TR 2990WX 2020 AMD Ryzen Threadripper 2990WX 32-Core testing with a ASUS ROG ZENITH EXTREME (1701 BIOS) and Gigabyte AMD Radeon RX 470/480/570/570X/580/580X/590 4GB on Ubuntu 20.10 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2012260-HA-TR2990WX254&grs .
TR 2990WX 2020 Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server Display Driver OpenGL Vulkan Compiler File-System Screen Resolution 1 2 3 AMD Ryzen Threadripper 2990WX 32-Core @ 3.00GHz (32 Cores / 64 Threads) ASUS ROG ZENITH EXTREME (1701 BIOS) AMD 17h 32GB Samsung SSD 970 EVO 500GB + 250GB Western Digital WDS250G2X0C-00L350 Gigabyte AMD Radeon RX 470/480/570/570X/580/580X/590 4GB (1244/1750MHz) Realtek ALC1220 LG Ultra HD Intel I211 + Qualcomm Atheros QCA6174 802.11ac + Wilocity Wil6200 802.11ad Ubuntu 20.10 5.8.0-33-generic (x86_64) GNOME Shell 3.38.1 X Server 1.20.9 modesetting 1.20.9 4.6 Mesa 20.2.1 (LLVM 11.0.0) 1.2.131 GCC 10.2.0 ext4 1920x1080 OpenBenchmarking.org Compiler Details - --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none=/build/gcc-10-JvwpWM/gcc-10-10.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-10-JvwpWM/gcc-10-10.2.0/debian/tmp-gcn/usr,hsa --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v Processor Details - Scaling Governor: acpi-cpufreq ondemand (Boost: Enabled) - CPU Microcode: 0x800820d Graphics Details - GLAMOR Java Details - OpenJDK Runtime Environment (build 11.0.9.1+1-Ubuntu-0ubuntu1.20.10) Python Details - Python 3.8.6 Security Details - 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: disabled RSB filling + srbds: Not affected + tsx_async_abort: Not affected
TR 2990WX 2020 build-ffmpeg: Time To Compile compress-zstd: 19 stockfish: Total Time yquake2: OpenGL 1.x - 1920 x 1080 x265: Bosphorus 4K redis: SADD sunflow: Global Illumination + Image Synthesis redis: SET redis: GET onednn: Recurrent Neural Network Training - u8s8f32 - CPU onednn: IP Shapes 3D - f32 - CPU hpcc: Rand Ring Bandwidth hpcc: G-Ffte vkmark: 1280 x 1024 compress-lz4: 9 - Compression Speed hpcc: G-HPL ai-benchmark: Device Training Score node-web-tooling: asmfish: 1024 Hash Memory, 26 Depth onednn: Recurrent Neural Network Inference - u8s8f32 - CPU vkmark: 1920 x 1080 simdjson: LargeRand embree: Pathtracer - Crown embree: Pathtracer - Asian Dragon onednn: Recurrent Neural Network Training - f32 - CPU hpcc: G-Rand Access yquake2: OpenGL 3.x - 1920 x 1080 embree: Pathtracer ISPC - Crown redis: LPUSH ai-benchmark: Device AI Score lammps: 20k Atoms simdjson: PartialTweets x265: Bosphorus 1080p onednn: IP Shapes 3D - u8s8f32 - CPU embree: Pathtracer ISPC - Asian Dragon rav1e: 1 libplacebo: av1_grain_lap compress-lz4: 1 - Decompression Speed brl-cad: VGR Performance Metric hmmer: Pfam Database Search onednn: Convolution Batch Shapes Auto - f32 - CPU onednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU gromacs: Water Benchmark compress-lz4: 1 - Compression Speed hpcc: Max Ping Pong Bandwidth indigobench: CPU - Bedroom compress-lz4: 3 - Compression Speed compress-lz4: 9 - Decompression Speed basis: ETC1S onednn: Recurrent Neural Network Training - bf16bf16bf16 - CPU ai-benchmark: Device Inference Score rav1e: 10 espeak: Text-To-Speech Synthesis onednn: IP Shapes 1D - f32 - CPU astcenc: Thorough onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPU mafft: Multiple Sequence Alignment - LSU RNA onednn: Recurrent Neural Network Inference - f32 - CPU kvazaar: Bosphorus 4K - Slow kvazaar: Bosphorus 4K - Medium numpy: yquake2: Software CPU - 1920 x 1080 onednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPU onednn: IP Shapes 1D - u8s8f32 - CPU kvazaar: Bosphorus 1080p - Slow kvazaar: Bosphorus 4K - Ultra Fast onednn: Deconvolution Batch shapes_1d - f32 - CPU compress-lz4: 3 - Decompression Speed indigobench: CPU - Supercar basis: UASTC Level 2 + RDO Post-Processing kvazaar: Bosphorus 1080p - Ultra Fast crafty: Elapsed Time encode-opus: WAV To Opus Encode kvazaar: Bosphorus 1080p - Medium astcenc: Medium vkfft: basis: UASTC Level 0 clomp: Static OMP Speedup astcenc: Fast sqlite-speedtest: Timed Time - Size 1,000 phpbench: PHP Benchmark Suite waifu2x-ncnn: 2x - 3 - Yes astcenc: Exhaustive basis: UASTC Level 3 rav1e: 5 rav1e: 6 libplacebo: polar_nocompute kvazaar: Bosphorus 4K - Very Fast build-eigen: Time To Compile onednn: Deconvolution Batch shapes_3d - f32 - CPU onednn: Convolution Batch Shapes Auto - u8s8f32 - CPU encode-ape: WAV To APE basis: UASTC Level 2 coremark: CoreMark Size 666 - Iterations Per Second libplacebo: deband_heavy kvazaar: Bosphorus 1080p - Very Fast hpcc: Rand Ring Latency build2: Time To Compile encode-wavpack: WAV To WavPack vkresample: 2x - Single libplacebo: hdr_peakdetect simdjson: DistinctUserID simdjson: Kostya waifu2x-ncnn: 2x - 3 - No ncnn: Vulkan GPU - regnety_400m ncnn: Vulkan GPU - squeezenet_ssd ncnn: Vulkan GPU - yolov4-tiny ncnn: Vulkan GPU - resnet50 ncnn: Vulkan GPU - alexnet ncnn: Vulkan GPU - resnet18 ncnn: Vulkan GPU - vgg16 ncnn: Vulkan GPU - googlenet ncnn: Vulkan GPU - blazeface ncnn: Vulkan GPU - efficientnet-b0 ncnn: Vulkan GPU - mnasnet ncnn: Vulkan GPU - shufflenet-v2 ncnn: Vulkan GPU-v3-v3 - mobilenet-v3 ncnn: Vulkan GPU-v2-v2 - mobilenet-v2 ncnn: Vulkan GPU - mobilenet ncnn: CPU - regnety_400m ncnn: CPU - squeezenet_ssd ncnn: CPU - yolov4-tiny ncnn: CPU - resnet50 ncnn: CPU - alexnet ncnn: CPU - resnet18 ncnn: CPU - vgg16 ncnn: CPU - googlenet ncnn: CPU - blazeface ncnn: CPU - efficientnet-b0 ncnn: CPU - mnasnet ncnn: CPU - shufflenet-v2 ncnn: CPU-v3-v3 - mobilenet-v3 ncnn: CPU-v2-v2 - mobilenet-v2 ncnn: CPU - mobilenet redis: LPOP build-clash: Time To Compile embree: Pathtracer ISPC - Asian Dragon Obj embree: Pathtracer - Asian Dragon Obj onednn: Matrix Multiply Batch Shapes Transformer - f32 - CPU onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPU compress-zstd: 3 lammps: Rhodopsin Protein hpcc: EP-STREAM Triad hpcc: G-Ptrans hpcc: EP-DGEMM betsy: ETC2 RGB - Highest betsy: ETC1 - Highest 1 2 3 46.084 33.9 49821254 683.8 16.62 1892562.21 0.789 1675529.38 2201882.03 13991.1 11.4694 0.57120 10.04669 5850 46.60 54.59837 966 8.47 73804001 3802.52 4384 0.39 24.0193 21.9564 13861.2 0.02819 949.7 22.7259 1368828.88 2270 15.501 0.51 41.02 3.45370 21.3156 0.315 448.24 9506.2 286847 163.935 19.9734 1.46928 1.834 8518.04 10862.041 5.036 47.46 9106.9 47.602 13826.2 1304 