 80 100
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Person Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 400 800 1200 1600 2000 SE +/- 2.40, N = 3 SE +/- 0.47, N = 3 1050.37 1889.40 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 604.91 / MAX: 1254.32 -mno-avx512f - MIN: 1032.21 / MAX: 2215.28 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5035 5229 5881 Without AVX-512 4947 5189 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.2 199.4 219.9 Without AVX-512 10.2 206.7 232.4 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.1 93.2 96.1 Without AVX-512 47.6 94.0 96.6 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Person Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 7.57 4.20 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Person Detection FP32 - Device: CPU Default, AVX-512 Enabled Without AVX-512 400 800 1200 1600 2000 SE +/- 3.05, N = 3 SE +/- 5.93, N = 3 1054.92 1887.13 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 641.2 / MAX: 1276.51 -mno-avx512f - MIN: 1054.92 / MAX: 2161.5 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5059 5219 5881 Without AVX-512 4917 5157 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.0 199.5 219.7 Without AVX-512 23.4 205.1 233.4 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.3 93.3 96.3 Without AVX-512 48.4 94.1 96.8 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Person Detection FP32 - Device: CPU Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 7.55 4.20 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
simdjson This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org GB/s, More Is Better simdjson 2.0 Throughput Test: DistinctUserID Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 10.26 7.74 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native GB/s Per Watt
OpenBenchmarking.org GB/s Per Watt, More Is Better simdjson 2.0 Throughput Test: DistinctUserID Default, AVX-512 Enabled Without AVX-512 0.0371 0.0742 0.1113 0.1484 0.1855 0.165 0.122 CPU Peak Freq (Highest CPU Core Frequency
OpenBenchmarking.org Megahertz, More Is Better simdjson 2.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor Without AVX-512 1300 2600 3900 5200 6500 5881 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.3 62.1 66.4 Without AVX-512 23.1 63.4 67.0 OpenBenchmarking.org Watts, Fewer Is Better simdjson 2.0 CPU Power Consumption Monitor 20 40 60 80 100 CPU Temp
Min Avg Max Default, AVX-512 Enabled 38.9 66.7 68.6 Without AVX-512 38.9 67.2 69.5 OpenBenchmarking.org Celsius, Fewer Is Better simdjson 2.0 CPU Temperature Monitor 20 40 60 80 100
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Face Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 300 600 900 1200 1500 SE +/- 0.48, N = 3 SE +/- 0.92, N = 3 549.94 1178.81 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 282.71 / MAX: 577.01 -mno-avx512f - MIN: 596.16 / MAX: 1250.72 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5038 5180 5881 Without AVX-512 4988 5151 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.3 209.6 235.1 Without AVX-512 20.6 212.4 235.9 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 40.9 93.1 95.9 Without AVX-512 41.8 93.0 96.3 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Face Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 4 8 12 16 20 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 14.50 6.74 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
simdjson This is a benchmark of SIMDJSON, a high performance JSON parser. SIMDJSON aims to be the fastest JSON parser and is used by projects like Microsoft FishStore, Yandex ClickHouse, Shopify, and others. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org GB/s, More Is Better simdjson 2.0 Throughput Test: PartialTweets Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 SE +/- 0.08, N = 3 SE +/- 0.01, N = 3 9.64 7.69 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native GB/s Per Watt
OpenBenchmarking.org GB/s Per Watt, More Is Better simdjson 2.0 Throughput Test: PartialTweets Default, AVX-512 Enabled Without AVX-512 0.0349 0.0698 0.1047 0.1396 0.1745 0.155 0.121 CPU Peak Freq (Highest CPU Core Frequency
OpenBenchmarking.org Megahertz, More Is Better simdjson 2.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor Without AVX-512 1300 2600 3900 5200 6500 5881 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.4 62.3 66.2 Without AVX-512 23.5 63.8 67.3 OpenBenchmarking.org Watts, Fewer Is Better simdjson 2.0 CPU Power Consumption Monitor 20 40 60 80 100 CPU Temp
Min Avg Max Default, AVX-512 Enabled 39.0 65.3 68.4 Without AVX-512 39.3 66.3 67.6 OpenBenchmarking.org Celsius, Fewer Is Better simdjson 2.0 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org GB/s, More Is Better simdjson 2.0 Throughput Test: TopTweet Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 SE +/- 0.09, N = 3 SE +/- 0.02, N = 3 9.96 7.90 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native GB/s Per Watt
OpenBenchmarking.org GB/s Per Watt, More Is Better simdjson 2.0 Throughput Test: TopTweet Default, AVX-512 Enabled Without AVX-512 0.0356 0.0712 0.1068 0.1424 0.178 0.158 0.125 CPU Peak Freq (Highest CPU Core Frequency
OpenBenchmarking.org Megahertz, More Is Better simdjson 2.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor Without AVX-512 1300 2600 3900 5200 6500 5881 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.0 62.9 66.5 Without AVX-512 23.2 63.4 67.0 OpenBenchmarking.org Watts, Fewer Is Better simdjson 2.0 CPU Power Consumption Monitor 20 40 60 80 100 CPU Temp
Min Avg Max Default, AVX-512 Enabled 39.0 65.4 66.6 Without AVX-512 39.5 66.7 67.9 OpenBenchmarking.org Celsius, Fewer Is Better simdjson 2.0 CPU Temperature Monitor 20 40 60 80 100
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Face Detection FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 130 260 390 520 650 SE +/- 0.17, N = 3 SE +/- 0.73, N = 3 280.86 585.81 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 263.49 / MAX: 294.78 -mno-avx512f - MIN: 556.1 / MAX: 596 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5069 5188 5881 Without AVX-512 4963 5101 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.2 207.0 227.9 Without AVX-512 22.9 206.6 233.0 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 45.8 93.2 96.0 Without AVX-512 48.3 93.3 95.6 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Face Detection FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 7 14 21 28 35 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 28.44 13.60 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Machine Translation EN To DE FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 30 60 90 120 150 SE +/- 0.18, N = 3 SE +/- 0.21, N = 3 59.39 121.06 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 26.85 / MAX: 70.78 -mno-avx512f - MIN: 57.76 / MAX: 141.8 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5139 5243 5881 Without AVX-512 4888 5140 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.6 191.8 217.6 Without AVX-512 23.2 205.1 225.9 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.3 87.2 90.1 Without AVX-512 48.6 93.7 96.8 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Machine Translation EN To DE FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 30 60 90 120 150 SE +/- 0.42, N = 3 SE +/- 0.12, N = 3 134.59 66.03 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Person Vehicle Bike Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 4.71 8.80 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 3.46 / MAX: 12.8 -mno-avx512f - MIN: 5.34 / MAX: 19.18 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5100 5188 5881 Without AVX-512 4899 5114 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 19.8 209.2 227.6 Without AVX-512 23.7 209.7 228.7 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.5 91.1 94.1 Without AVX-512 48.3 93.4 96.0 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Person Vehicle Bike Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 400 800 1200 1600 2000 SE +/- 3.05, N = 3 SE +/- 1.39, N = 3 1696.82 907.36 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