2.859 30.847 6.27715 9.54 3.43727 12.153 3749.18 10.29 10.47 296.80 106.6 3808.57 2.12408 26.99 39.68 2.99559 9128.2 11.035 651.213 116.80 7329531 7.753 27.69 6.34 9412 7.595 51.7 5.22 67.207 577229 10.038 73.36 25.229 0.954 1.270 268.72 23.54 93.252 5.92604 25.1643 14.029 15.785 1146059.294691 192.54 60.69 1.53681 96.200 13.199 49.012 1762.98 0.52 0.44 2.052 101.01 41.28 50.25 72.06 33.56 55.89 103.41 41.08 9.82 19.14 15.64 14.92 14.58 15.59 37.30 102.93 41.89 50.19 79.53 37.72 59.91 102.38 47.87 6.83 20.01 13.60 15.28 14.44 15.67 37.24 2431710.35 463.220 17.6791 18.2418 1.65767 3.69657 3606.7 13.399 1.30966 4.00164 12.61727 12.697 10.754 30.918 42.0 55404772 619.9 15.26 1826296.85 0.731 1654163.60 2099666.52 13295.0 11.2085 0.58300 9.74083 5644 47.55 52.77753 998 8.25 74895187 3821.24 4267 0.39 24.6155 22.4966 13529.0 0.02798 972.2 22.2365 1340079.29 2317 15.418 0.50 40.36 3.49064 21.6674 0.320 441.37 9647.9 291013 162.561 20.1538 1.48549 1.858 8619.84 10831.974 4.994 47.12 9217.2 47.602 13783.3 1319 2.830 31.153 6.25011 9.49 3.44980 12.264 3782.36 10.38 10.56 298.31 107.1 3780.47 2.10893 27.13 39.67 3.01721 9191.9 10.963 647.924 116.09 7371374 7.795 27.84 6.32 9451 7.626 51.6 5.22 67.458 579360 10.055 73.14 25.217 0.951 1.273 268.36 23.57 92.981 5.93856 25.1026 14.010 15.790 1148000.044932 192.43 60.69 1.53453 96.340 13.215 48.996 1762.78 0.52 0.44 100.26 40.01 50.57 73.86 32.39 53.19 100.73 38.90 7.40 19.28 16.26 14.65 14.22 16.30 33.78 101.47 40.80 49.69 75.03 34.34 54.29 92.98 42.04 6.67 20.92 14.50 14.42 15.70 15.60 34.56 2058658.41 477.084 17.2357 19.2992 2.09567 3.76164 3798.9 12.913 1.30771 4.34106 8.80303 12.543 10.519 32.699 36.2 54984012 623.8 16.55 1974193.46 1577289.79 2091130.67 13669.2 10.9285 0.59385 9.67646 5640 45.94 54.30507 8.21 72630737 3711.96 4268 0.40 24.3636 22.2655 13753.9 0.02866 972.5 22.6453 1366164.67 15.187 0.51 41.06 3.51087 21.5622 0.318 446.38 9540.0 161.632 20.2543 1.48957 1.837 8508.64 10972.494 5.057 46.87 9130.5 48.164 13668.4 2.853 31.028 6.21623 9.45 3.46909 12.157 3783.40 10.32 10.53 295.79 106.2 3777.65 2.12611 26.92 39.39 3.01645 9188.3 11.036 647.089 116.68 7366701 7.774 27.78 6.35 9417 7.600 51.8 5.20 67.247 10.074 73.10 25.144 0.953 1.274 267.90 23.61 93.058 5.92300 25.1041 14.041 15.758 1146667.406114 192.22 60.79 1.53509 49.021 1762.55 0.52 0.44 101.27 39.14 48.23 70.92 35.93 56.35 100.17 38.32 7.09 20.49 13.77 15.75 15.95 14.97 34.39 103.89 39.13 49.25 73.65 35.69 64.49 95.61 42.95 7.65 18.97 14.17 15.30 15.21 15.78 34.91 1998586.16 485.485 17.3999 19.4014 2.00984 3.41699 3781.2 12.551 1.33820 4.17428 13.03417 12.553 10.529 OpenBenchmarking.org
Timed FFmpeg Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed FFmpeg Compilation 4.2.2 Time To Compile 1 2 3 10 20 30 40 50 SE +/- 0.18, N = 3 SE +/- 0.15, N = 3 SE +/- 0.18, N = 3 46.08 30.92 32.70
Zstd Compression Compression Level: 19 OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.5 Compression Level: 19 1 2 3 10 20 30 40 50 SE +/- 0.37, N = 15 SE +/- 0.34, N = 3 SE +/- 0.39, N = 3 33.9 42.0 36.2 1. (CC) gcc options: -O3 -pthread -lz -llzma
Stockfish Total Time OpenBenchmarking.org Nodes Per Second, More Is Better Stockfish 12 Total Time 1 2 3 12M 24M 36M 48M 60M SE +/- 796222.48, N = 3 SE +/- 577141.31, N = 3 SE +/- 749500.39, N = 3 49821254 55404772 54984012 1. (CXX) g++ options: -m64 -lpthread -fno-exceptions -std=c++17 -pedantic -O3 -msse -msse3 -mpopcnt -msse4.1 -mssse3 -msse2 -flto -flto=jobserver
yquake2 Renderer: OpenGL 1.x - Resolution: 1920 x 1080 OpenBenchmarking.org Frames Per Second, More Is Better yquake2 7.45 Renderer: OpenGL 1.x - Resolution: 1920 x 1080 1 2 3 150 300 450 600 750 SE +/- 4.44, N = 3 SE +/- 6.33, N = 8 SE +/- 5.54, N = 15 683.8 619.9 623.8 1. (CC) gcc options: -lm -ldl -rdynamic -shared -lSDL2 -O2 -pipe -fomit-frame-pointer -std=gnu99 -fno-strict-aliasing -fwrapv -fvisibility=hidden -MMD -mfpmath=sse -fPIC
x265 Video Input: Bosphorus 4K OpenBenchmarking.org Frames Per Second, More Is Better x265 3.4 Video Input: Bosphorus 4K 1 2 3 4 8 12 16 20 SE +/- 0.03, N = 3 SE +/- 0.05, N = 3 SE +/- 0.06, N = 3 16.62 15.26 16.55 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
Redis Test: SADD OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: SADD 1 2 3 400K 800K 1200K 1600K 2000K SE +/- 20792.00, N = 15 SE +/- 25792.40, N = 15 SE +/- 16528.72, N = 3 1892562.21 1826296.85 1974193.46 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
Sunflow Rendering System Global Illumination + Image Synthesis OpenBenchmarking.org Seconds, Fewer Is Better Sunflow Rendering System 0.07.2 Global Illumination + Image Synthesis 1 2 0.1775 0.355 0.5325 0.71 0.8875 SE +/- 0.004, N = 3 SE +/- 0.011, N = 15 0.789 0.731 MIN: 0.57 / MAX: 1.63 MIN: 0.49 / MAX: 1.57
Redis Test: SET OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: SET 1 2 3 400K 800K 1200K 1600K 2000K SE +/- 19958.53, N = 3 SE +/- 23714.82, N = 4 SE +/- 22941.42, N = 3 1675529.38 1654163.60 1577289.79 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
Redis Test: GET OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: GET 1 2 3 500K 1000K 1500K 2000K 2500K SE +/- 31653.62, N = 15 SE +/- 31097.88, N = 15 SE +/- 27660.05, N = 15 2201882.03 2099666.52 2091130.67 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
oneDNN Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU 1 2 3 3K 6K 9K 12K 15K SE +/- 48.11, N = 3 SE +/- 123.89, N = 10 SE +/- 122.96, N = 11 13991.1 13295.0 13669.2 MIN: 13812.5 MIN: 12213.4 MIN: 12754.3 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
oneDNN Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU 1 2 3 3 6 9 12 15 SE +/- 0.19, N = 12 SE +/- 0.15, N = 15 SE +/- 0.01, N = 3 11.47 11.21 10.93 MIN: 10.94 MIN: 10.67 MIN: 10.81 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
HPC Challenge Test / Class: Random Ring Bandwidth OpenBenchmarking.org GB/s, More Is Better HPC Challenge 1.5.0 Test / Class: Random Ring Bandwidth 1 2 3 0.1336 0.2672 0.4008 0.5344 0.668 SE +/- 0.00203, N = 3 SE +/- 0.01337, N = 3 SE +/- 0.00691, N = 3 0.57120 0.58300 0.59385 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
HPC Challenge Test / Class: G-Ffte OpenBenchmarking.org GFLOPS, More Is Better HPC Challenge 1.5.0 Test / Class: G-Ffte 1 2 3 3 6 9 12 15 SE +/- 0.04480, N = 3 SE +/- 0.17781, N = 3 SE +/- 0.14488, N = 3 10.04669 9.74083 9.67646 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