NCNN NCNN is a high performance neural network inference framework optimized for mobile and other platforms developed by Tencent. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better NCNN 20220729 Target: CPU - Model: vision_transformer Default, AVX-512 Enabled Without AVX-512 14 28 42 56 70 SE +/- 0.03, N = 3 SE +/- 0.91, N = 3 39.57 62.74 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 39.32 / MAX: 41.99 -mno-avx512f - MIN: 61.57 / MAX: 65.21 1. (CXX) g++ options: -O3 -march=native -rdynamic -lgomp -lpthread
OpenVINO This is a test of the Intel OpenVINO, a toolkit around neural networks, using its built-in benchmarking support and analyzing the throughput and latency for various models. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 0.0923 0.1846 0.2769 0.3692 0.4615 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.24 0.41 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.15 / MAX: 7.55 -mno-avx512f - MIN: 0.22 / MAX: 7.87 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5062 5142 5881 Without AVX-512 4999 5114 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 20.0 217.4 237.8 Without AVX-512 23.1 208.4 234.1 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 48.0 93.2 95.5 Without AVX-512 48.5 93.7 95.6 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Age Gender Recognition Retail 0013 FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 14K 28K 42K 56K 70K SE +/- 24.16, N = 3 SE +/- 43.53, N = 3 64463.94 38316.87 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Vehicle Detection FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 4.23 8.38 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 2.63 / MAX: 12.45 -mno-avx512f - MIN: 4.65 / MAX: 17.13 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5066 5147 5881 Without AVX-512 5027 5125 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.8 212.9 234.8 Without AVX-512 23.5 209.4 234.1 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.3 93.3 95.4 Without AVX-512 48.3 93.5 95.4 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Vehicle Detection FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 400 800 1200 1600 2000 SE +/- 2.10, N = 3 SE +/- 0.65, N = 3 1890.25 953.79 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 0.1575 0.315 0.4725 0.63 0.7875 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.35 0.70 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.21 / MAX: 7.78 -mno-avx512f - MIN: 0.39 / MAX: 8.82 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5006 5107 5881 Without AVX-512 4945 5053 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.2 214.9 234.4 Without AVX-512 23.3 209.3 228.7 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.0 93.2 95.5 Without AVX-512 48.6 93.2 95.5 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Age Gender Recognition Retail 0013 FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 10K 20K 30K 40K 50K SE +/- 43.30, N = 3 SE +/- 51.81, N = 3 45555.67 22781.70 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Vehicle Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 4 8 12 16 20 SE +/- 0.12, N = 3 SE +/- 0.16, N = 3 10.76 15.92 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 4.25 / MAX: 23.39 -mno-avx512f - MIN: 6.75 / MAX: 34.35 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5144 5265 5881 Without AVX-512 5062 5173 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 21.9 188.8 207.3 Without AVX-512 20.6 207.2 225.3 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.5 83.4 86.0 Without AVX-512 47.9 91.2 94.1 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Vehicle Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 160 320 480 640 800 SE +/- 8.04, N = 3 SE +/- 5.20, N = 3 742.90 502.28 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Weld Porosity Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 5.49 11.73 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 2.89 / MAX: 13.22 -mno-avx512f - MIN: 6.24 / MAX: 20.45 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5075 5159 5881 Without AVX-512 5008 5103 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.8 208.9 234.0 Without AVX-512 23.1 212.2 232.9 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 47.8 93.3 95.4 Without AVX-512 48.0 93.9 95.8 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Weld Porosity Detection FP16 - Device: CPU Default, AVX-512 Enabled Without AVX-512 300 600 900 1200 1500 SE +/- 1.06, N = 3 SE +/- 0.52, N = 3 1455.45 681.37 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