VKMark Resolution: 1280 x 1024 OpenBenchmarking.org VKMark Score, More Is Better VKMark 2020-05-21 Resolution: 1280 x 1024 1 2 3 1300 2600 3900 5200 6500 SE +/- 2.19, N = 3 SE +/- 5.90, N = 3 SE +/- 2.33, N = 3 5850 5644 5640 1. (CXX) g++ options: -ldl -pipe -std=c++14 -fPIC -MD -MQ -MF
LZ4 Compression Compression Level: 9 - Compression Speed OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 9 - Compression Speed 1 2 3 11 22 33 44 55 SE +/- 0.43, N = 15 SE +/- 0.77, N = 3 SE +/- 0.01, N = 3 46.60 47.55 45.94 1. (CC) gcc options: -O3
HPC Challenge Test / Class: G-HPL OpenBenchmarking.org GFLOPS, More Is Better HPC Challenge 1.5.0 Test / Class: G-HPL 1 2 3 12 24 36 48 60 SE +/- 0.14, N = 3 SE +/- 0.10, N = 3 SE +/- 0.24, N = 3 54.60 52.78 54.31 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
AI Benchmark Alpha Device Training Score OpenBenchmarking.org Score, More Is Better AI Benchmark Alpha 0.1.2 Device Training Score 1 2 200 400 600 800 1000 966 998
Node.js V8 Web Tooling Benchmark OpenBenchmarking.org runs/s, More Is Better Node.js V8 Web Tooling Benchmark 1 2 3 2 4 6 8 10 SE +/- 0.03, N = 3 SE +/- 0.12, N = 4 SE +/- 0.04, N = 3 8.47 8.25 8.21 1. Nodejs
v12.18.2
asmFish 1024 Hash Memory, 26 Depth OpenBenchmarking.org Nodes/second, More Is Better asmFish 2018-07-23 1024 Hash Memory, 26 Depth 1 2 3 16M 32M 48M 64M 80M SE +/- 738089.20, N = 3 SE +/- 126998.98, N = 3 SE +/- 633798.86, N = 3 73804001 74895187 72630737
oneDNN Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU 1 2 3 800 1600 2400 3200 4000 SE +/- 10.92, N = 3 SE +/- 46.19, N = 3 SE +/- 38.84, N = 8 3802.52 3821.24 3711.96 MIN: 3624.28 MIN: 3746.24 MIN: 3503.46 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
VKMark Resolution: 1920 x 1080 OpenBenchmarking.org VKMark Score, More Is Better VKMark 2020-05-21 Resolution: 1920 x 1080 1 2 3 900 1800 2700 3600 4500 SE +/- 2.96, N = 3 SE +/- 2.65, N = 3 SE +/- 2.85, N = 3 4384 4267 4268 1. (CXX) g++ options: -ldl -pipe -std=c++14 -fPIC -MD -MQ -MF
simdjson Throughput Test: LargeRandom OpenBenchmarking.org GB/s, More Is Better simdjson 0.7.1 Throughput Test: LargeRandom 1 2 3 0.09 0.18 0.27 0.36 0.45 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.39 0.39 0.40 1. (CXX) g++ options: -O3 -pthread
Embree Binary: Pathtracer - Model: Crown OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.9.0 Binary: Pathtracer - Model: Crown 1 2 3 6 12 18 24 30 SE +/- 0.33, N = 15 SE +/- 0.22, N = 3 SE +/- 0.31, N = 3 24.02 24.62 24.36 MIN: 19.33 / MAX: 25.81 MIN: 23.95 / MAX: 25.61 MIN: 23 / MAX: 25.59
Embree Binary: Pathtracer - Model: Asian Dragon OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.9.0 Binary: Pathtracer - Model: Asian Dragon 1 2 3 5 10 15 20 25 SE +/- 0.31, N = 15 SE +/- 0.25, N = 15 SE +/- 0.26, N = 15 21.96 22.50 22.27 MIN: 19.03 / MAX: 24.66 MIN: 20.59 / MAX: 25.14 MIN: 20.37 / MAX: 25.26
oneDNN Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU 1 2 3 3K 6K 9K 12K 15K SE +/- 110.93, N = 3 SE +/- 161.18, N = 3 SE +/- 132.58, N = 3 13861.2 13529.0 13753.9 MIN: 13490.9 MIN: 13190.1 MIN: 12649.5 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
HPC Challenge Test / Class: G-Random Access OpenBenchmarking.org GUP/s, More Is Better HPC Challenge 1.5.0 Test / Class: G-Random Access 1 2 3 0.0064 0.0128 0.0192 0.0256 0.032 SE +/- 0.00038, N = 3 SE +/- 0.00067, N = 3 SE +/- 0.00018, N = 3 0.02819 0.02798 0.02866 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
yquake2 Renderer: OpenGL 3.x - Resolution: 1920 x 1080 OpenBenchmarking.org Frames Per Second, More Is Better yquake2 7.45 Renderer: OpenGL 3.x - Resolution: 1920 x 1080 1 2 3 200 400 600 800 1000 SE +/- 12.14, N = 3 SE +/- 13.22, N = 3 SE +/- 10.37, N = 3 949.7 972.2 972.5 1. (CC) gcc options: -lm -ldl -rdynamic -shared -lSDL2 -O2 -pipe -fomit-frame-pointer -std=gnu99 -fno-strict-aliasing -fwrapv -fvisibility=hidden -MMD -mfpmath=sse -fPIC
Embree Binary: Pathtracer ISPC - Model: Crown OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.9.0 Binary: Pathtracer ISPC - Model: Crown 1 2 3 5 10 15 20 25 SE +/- 0.21, N = 3 SE +/- 0.38, N = 3 SE +/- 0.22, N = 3 22.73 22.24 22.65 MIN: 21.97 / MAX: 23.66 MIN: 20.82 / MAX: 23.48 MIN: 21.43 / MAX: 23.62
Redis Test: LPUSH OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: LPUSH 1 2 3 300K 600K 900K 1200K 1500K SE +/- 18400.26, N = 3 SE +/- 19080.84, N = 4 SE +/- 13973.47, N = 3 1368828.88 1340079.29 1366164.67 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
AI Benchmark Alpha Device AI Score OpenBenchmarking.org Score, More Is Better AI Benchmark Alpha 0.1.2 Device AI Score 1 2 500 1000 1500 2000 2500 2270 2317
LAMMPS Molecular Dynamics Simulator Model: 20k Atoms OpenBenchmarking.org ns/day, More Is Better LAMMPS Molecular Dynamics Simulator 29Oct2020 Model: 20k Atoms 1 2 3 4 8 12 16 20 SE +/- 0.21, N = 3 SE +/- 0.02, N = 3 SE +/- 0.21, N = 4 15.50 15.42 15.19 1. (CXX) g++ options: -O3 -pthread -lm
simdjson Throughput Test: PartialTweets OpenBenchmarking.org GB/s, More Is Better simdjson 0.7.1 Throughput Test: PartialTweets 1 2 3 0.1148 0.2296 0.3444 0.4592 0.574 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 0.51 0.50 0.51 1. (CXX) g++ options: -O3 -pthread
x265 Video Input: Bosphorus 1080p OpenBenchmarking.org Frames Per Second, More Is Better x265 3.4 Video Input: Bosphorus 1080p 1 2 3 9 18 27 36 45 SE +/- 0.23, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 41.02 40.36 41.06 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
oneDNN Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU 1 2 3 0.7899 1.5798 2.3697 3.1596 3.9495 SE +/- 0.02421, N = 3 SE +/- 0.04824, N = 3 SE +/- 0.03734, N = 3 3.45370 3.49064 3.51087 MIN: 1.93 MIN: 2.01 MIN: 2 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
Embree Binary: Pathtracer ISPC - Model: Asian Dragon OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.9.0 Binary: Pathtracer ISPC - Model: Asian Dragon 1 2 3 5 10 15 20 25 SE +/- 0.09, N = 3 SE +/- 0.26, N = 15 SE +/- 0.19, N = 15 21.32 21.67 21.56 MIN: 20.36 / MAX: 22.38 MIN: 19.41 / MAX: 23.55 MIN: 19.33 / MAX: 23.88
rav1e Speed: 1 OpenBenchmarking.org Frames Per Second, More Is Better rav1e 0.4 Alpha Speed: 1 1 2 3 0.072 0.144 0.216 0.288 0.36 SE +/- 0.001, N = 3 SE +/- 0.000, N = 3 SE +/- 0.001, N = 3 0.315 0.320 0.318
Libplacebo Test: av1_grain_lap OpenBenchmarking.org FPS, More Is Better Libplacebo 2.72.2 Test: av1_grain_lap 1 2 3 100 200 300 400 500 SE +/- 3.42, N = 3 SE +/- 3.25, N = 3 SE +/- 1.25, N = 3 448.24 441.37 446.38 1. (CXX) g++ options: -lm -lglslang -lHLSL -lOGLCompiler -lOSDependent -lSPIRV -lSPVRemapper -lSPIRV-Tools -lSPIRV-Tools-opt -lpthread -pthread -pipe -std=c++11 -fvisibility=hidden -fPIC -MD -MQ -MF
LZ4 Compression Compression Level: 1 - Decompression Speed OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 1 - Decompression Speed 1 2 3 2K 4K 6K 8K 10K SE +/- 34.06, N = 3 SE +/- 9.81, N = 3 SE +/- 47.16, N = 3 9506.2 9647.9 9540.0 1. (CC) gcc options: -O3
BRL-CAD VGR Performance Metric OpenBenchmarking.org VGR Performance Metric, More Is Better BRL-CAD 7.30.8 VGR Performance Metric 1 2 60K 120K 180K 240K 300K 286847 291013 1. (CXX) g++ options: -std=c++11 -pipe -fno-strict-aliasing -fno-common -fexceptions -ftemplate-depth-128 -m64 -ggdb3 -O3 -fipa-pta -fstrength-reduce -finline-functions -flto -pedantic -rdynamic -lSM -lICE -lXi -lGLU -lGL -lGLdispatch -lX11 -lXext -lXrender -lpthread -ldl -luuid -lm
Timed HMMer Search Pfam Database Search OpenBenchmarking.org Seconds, Fewer Is Better Timed HMMer Search 3.3.1 Pfam Database Search 1 2 3 40 80 120 160 200 SE +/- 1.51, N = 10 SE +/- 0.23, N = 3 SE +/- 0.19, N = 3 163.94 162.56 161.63 1. (CC) gcc options: -O3 -pthread -lhmmer -leasel -lm
oneDNN Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU 1 2 3 5 10 15 20 25 SE +/- 0.13, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 19.97 20.15 20.25 MIN: 14.86 MIN: 19.19 MIN: 19.16 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU 1 2 3 0.3352 0.6704 1.0056 1.3408 1.676 SE +/- 0.01123, N = 15 SE +/- 0.00191, N = 3 SE +/- 0.00013, N = 3 1.46928 1.48549 1.48957 MIN: 1.34 MIN: 1.45 MIN: 1.44 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
GROMACS Water Benchmark OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2020.3 Water Benchmark 1 2 3 0.4181 0.8362 1.2543 1.6724 2.0905 SE +/- 0.004, N = 3 SE +/- 0.004, N = 3 SE +/- 0.003, N = 2 1.834 1.858 1.837 1. (CXX) g++ options: -O3 -pthread -lrt -lpthread -lm
LZ4 Compression Compression Level: 1 - Compression Speed OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 1 - Compression Speed 1 2 3 2K 4K 6K 8K 10K SE +/- 65.90, N = 3 SE +/- 79.09, N = 3 SE +/- 51.76, N = 3 8518.04 8619.84 8508.64 1. (CC) gcc options: -O3
HPC Challenge Test / Class: Max Ping Pong Bandwidth OpenBenchmarking.org MB/s, More Is Better HPC Challenge 1.5.0 Test / Class: Max Ping Pong Bandwidth 1 2 3 2K 4K 6K 8K 10K SE +/- 250.74, N = 3 SE +/- 181.12, N = 3 SE +/- 206.43, N = 3 10862.04 10831.97 10972.49 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
IndigoBench Acceleration: CPU - Scene: Bedroom OpenBenchmarking.org M samples/s, More Is Better IndigoBench 4.4 Acceleration: CPU - Scene: Bedroom 1 2 3 1.1378 2.2756 3.4134 4.5512 5.689 SE +/- 0.025, N = 3 SE +/- 0.054, N = 12 SE +/- 0.019, N = 3 5.036 4.994 5.057
LZ4 Compression Compression Level: 3 - Compression Speed OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 3 - Compression Speed 1 2 3 11 22 33 44 55 SE +/- 0.54, N = 3 SE +/- 0.25, N = 3 SE +/- 0.29, N = 3 47.46 47.12 46.87 1. (CC) gcc options: -O3
LZ4 Compression Compression Level: 9 - Decompression Speed OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 9 - Decompression Speed 1 2 3 2K 4K 6K 8K 10K SE +/- 20.12, N = 15 SE +/- 4.46, N = 3 SE +/- 18.60, N = 3 9106.9 9217.2 9130.5 1. (CC) gcc options: -O3
Basis Universal Settings: ETC1S OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: ETC1S 1 2 3 11 22 33 44 55 SE +/- 0.22, N = 3 SE +/- 0.10, N = 3 SE +/- 0.24, N = 3 47.60 47.60 48.16 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
oneDNN Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU 1 2 3 3K 6K 9K 12K 15K SE +/- 141.29, N = 3 SE +/- 53.39, N = 3 SE +/- 128.38, N = 12 13826.2 13783.3 13668.4 MIN: 13450.2 MIN: 13567.9 MIN: 12362.1 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
AI Benchmark Alpha Device Inference Score OpenBenchmarking.org Score, More Is Better AI Benchmark Alpha 0.1.2 Device Inference Score 1 2 300 600 900 1200 1500 1304 1319
rav1e Speed: 10 OpenBenchmarking.org Frames Per Second, More Is Better rav1e 0.4 Alpha Speed: 10 1 2 3 0.6433 1.2866 1.9299 2.5732 3.2165 SE +/- 0.006, N = 3 SE +/- 0.018, N = 3 SE +/- 0.012, N = 3 2.859 2.830 2.853
eSpeak-NG Speech Engine Text-To-Speech Synthesis OpenBenchmarking.org Seconds, Fewer Is Better eSpeak-NG Speech Engine 20200907 Text-To-Speech Synthesis 1 2 3 7 14 21 28 35 SE +/- 0.06, N = 4 SE +/- 0.17, N = 4 SE +/- 0.07, N = 4 30.85 31.15 31.03 1. (CC) gcc options: -O2 -std=c99
oneDNN Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU 1 2 3 2 4 6 8 10 SE +/- 0.05437, N = 3 SE +/- 0.08570, N = 3 SE +/- 0.09143, N = 3 6.27715 6.25011 6.21623 MIN: 5.63 MIN: 5.72 MIN: 5.67 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
ASTC Encoder Preset: Thorough OpenBenchmarking.org Seconds, Fewer Is Better ASTC Encoder 2.0 Preset: Thorough 1 2 3 3 6 9 12 15 SE +/- 0.08, N = 3 SE +/- 0.05, N = 3 SE +/- 0.01, N = 3 9.54 9.49 9.45 1. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU 1 2 3 0.7805 1.561 2.3415 3.122 3.9025 SE +/- 0.00942, N = 3 SE +/- 0.00155, N = 3 SE +/- 0.02456, N = 3 3.43727 3.44980 3.46909 MIN: 3.34 MIN: 3.36 MIN: 3.35 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
Timed MAFFT Alignment Multiple Sequence Alignment - LSU RNA OpenBenchmarking.org Seconds, Fewer Is Better Timed MAFFT Alignment 7.471 Multiple Sequence Alignment - LSU RNA 1 2 3 3 6 9 12 15 SE +/- 0.04, N = 3 SE +/- 0.15, N = 6 SE +/- 0.12, N = 3 12.15 12.26 12.16 1. (CC) gcc options: -std=c99 -O3 -lm -lpthread
oneDNN Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU 1 2 3 800 1600 2400 3200 4000 SE +/- 16.86, N = 3 SE +/- 8.72, N = 3 SE +/- 8.76, N = 3 3749.18 3782.36 3783.40 MIN: 3631.66 MIN: 3756.39 MIN: 3758.06 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