Result
OpenBenchmarking.org ms, Fewer Is Better OpenVINO 2022.2.dev Model: Weld Porosity Detection FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 5.44 11.81 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 2.88 / MAX: 13.32 -mno-avx512f - MIN: 6.9 / MAX: 19.2 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 4582 5163 5881 Without AVX-512 4982 5093 5881 OpenBenchmarking.org Megahertz, More Is Better OpenVINO 2022.2.dev CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.3 212.3 236.0 Without AVX-512 23.1 211.2 232.0 OpenBenchmarking.org Watts, Fewer Is Better OpenVINO 2022.2.dev CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 45.9 93.1 95.3 Without AVX-512 47.9 93.5 95.4 OpenBenchmarking.org Celsius, Fewer Is Better OpenVINO 2022.2.dev CPU Temperature Monitor 20 40 60 80 100
OpenBenchmarking.org FPS, More Is Better OpenVINO 2022.2.dev Model: Weld Porosity Detection FP16-INT8 - Device: CPU Default, AVX-512 Enabled Without AVX-512 600 1200 1800 2400 3000 SE +/- 4.82, N = 3 SE +/- 2.11, N = 3 2938.13 1353.84 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -fPIC -O3 -march=native -fsigned-char -ffunction-sections -fdata-sections -fno-strict-overflow -fwrapv -flto -shared
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 Intel oneAPI. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 0.1408 0.2816 0.4224 0.5632 0.704 SE +/- 0.007969, N = 15 SE +/- 0.006181, N = 5 0.417196 0.625942 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.34 -mno-avx512f - MIN: 0.58 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5127 5419 5881 Without AVX-512 5125 5367 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 20.1 124.9 214.3 Without AVX-512 23.9 140.4 221.2 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 48.4 70.8 87.5 Without AVX-512 48.1 80.2 93.6 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
Cpuminer-Opt Cpuminer-Opt is a fork of cpuminer-multi that carries a wide range of CPU performance optimizations for measuring the potential cryptocurrency mining performance of the CPU/processor with a wide variety of cryptocurrencies. The benchmark reports the hash speed for the CPU mining performance for the selected cryptocurrency. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 3.18 Algorithm: LBC, LBRY Credits Default, AVX-512 Enabled Without AVX-512 30K 60K 90K 120K 150K SE +/- 177.76, N = 3 SE +/- 40.41, N = 3 151720 76960 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -lcurl -lz -lpthread -lssl -lcrypto -lgmp kH/s Per Watt
OpenBenchmarking.org kH/s Per Watt, More Is Better Cpuminer-Opt 3.18 Algorithm: LBC, LBRY Credits Default, AVX-512 Enabled Without AVX-512 300 600 900 1200 1500 1370.92 579.30 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5208 5520 5881 Without AVX-512 5164 5459 5881 OpenBenchmarking.org Megahertz, More Is Better Cpuminer-Opt 3.18 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.3 110.7 206.7 Without AVX-512 23.5 132.8 181.2 OpenBenchmarking.org Watts, Fewer Is Better Cpuminer-Opt 3.18 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 40.5 66.6 83.9 Without AVX-512 41.9 73.0 86.5 OpenBenchmarking.org Celsius, Fewer Is Better Cpuminer-Opt 3.18 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 3.18 Algorithm: Quad SHA-256, Pyrite Default, AVX-512 Enabled Without AVX-512 70K 140K 210K 280K 350K SE +/- 1040.42, N = 3 SE +/- 177.04, N = 3 321503 192667 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -lcurl -lz -lpthread -lssl -lcrypto -lgmp kH/s Per Watt
OpenBenchmarking.org kH/s Per Watt, More Is Better Cpuminer-Opt 3.18 Algorithm: Quad SHA-256, Pyrite Default, AVX-512 Enabled Without AVX-512 600 1200 1800 2400 3000 3025.64 1788.77 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5224 5540 5881 Without AVX-512 5207 5522 5881 OpenBenchmarking.org Megahertz, More Is Better Cpuminer-Opt 3.18 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.3 106.3 178.0 Without AVX-512 23.2 107.7 185.5 OpenBenchmarking.org Watts, Fewer Is Better Cpuminer-Opt 3.18 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 41.4 64.7 81.5 Without AVX-512 44.0 67.7 80.0 OpenBenchmarking.org Celsius, Fewer Is Better Cpuminer-Opt 3.18 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org kH/s, More Is Better Cpuminer-Opt 3.18 Algorithm: Skeincoin Default, AVX-512 Enabled Without AVX-512 60K 120K 180K 240K 300K SE +/- 424.91, N = 3 SE +/- 1803.62, N = 3 284667 204320 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mno-avx512f 1. (CXX) g++ options: -O3 -march=native -lcurl -lz -lpthread -lssl -lcrypto -lgmp kH/s Per Watt
OpenBenchmarking.org kH/s Per Watt, More Is Better Cpuminer-Opt 3.18 Algorithm: Skeincoin Default, AVX-512 Enabled Without AVX-512 600 1200 1800 2400 3000 2696.44 1926.96 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5157 5535 5881 Without AVX-512 5189 5517 5881 OpenBenchmarking.org Megahertz, More Is Better Cpuminer-Opt 3.18 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.7 105.6 218.2 Without AVX-512 22.9 106.0 219.4 OpenBenchmarking.org Watts, Fewer Is Better Cpuminer-Opt 3.18 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 43.3 66.6 85.6 Without AVX-512 44.8 68.5 84.1 OpenBenchmarking.org Celsius, Fewer Is Better Cpuminer-Opt 3.18 CPU Temperature Monitor 20 40 60 80 100
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 Intel oneAPI. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 0.18 0.36 0.54 0.72 0.9 SE +/- 0.000107, N = 3 SE +/- 0.010456, N = 3 0.415940 0.800116 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.4 -mno-avx512f - MIN: 0.76 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5180 5381 5881 Without AVX-512 4992 5348 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.0 149.1 219.8 Without AVX-512 17.5 152.3 233.2 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 41.3 75.8 86.6 Without AVX-512 40.9 81.9 94.9 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
Embree Intel Embree is a collection of high-performance ray-tracing kernels for execution on CPUs and supporting instruction sets such as SSE, AVX, AVX2, and AVX-512. Embree also supports making use of the Intel SPMD Program Compiler (ISPC). Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.13 Binary: Pathtracer ISPC - Model: Crown Default, AVX-512 Enabled Without AVX-512 8 16 24 32 40 SE +/- 0.08, N = 3 SE +/- 0.11, N = 3 35.23 29.12 MIN: 34.75 / MAX: 36.08 MIN: 28.7 / MAX: 29.86 Frames Per Second Per Watt
OpenBenchmarking.org Frames Per Second Per Watt, More Is Better Embree 3.13 Binary: Pathtracer ISPC - Model: Crown Default, AVX-512 Enabled Without AVX-512 0.0423 0.0846 0.1269 0.1692 0.2115 0.188 0.153 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5146 5324 5881 Without AVX-512 5152 5300 5881 OpenBenchmarking.org Megahertz, More Is Better Embree 3.13 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.6 187.3 235.9 Without AVX-512 23.3 190.5 227.8 OpenBenchmarking.org Watts, Fewer Is Better Embree 3.13 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 40.3 83.2 90.9 Without AVX-512 40.0 84.9 91.0 OpenBenchmarking.org Celsius, Fewer Is Better Embree 3.13 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.13 Binary: Pathtracer ISPC - Model: Asian Dragon Default, AVX-512 Enabled Without AVX-512 8 16 24 32 40 SE +/- 0.08, N = 3 SE +/- 0.16, N = 3 35.12 30.39 MIN: 34.64 / MAX: 36.13 MIN: 29.85 / MAX: 30.94 Frames Per Second Per Watt
OpenBenchmarking.org Frames Per Second Per Watt, More Is Better Embree 3.13 Binary: Pathtracer ISPC - Model: Asian Dragon Default, AVX-512 Enabled Without AVX-512 0.0477 0.0954 0.1431 0.1908 0.2385 0.212 0.171 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5222 5375 5881 Without AVX-512 5197 5354 5881 OpenBenchmarking.org Megahertz, More Is Better Embree 3.13 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 22.6 166.0 203.7 Without AVX-512 24.3 177.3 210.0 OpenBenchmarking.org Watts, Fewer Is Better Embree 3.13 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 44.1 75.1 80.8 Without AVX-512 44.6 81.0 85.9 OpenBenchmarking.org Celsius, Fewer Is Better Embree 3.13 CPU Temperature Monitor 20 40 60 80 100
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.