Kvazaar Video Input: Bosphorus 4K - Video Preset: Slow OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 4K - Video Preset: Slow 1 2 3 3 6 9 12 15 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 10.29 10.38 10.32 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Kvazaar Video Input: Bosphorus 4K - Video Preset: Medium OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 4K - Video Preset: Medium 1 2 3 3 6 9 12 15 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 10.47 10.56 10.53 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Numpy Benchmark OpenBenchmarking.org Score, More Is Better Numpy Benchmark 1 2 3 60 120 180 240 300 SE +/- 0.54, N = 3 SE +/- 0.29, N = 3 SE +/- 0.70, N = 3 296.80 298.31 295.79
yquake2 Renderer: Software CPU - Resolution: 1920 x 1080 OpenBenchmarking.org Frames Per Second, More Is Better yquake2 7.45 Renderer: Software CPU - Resolution: 1920 x 1080 1 2 3 20 40 60 80 100 SE +/- 1.02, N = 3 SE +/- 1.01, N = 3 SE +/- 0.93, N = 3 106.6 107.1 106.2 1. (CC) gcc options: -lm -ldl -rdynamic -shared -lSDL2 -O2 -pipe -fomit-frame-pointer -std=gnu99 -fno-strict-aliasing -fwrapv -fvisibility=hidden -MMD -mfpmath=sse -fPIC
oneDNN Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU 1 2 3 800 1600 2400 3200 4000 SE +/- 3.43, N = 3 SE +/- 13.99, N = 3 SE +/- 10.39, N = 3 3808.57 3780.47 3777.65 MIN: 3705.89 MIN: 3731.84 MIN: 3644.35 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
oneDNN Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU 1 2 3 0.4784 0.9568 1.4352 1.9136 2.392 SE +/- 0.00488, N = 3 SE +/- 0.00369, N = 3 SE +/- 0.00796, N = 3 2.12408 2.10893 2.12611 MIN: 2.04 MIN: 2.03 MIN: 2.05 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
Kvazaar Video Input: Bosphorus 1080p - Video Preset: Slow OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 1080p - Video Preset: Slow 1 2 3 6 12 18 24 30 SE +/- 0.04, N = 3 SE +/- 0.07, N = 3 SE +/- 0.18, N = 3 26.99 27.13 26.92 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Kvazaar Video Input: Bosphorus 4K - Video Preset: Ultra Fast OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 4K - Video Preset: Ultra Fast 1 2 3 9 18 27 36 45 SE +/- 0.28, N = 3 SE +/- 0.15, N = 3 SE +/- 0.03, N = 3 39.68 39.67 39.39 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU 1 2 3 0.6789 1.3578 2.0367 2.7156 3.3945 SE +/- 0.02534, N = 3 SE +/- 0.01913, N = 3 SE +/- 0.01010, N = 3 2.99559 3.01721 3.01645 MIN: 2.82 MIN: 2.84 MIN: 2.85 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
LZ4 Compression Compression Level: 3 - Decompression Speed OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 3 - Decompression Speed 1 2 3 2K 4K 6K 8K 10K SE +/- 64.14, N = 3 SE +/- 19.51, N = 3 SE +/- 57.74, N = 3 9128.2 9191.9 9188.3 1. (CC) gcc options: -O3
IndigoBench Acceleration: CPU - Scene: Supercar OpenBenchmarking.org M samples/s, More Is Better IndigoBench 4.4 Acceleration: CPU - Scene: Supercar 1 2 3 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.03, N = 3 11.04 10.96 11.04
Basis Universal Settings: UASTC Level 2 + RDO Post-Processing OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 2 + RDO Post-Processing 1 2 3 140 280 420 560 700 SE +/- 1.61, N = 3 SE +/- 0.39, N = 3 SE +/- 0.07, N = 3 651.21 647.92 647.09 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
Kvazaar Video Input: Bosphorus 1080p - Video Preset: Ultra Fast OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 1080p - Video Preset: Ultra Fast 1 2 3 30 60 90 120 150 SE +/- 0.15, N = 3 SE +/- 0.66, N = 3 SE +/- 0.26, N = 3 116.80 116.09 116.68 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Crafty Elapsed Time OpenBenchmarking.org Nodes Per Second, More Is Better Crafty 25.2 Elapsed Time 1 2 3 1.6M 3.2M 4.8M 6.4M 8M SE +/- 7194.80, N = 3 SE +/- 6808.24, N = 3 SE +/- 5337.04, N = 3 7329531 7371374 7366701 1. (CC) gcc options: -pthread -lstdc++ -fprofile-use -lm
Opus Codec Encoding WAV To Opus Encode OpenBenchmarking.org Seconds, Fewer Is Better Opus Codec Encoding 1.3.1 WAV To Opus Encode 1 2 3 2 4 6 8 10 SE +/- 0.012, N = 5 SE +/- 0.012, N = 5 SE +/- 0.021, N = 5 7.753 7.795 7.774 1. (CXX) g++ options: -fvisibility=hidden -logg -lm
Kvazaar Video Input: Bosphorus 1080p - Video Preset: Medium OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 1080p - Video Preset: Medium 1 2 3 7 14 21 28 35 SE +/- 0.07, N = 3 SE +/- 0.06, N = 3 SE +/- 0.15, N = 3 27.69 27.84 27.78 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
ASTC Encoder Preset: Medium OpenBenchmarking.org Seconds, Fewer Is Better ASTC Encoder 2.0 Preset: Medium 1 2 3 2 4 6 8 10 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 6.34 6.32 6.35 1. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
VkFFT OpenBenchmarking.org Benchmark Score, More Is Better VkFFT 1.1.1 1 2 3 2K 4K 6K 8K 10K SE +/- 13.38, N = 3 SE +/- 49.17, N = 3 SE +/- 6.74, N = 3 9412 9451 9417 1. (CXX) g++ options: -O3 -pthread
Basis Universal Settings: UASTC Level 0 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 0 1 2 3 2 4 6 8 10 SE +/- 0.057, N = 3 SE +/- 0.021, N = 3 SE +/- 0.019, N = 3 7.595 7.626 7.600 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
CLOMP Static OMP Speedup OpenBenchmarking.org Speedup, More Is Better CLOMP 1.2 Static OMP Speedup 1 2 3 12 24 36 48 60 SE +/- 0.57, N = 3 SE +/- 0.20, N = 2 SE +/- 0.68, N = 3 51.7 51.6 51.8 1. (CC) gcc options: -fopenmp -O3 -lm
ASTC Encoder Preset: Fast OpenBenchmarking.org Seconds, Fewer Is Better ASTC Encoder 2.0 Preset: Fast 1 2 3 1.1745 2.349 3.5235 4.698 5.8725 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 5.22 5.22 5.20 1. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
SQLite Speedtest Timed Time - Size 1,000 OpenBenchmarking.org Seconds, Fewer Is Better SQLite Speedtest 3.30 Timed Time - Size 1,000 1 2 3 15 30 45 60 75 SE +/- 0.47, N = 3 SE +/- 0.25, N = 3 SE +/- 0.13, N = 3 67.21 67.46 67.25 1. (CC) gcc options: -O2 -ldl -lz -lpthread
PHPBench PHP Benchmark Suite OpenBenchmarking.org Score, More Is Better PHPBench 0.8.1 PHP Benchmark Suite 1 2 120K 240K 360K 480K 600K SE +/- 2547.76, N = 3 SE +/- 145.08, N = 3 577229 579360
Waifu2x-NCNN Vulkan Scale: 2x - Denoise: 3 - TAA: Yes OpenBenchmarking.org Seconds, Fewer Is Better Waifu2x-NCNN Vulkan 20200818 Scale: 2x - Denoise: 3 - TAA: Yes 1 2 3 3 6 9 12 15 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 10.04 10.06 10.07
ASTC Encoder Preset: Exhaustive OpenBenchmarking.org Seconds, Fewer Is Better ASTC Encoder 2.0 Preset: Exhaustive 1 2 3 16 32 48 64 80 SE +/- 0.09, N = 3 SE +/- 0.12, N = 3 SE +/- 0.10, N = 3 73.36 73.14 73.10 1. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -mfpmath=sse -mavx2 -mpopcnt -lpthread
Basis Universal Settings: UASTC Level 3 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 3 1 2 3 6 12 18 24 30 SE +/- 0.06, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 25.23 25.22 25.14 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
rav1e Speed: 5 OpenBenchmarking.org Frames Per Second, More Is Better rav1e 0.4 Alpha Speed: 5 1 2 3 0.2147 0.4294 0.6441 0.8588 1.0735 SE +/- 0.002, N = 3 SE +/- 0.002, N = 3 SE +/- 0.004, N = 3 0.954 0.951 0.953
rav1e Speed: 6 OpenBenchmarking.org Frames Per Second, More Is Better rav1e 0.4 Alpha Speed: 6 1 2 3 0.2867 0.5734 0.8601 1.1468 1.4335 SE +/- 0.001, N = 3 SE +/- 0.002, N = 3 SE +/- 0.002, N = 3 1.270 1.273 1.274
Libplacebo Test: polar_nocompute OpenBenchmarking.org FPS, More Is Better Libplacebo 2.72.2 Test: polar_nocompute 1 2 3 60 120 180 240 300 SE +/- 0.83, N = 3 SE +/- 0.53, N = 3 SE +/- 0.38, N = 3 268.72 268.36 267.90 1. (CXX) g++ options: -lm -lglslang -lHLSL -lOGLCompiler -lOSDependent -lSPIRV -lSPVRemapper -lSPIRV-Tools -lSPIRV-Tools-opt -lpthread -pthread -pipe -std=c++11 -fvisibility=hidden -fPIC -MD -MQ -MF
Kvazaar Video Input: Bosphorus 4K - Video Preset: Very Fast OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 4K - Video Preset: Very Fast 1 2 3 6 12 18 24 30 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 23.54 23.57 23.61 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
Timed Eigen Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed Eigen Compilation 3.3.9 Time To Compile 1 2 3 20 40 60 80 100 SE +/- 0.06, N = 3 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 93.25 92.98 93.06
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU 1 2 3 1.3362 2.6724 4.0086 5.3448 6.681 SE +/- 0.01022, N = 3 SE +/- 0.00211, N = 3 SE +/- 0.00240, N = 3 5.92604 5.93856 5.92300 MIN: 5.7 MIN: 5.71 MIN: 5.5 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
oneDNN Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU 1 2 3 6 12 18 24 30 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 25.16 25.10 25.10 MIN: 24.12 MIN: 24.22 MIN: 23.92 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
Monkey Audio Encoding WAV To APE OpenBenchmarking.org Seconds, Fewer Is Better Monkey Audio Encoding 3.99.6 WAV To APE 1 2 3 4 8 12 16 20 SE +/- 0.06, N = 5 SE +/- 0.06, N = 5 SE +/- 0.06, N = 5 14.03 14.01 14.04 1. (CXX) g++ options: -O3 -pedantic -rdynamic -lrt
Basis Universal Settings: UASTC Level 2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 2 1 2 3 4 8 12 16 20 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 15.79 15.79 15.76 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
Coremark CoreMark Size 666 - Iterations Per Second OpenBenchmarking.org Iterations/Sec, More Is Better Coremark 1.0 CoreMark Size 666 - Iterations Per Second 1 2 3 200K 400K 600K 800K 1000K SE +/- 2731.02, N = 3 SE +/- 686.93, N = 3 SE +/- 1713.53, N = 3 1146059.29 1148000.04 1146667.41 1. (CC) gcc options: -O2 -lrt" -lrt
Libplacebo Test: deband_heavy OpenBenchmarking.org FPS, More Is Better Libplacebo 2.72.2 Test: deband_heavy 1 2 3 40 80 120 160 200 SE +/- 0.36, N = 3 SE +/- 0.37, N = 3 SE +/- 0.29, N = 3 192.54 192.43 192.22 1. (CXX) g++ options: -lm -lglslang -lHLSL -lOGLCompiler -lOSDependent -lSPIRV -lSPVRemapper -lSPIRV-Tools -lSPIRV-Tools-opt -lpthread -pthread -pipe -std=c++11 -fvisibility=hidden -fPIC -MD -MQ -MF
Kvazaar Video Input: Bosphorus 1080p - Video Preset: Very Fast OpenBenchmarking.org Frames Per Second, More Is Better Kvazaar 2.0 Video Input: Bosphorus 1080p - Video Preset: Very Fast 1 2 3 14 28 42 56 70 SE +/- 0.21, N = 3 SE +/- 0.20, N = 3 SE +/- 0.26, N = 3 60.69 60.69 60.79 1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O2 -lpthread -lm -lrt
HPC Challenge Test / Class: Random Ring Latency OpenBenchmarking.org usecs, Fewer Is Better HPC Challenge 1.5.0 Test / Class: Random Ring Latency 1 2 3 0.3458 0.6916 1.0374 1.3832 1.729 SE +/- 0.00435, N = 3 SE +/- 0.00474, N = 3 SE +/- 0.00358, N = 3 1.53681 1.53453 1.53509 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
Build2 Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Build2 0.13 Time To Compile 1 2 20 40 60 80 100 SE +/- 0.24, N = 3 SE +/- 0.21, N = 3 96.20 96.34
WavPack Audio Encoding WAV To WavPack OpenBenchmarking.org Seconds, Fewer Is Better WavPack Audio Encoding 5.3 WAV To WavPack 1 2 3 6 9 12 15 SE +/- 0.01, N = 5 SE +/- 0.01, N = 5 13.20 13.22 1. (CXX) g++ options: -rdynamic
VkResample Upscale: 2x - Precision: Single OpenBenchmarking.org ms, Fewer Is Better VkResample 1.0 Upscale: 2x - Precision: Single 1 2 3 11 22 33 44 55 SE +/- 0.04, N = 3 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 49.01 49.00 49.02 1. (CXX) g++ options: -O3 -pthread
Libplacebo Test: hdr_peakdetect OpenBenchmarking.org FPS, More Is Better Libplacebo 2.72.2 Test: hdr_peakdetect 1 2 3 400 800 1200 1600 2000 SE +/- 0.03, N = 3 SE +/- 0.18, N = 3 SE +/- 0.10, N = 3 1762.98 1762.78 1762.55 1. (CXX) g++ options: -lm -lglslang -lHLSL -lOGLCompiler -lOSDependent -lSPIRV -lSPVRemapper -lSPIRV-Tools -lSPIRV-Tools-opt -lpthread -pthread -pipe -std=c++11 -fvisibility=hidden -fPIC -MD -MQ -MF
simdjson Throughput Test: DistinctUserID OpenBenchmarking.org GB/s, More Is Better simdjson 0.7.1 Throughput Test: DistinctUserID 1 2 3 0.117 0.234 0.351 0.468 0.585 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.52 0.52 0.52 1. (CXX) g++ options: -O3 -pthread
simdjson Throughput Test: Kostya OpenBenchmarking.org GB/s, More Is Better simdjson 0.7.1 Throughput Test: Kostya 1 2 3 0.099 0.198 0.297 0.396 0.495 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.44 0.44 0.44 1. (CXX) g++ options: -O3 -pthread
Waifu2x-NCNN Vulkan Scale: 2x - Denoise: 3 - TAA: No OpenBenchmarking.org Seconds, Fewer Is Better Waifu2x-NCNN Vulkan 20200818 Scale: 2x - Denoise: 3 - TAA: No 1 0.4617 0.9234 1.3851 1.8468 2.3085 2.052