Result
OpenBenchmarking.org FPS, More Is Better dav1d 1.0 Video Input: Summer Nature 4K Default, AVX-512 Enabled Without AVX-512 90 180 270 360 450 SE +/- 1.37, N = 5 SE +/- 3.22, N = 5 393.40 380.32 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -lm -mno-avx512f 1. (CC) gcc options: -O3 -march=native -pthread FPS Per Watt
OpenBenchmarking.org FPS Per Watt, More Is Better dav1d 1.0 Video Input: Summer Nature 4K Default, AVX-512 Enabled Without AVX-512 0.7018 1.4036 2.1054 2.8072 3.509 3.119 2.937 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5239 5493 5881 Without AVX-512 5258 5492 5881 OpenBenchmarking.org Megahertz, More Is Better dav1d 1.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.2 126.1 185.4 Without AVX-512 23.9 129.5 187.9 OpenBenchmarking.org Watts, Fewer Is Better dav1d 1.0 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 38.6 70.0 79.0 Without AVX-512 38.8 70.9 80.3 OpenBenchmarking.org Celsius, Fewer Is Better dav1d 1.0 CPU Temperature Monitor 20 40 60 80 100
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 Intel oneAPI. Learn more via the OpenBenchmarking.org test page.
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 2 4 6 8 10 SE +/- 0.00255, N = 7 SE +/- 0.00637, N = 7 5.32127 7.44936 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 5.26 -mno-avx512f - MIN: 7.29 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 4910 5475 5881 Without AVX-512 5318 5585 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 11.5 107.7 218.9 Without AVX-512 23.5 93.1 166.1 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 42.0 69.7 82.9 Without AVX-512 45.1 58.8 65.6 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
Result
OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU Default, AVX-512 Enabled Without AVX-512 0.2439 0.4878 0.7317 0.9756 1.2195 SE +/- 0.001581, N = 9 SE +/- 0.006665, N = 9 0.576432 1.083960 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.56 -mno-avx512f - MIN: 1.03 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5208 5613 5881 Without AVX-512 5157 5585 5881 OpenBenchmarking.org Megahertz, More Is Better oneDNN 2.6 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.1 102.3 213.6 Without AVX-512 22.8 108.1 230.7 OpenBenchmarking.org Watts, Fewer Is Better oneDNN 2.6 CPU Power Consumption Monitor 60 120 180 240 300 CPU Temp
Min Avg Max Default, AVX-512 Enabled 43.5 72.4 82.6 Without AVX-512 44.8 77.9 91.6 OpenBenchmarking.org Celsius, Fewer Is Better oneDNN 2.6 CPU Temperature Monitor 20 40 60 80 100
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.