NCNN Target: Vulkan GPU - Model: regnety_400m OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: regnety_400m 1 2 3 20 40 60 80 100 SE +/- 1.78, N = 12 SE +/- 2.13, N = 12 SE +/- 2.24, N = 9 101.01 100.26 101.27 MIN: 90.72 / MAX: 1587.31 MIN: 90.99 / MAX: 1833.25 MIN: 89.99 / MAX: 2458.9 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: squeezenet_ssd OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: squeezenet_ssd 1 2 3 9 18 27 36 45 SE +/- 1.34, N = 12 SE +/- 1.39, N = 12 SE +/- 1.29, N = 9 41.28 40.01 39.14 MIN: 31.56 / MAX: 448.33 MIN: 31.52 / MAX: 514.49 MIN: 31.19 / MAX: 435.95 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: yolov4-tiny OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: yolov4-tiny 1 2 3 11 22 33 44 55 SE +/- 1.15, N = 12 SE +/- 1.28, N = 12 SE +/- 1.42, N = 9 50.25 50.57 48.23 MIN: 39.88 / MAX: 213.87 MIN: 39.57 / MAX: 214.93 MIN: 39.54 / MAX: 230.79 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: resnet50 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: resnet50 1 2 3 16 32 48 64 80 SE +/- 3.10, N = 12 SE +/- 5.68, N = 12 SE +/- 6.02, N = 9 72.06 73.86 70.92 MIN: 39.39 / MAX: 562.21 MIN: 40.4 / MAX: 546.35 MIN: 39.2 / MAX: 557.13 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: alexnet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: alexnet 1 2 3 8 16 24 32 40 SE +/- 1.15, N = 12 SE +/- 1.09, N = 12 SE +/- 1.85, N = 9 33.56 32.39 35.93 MIN: 15.36 / MAX: 104.32 MIN: 17.56 / MAX: 91.67 MIN: 17.55 / MAX: 96.42 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: resnet18 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: resnet18 1 2 3 13 26 39 52 65 SE +/- 3.22, N = 12 SE +/- 5.95, N = 12 SE +/- 3.64, N = 9 55.89 53.19 56.35 MIN: 23.8 / MAX: 226.22 MIN: 21.74 / MAX: 222.51 MIN: 22.77 / MAX: 219.04 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: vgg16 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: vgg16 1 2 3 20 40 60 80 100 SE +/- 2.14, N = 12 SE +/- 2.43, N = 12 SE +/- 2.35, N = 9 103.41 100.73 100.17 MIN: 63 / MAX: 216.59 MIN: 65.03 / MAX: 221.49 MIN: 63.35 / MAX: 242.13 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: googlenet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: googlenet 1 2 3 9 18 27 36 45 SE +/- 2.33, N = 12 SE +/- 1.77, N = 12 SE +/- 1.78, N = 9 41.08 38.90 38.32 MIN: 28.69 / MAX: 513.63 MIN: 28.91 / MAX: 532.2 MIN: 28.17 / MAX: 505.55 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: blazeface OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: blazeface 1 2 3 3 6 9 12 15 SE +/- 1.94, N = 12 SE +/- 0.76, N = 12 SE +/- 0.51, N = 9 9.82 7.40 7.09 MIN: 6.19 / MAX: 229.58 MIN: 6.14 / MAX: 215.25 MIN: 6.15 / MAX: 204.55 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: efficientnet-b0 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: efficientnet-b0 1 2 3 5 10 15 20 25 SE +/- 0.25, N = 12 SE +/- 0.70, N = 12 SE +/- 0.66, N = 9 19.14 19.28 20.49 MIN: 17.51 / MAX: 352.05 MIN: 16.96 / MAX: 430.49 MIN: 17.45 / MAX: 438.66 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: mnasnet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: mnasnet 1 2 3 4 8 12 16 20 SE +/- 0.70, N = 11 SE +/- 1.06, N = 12 SE +/- 0.35, N = 9 15.64 16.26 13.77 MIN: 12.3 / MAX: 352.98 MIN: 12.28 / MAX: 390.46 MIN: 12.4 / MAX: 378.94 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: shufflenet-v2 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: shufflenet-v2 1 2 3 4 8 12 16 20 SE +/- 0.51, N = 12 SE +/- 0.14, N = 11 SE +/- 0.93, N = 9 14.92 14.65 15.75 MIN: 13.15 / MAX: 283.76 MIN: 13.43 / MAX: 114.64 MIN: 12.97 / MAX: 295.26 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU-v3-v3 - Model: mobilenet-v3 1 2 3 4 8 12 16 20 SE +/- 0.55, N = 12 SE +/- 0.34, N = 12 SE +/- 1.16, N = 9 14.58 14.22 15.95 MIN: 12.63 / MAX: 382.37 MIN: 12.46 / MAX: 387.96 MIN: 12.47 / MAX: 359.98 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU-v2-v2 - Model: mobilenet-v2 1 2 3 4 8 12 16 20 SE +/- 0.35, N = 12 SE +/- 0.63, N = 12 SE +/- 0.24, N = 9 15.59 16.30 14.97 MIN: 12.98 / MAX: 359.55 MIN: 13.31 / MAX: 389.24 MIN: 13.4 / MAX: 357.21 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: Vulkan GPU - Model: mobilenet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: Vulkan GPU - Model: mobilenet 1 2 3 9 18 27 36 45 SE +/- 1.11, N = 12 SE +/- 0.72, N = 12 SE +/- 1.11, N = 9 37.30 33.78 34.39 MIN: 29.33 / MAX: 427.64 MIN: 29.45 / MAX: 408.23 MIN: 29.33 / MAX: 412.06 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: regnety_400m OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: regnety_400m 1 2 3 20 40 60 80 100 SE +/- 1.12, N = 12 SE +/- 1.29, N = 12 SE +/- 3.17, N = 12 102.93 101.47 103.89 MIN: 90.68 / MAX: 1519.43 MIN: 90.53 / MAX: 2458.71 MIN: 90.75 / MAX: 3380.17 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: squeezenet_ssd OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: squeezenet_ssd 1 2 3 10 20 30 40 50 SE +/- 1.53, N = 12 SE +/- 1.74, N = 12 SE +/- 1.90, N = 12 41.89 40.80 39.13 MIN: 31.38 / MAX: 429.4 MIN: 31.26 / MAX: 438.66 MIN: 32.02 / MAX: 459.23 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: yolov4-tiny OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: yolov4-tiny 1 2 3 11 22 33 44 55 SE +/- 0.68, N = 12 SE +/- 1.04, N = 12 SE +/- 0.86, N = 12 50.19 49.69 49.25 MIN: 39.36 / MAX: 224.81 MIN: 39.24 / MAX: 224.21 MIN: 39.85 / MAX: 267.24 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: resnet50 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: resnet50 1 2 3 20 40 60 80 100 SE +/- 5.46, N = 12 SE +/- 2.78, N = 12 SE +/- 4.79, N = 12 79.53 75.03 73.65 MIN: 38.42 / MAX: 565.15 MIN: 38.58 / MAX: 559.97 MIN: 39.25 / MAX: 638.31 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: alexnet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: alexnet 1 2 3 9 18 27 36 45 SE +/- 1.89, N = 12 SE +/- 2.63, N = 12 SE +/- 1.62, N = 12 37.72 34.34 35.69 MIN: 15.66 / MAX: 106.73 MIN: 15.18 / MAX: 103.77 MIN: 16.27 / MAX: 106.47 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: resnet18 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: resnet18 1 2 3 14 28 42 56 70 SE +/- 4.97, N = 12 SE +/- 5.28, N = 12 SE +/- 4.57, N = 12 59.91 54.29 64.49 MIN: 23.15 / MAX: 227.75 MIN: 22 / MAX: 230.1 MIN: 21.25 / MAX: 228.14 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: vgg16 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: vgg16 1 2 3 20 40 60 80 100 SE +/- 3.95, N = 12 SE +/- 1.61, N = 12 SE +/- 2.05, N = 12 102.38 92.98 95.61 MIN: 62.04 / MAX: 220.23 MIN: 61.58 / MAX: 223.7 MIN: 64.1 / MAX: 227.91 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: googlenet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: googlenet 1 2 3 11 22 33 