Result
OpenBenchmarking.org FPS, More Is Better dav1d 1.0 Video Input: Summer Nature 1080p Default, AVX-512 Enabled Without AVX-512 300 600 900 1200 1500 SE +/- 1.72, N = 10 SE +/- 2.53, N = 10 1451.94 1412.27 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -lm -mno-avx512f 1. (CC) gcc options: -O3 -march=native -pthread FPS Per Watt
OpenBenchmarking.org FPS Per Watt, More Is Better dav1d 1.0 Video Input: Summer Nature 1080p Default, AVX-512 Enabled Without AVX-512 4 8 12 16 20 18.11 17.21 CPU Peak Freq (Highest CPU Core Frequency
Min Avg Max Default, AVX-512 Enabled 5282 5674 5881 Without AVX-512 5263 5686 5881 OpenBenchmarking.org Megahertz, More Is Better dav1d 1.0 CPU Peak Freq (Highest CPU Core Frequency) Monitor 1600 3200 4800 6400 8000 CPU Power Consumption
Min Avg Max Default, AVX-512 Enabled 23.2 80.2 178.7 Without AVX-512 23.6 82.1 180.6 OpenBenchmarking.org Watts, Fewer Is Better dav1d 1.0 CPU Power Consumption Monitor 50 100 150 200 250 CPU Temp
Min Avg Max Default, AVX-512 Enabled 41.3 64.5 73.3 Without AVX-512 41.3 65.1 74.4 OpenBenchmarking.org Celsius, Fewer Is Better dav1d 1.0 CPU Temperature Monitor 20 40 60 80 100
CPU Temperature Monitor OpenBenchmarking.org Celsius CPU Temperature Monitor Phoronix Test Suite System Monitoring Default, AVX-512 Enabled Without AVX-512 20 40 60 80 100 Min: 33.63 / Avg: 81.24 / Max: 96.25 Min: 36.13 / Avg: 84.12 / Max: 96.75
CPU Power Consumption Monitor OpenBenchmarking.org Watts CPU Power Consumption Monitor Phoronix Test Suite System Monitoring Default, AVX-512 Enabled Without AVX-512 40 80 120 160 200 Min: 11.53 / Avg: 158.06 / Max: 237.76 Min: 10.2 / Avg: 161.11 / Max: 235.88
CPU Peak Freq (Highest CPU Core Frequency) Monitor OpenBenchmarking.org Megahertz CPU Peak Freq (Highest CPU Core Frequency) Monitor Phoronix Test Suite System Monitoring Default, AVX-512 Enabled Without AVX-512 1000 2000 3000 4000 5000 Min: 4421 / Avg: 5372.19 / Max: 5881 Min: 4495 / Avg: 5334.93 / Max: 5881
Default, AVX-512 Enabled Kernel Notes: Transparent Huge Pages: madviseEnvironment Notes: CXXFLAGS="-O3 -march=native -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mprefer-vector-width=512" CFLAGS="-O3 -march=native -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -mprefer-vector-width=512"Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate performance (Boost: Enabled) - CPU Microcode: 0xa601203Python Notes: Python 3.10.4Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 17 September 2022 20:32 by user phoronix.
Without AVX-512 Processor: AMD Ryzen 9 7950X 16-Core @ 5.88GHz (16 Cores / 32 Threads), Motherboard: ASUS ROG CROSSHAIR X670E HERO (0604 BIOS), Chipset: AMD Device 14d8, Memory: 32GB, Disk: 2000GB Samsung SSD 980 PRO 2TB + 2000GB, Graphics: AMD Radeon RX 6800 XT 16GB (2575/1000MHz), Audio: AMD Navi 21 HDMI Audio, Monitor: ASUS VP28U, Network: Intel I225-V + Intel Wi-Fi 6 AX210/AX211/AX411
OS: Ubuntu 22.04, Kernel: 6.0.0-060000rc1daily20220820-generic (x86_64), Desktop: GNOME Shell 42.2, Display Server: X Server + Wayland, OpenGL: 4.6 Mesa 22.3.0-devel (git-4685385 2022-08-23 jammy-oibaf-ppa) (LLVM 14.0.6 DRM 3.48), Vulkan: 1.3.224, Compiler: GCC 12.0.1 20220319, File-System: ext4, Screen Resolution: 3840x2160
Kernel Notes: Transparent Huge Pages: madviseEnvironment Notes: CXXFLAGS="-O3 -march=native -mno-avx512f" CFLAGS="-O3 -march=native -mno-avx512f"Compiler Notes: --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-12-OcsLtf/gcc-12-12-20220319/debian/tmp-gcn/usr --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -vProcessor Notes: Scaling Governor: amd-pstate performance (Boost: Enabled) - CPU Microcode: 0xa601203Python Notes: Python 3.10.4Security Notes: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 17 September 2022 15:47 by user phoronix.