44 55 SE +/- 2.92, N = 12 SE +/- 2.39, N = 12 SE +/- 3.13, N = 12 47.87 42.04 42.95 MIN: 27.73 / MAX: 542.74 MIN: 28.93 / MAX: 517.65 MIN: 27.93 / MAX: 530.03 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: blazeface OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: blazeface 1 2 3 2 4 6 8 10 SE +/- 0.16, N = 12 SE +/- 0.10, N = 12 SE +/- 0.60, N = 12 6.83 6.67 7.65 MIN: 6.11 / MAX: 191.61 MIN: 6.12 / MAX: 175.66 MIN: 6.15 / MAX: 211.21 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: efficientnet-b0 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: efficientnet-b0 1 2 3 5 10 15 20 25 SE +/- 0.67, N = 12 SE +/- 1.79, N = 12 SE +/- 0.37, N = 12 20.01 20.92 18.97 MIN: 17.38 / MAX: 456.25 MIN: 17.12 / MAX: 465.23 MIN: 17.3 / MAX: 414.55 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: mnasnet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: mnasnet 1 2 3 4 8 12 16 20 SE +/- 0.16, N = 12 SE +/- 0.64, N = 12 SE +/- 0.31, N = 12 13.60 14.50 14.17 MIN: 12.42 / MAX: 189.74 MIN: 12.55 / MAX: 348.4 MIN: 12.78 / MAX: 347.15 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: shufflenet-v2 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: shufflenet-v2 1 2 3 4 8 12 16 20 SE +/- 0.85, N = 12 SE +/- 0.15, N = 12 SE +/- 0.64, N = 12 15.28 14.42 15.30 MIN: 13.6 / MAX: 309.56 MIN: 13.2 / MAX: 104.86 MIN: 12.9 / MAX: 306.7 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU-v3-v3 - Model: mobilenet-v3 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU-v3-v3 - Model: mobilenet-v3 1 2 3 4 8 12 16 20 SE +/- 0.50, N = 12 SE +/- 1.39, N = 12 SE +/- 1.31, N = 12 14.44 15.70 15.21 MIN: 12.49 / MAX: 358.11 MIN: 12.6 / MAX: 382.98 MIN: 12.52 / MAX: 378.12 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU-v2-v2 - Model: mobilenet-v2 OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU-v2-v2 - Model: mobilenet-v2 1 2 3 4 8 12 16 20 SE +/- 0.44, N = 12 SE +/- 0.34, N = 12 SE +/- 0.73, N = 12 15.67 15.60 15.78 MIN: 13.15 / MAX: 357.88 MIN: 13.33 / MAX: 383.7 MIN: 13.2 / MAX: 388.68 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
NCNN Target: CPU - Model: mobilenet OpenBenchmarking.org ms, Fewer Is Better NCNN 20201218 Target: CPU - Model: mobilenet 1 2 3 9 18 27 36 45 SE +/- 1.43, N = 12 SE +/- 0.55, N = 12 SE +/- 0.99, N = 12 37.24 34.56 34.91 MIN: 29.37 / MAX: 419.38 MIN: 30.05 / MAX: 404.5 MIN: 29.38 / MAX: 419.39 1. (CXX) g++ options: -O3 -rdynamic -lgomp -lpthread
Redis Test: LPOP OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: LPOP 1 2 3 500K 1000K 1500K 2000K 2500K SE +/- 44065.81, N = 15 SE +/- 132410.87, N = 12 SE +/- 139985.77, N = 12 2431710.35 2058658.41 1998586.16 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
Timed Clash Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed Clash Compilation Time To Compile 1 2 3 110 220 330 440 550 SE +/- 13.67, N = 9 SE +/- 4.96, N = 9 SE +/- 0.58, N = 3 463.22 477.08 485.49
Embree Binary: Pathtracer ISPC - Model: Asian Dragon Obj OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.9.0 Binary: Pathtracer ISPC - Model: Asian Dragon Obj 1 2 3 4 8 12 16 20 SE +/- 0.39, N = 12 SE +/- 0.39, N = 12 SE +/- 0.30, N = 15 17.68 17.24 17.40 MIN: 15.66 / MAX: 20.6 MIN: 14.95 / MAX: 20.29 MIN: 15.12 / MAX: 19.79
Embree Binary: Pathtracer - Model: Asian Dragon Obj OpenBenchmarking.org Frames Per Second, More Is Better Embree 3.9.0 Binary: Pathtracer - Model: Asian Dragon Obj 1 2 3 5 10 15 20 25 SE +/- 0.25, N = 15 SE +/- 0.37, N = 15 SE +/- 0.44, N = 12 18.24 19.30 19.40 MIN: 16.42 / MAX: 21.1 MIN: 16.68 / MAX: 22.31 MIN: 16.59 / MAX: 22.2
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU 1 2 3 0.4715 0.943 1.4145 1.886 2.3575 SE +/- 0.05725, N = 15 SE +/- 0.05017, N = 15 SE +/- 0.03417, N = 15 1.65767 2.09567 2.00984 MIN: 1.11 MIN: 1.4 MIN: 1.39 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.0 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU 1 2 3 0.8464 1.6928 2.5392 3.3856 4.232 SE +/- 0.07703, N = 15 SE +/- 0.07933, N = 15 SE +/- 0.01146, N = 3 3.69657 3.76164 3.41699 MIN: 3.25 MIN: 3.26 MIN: 3.22 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread
Zstd Compression Compression Level: 3 OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.5 Compression Level: 3 1 2 3 800 1600 2400 3200 4000 SE +/- 88.27, N = 12 SE +/- 129.73, N = 12 SE +/- 107.86, N = 15 3606.7 3798.9 3781.2 1. (CC) gcc options: -O3 -pthread -lz -llzma
LAMMPS Molecular Dynamics Simulator Model: Rhodopsin Protein OpenBenchmarking.org ns/day, More Is Better LAMMPS Molecular Dynamics Simulator 29Oct2020 Model: Rhodopsin Protein 1 2 3 3 6 9 12 15 SE +/- 0.36, N = 12 SE +/- 0.30, N = 15 SE +/- 0.15, N = 3 13.40 12.91 12.55 1. (CXX) g++ options: -O3 -pthread -lm
HPC Challenge Test / Class: EP-STREAM Triad OpenBenchmarking.org GB/s, More Is Better HPC Challenge 1.5.0 Test / Class: EP-STREAM Triad 1 2 3 0.3011 0.6022 0.9033 1.2044 1.5055 SE +/- 0.05676, N = 3 SE +/- 0.03714, N = 3 SE +/- 0.07867, N = 3 1.30966 1.30771 1.33820 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
HPC Challenge Test / Class: G-Ptrans OpenBenchmarking.org GB/s, More Is Better HPC Challenge 1.5.0 Test / Class: G-Ptrans 1 2 3 0.9767 1.9534 2.9301 3.9068 4.8835 SE +/- 0.25728, N = 3 SE +/- 0.03755, N = 3 SE +/- 0.08946, N = 3 4.00164 4.34106 4.17428 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
HPC Challenge Test / Class: EP-DGEMM OpenBenchmarking.org GFLOPS, More Is Better HPC Challenge 1.5.0 Test / Class: EP-DGEMM 1 2 3 3 6 9 12 15 SE +/- 0.86774, N = 3 SE +/- 0.14154, N = 3 SE +/- 0.34471, N = 3 12.61727 8.80303 13.03417 1. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops 2. ATLAS + Open MPI 4.0.3
Betsy GPU Compressor Codec: ETC2 RGB - Quality: Highest OpenBenchmarking.org Seconds, Fewer Is Better Betsy GPU Compressor 1.1 Beta Codec: ETC2 RGB - Quality: Highest 1 2 3 3 6 9 12 15 SE +/- 0.24, N = 15 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 12.70 12.54 12.55 1. (CXX) g++ options: -O3 -O2 -lpthread -ldl
Betsy GPU Compressor Codec: ETC1 - Quality: Highest OpenBenchmarking.org Seconds, Fewer Is Better Betsy GPU Compressor 1.1 Beta Codec: ETC1 - Quality: Highest 1 2 3 3 6 9 12 15 SE +/- 0.23, N = 15 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 10.75 10.52 10.53 1. (CXX) g++ options: -O3 -O2 -lpthread -ldl
Phoronix Test Suite v10.8.4