2 x AMD EPYC 75F3 32-Core testing with a ASRockRack ROME2D16-2T (P3.30 BIOS) and ASPEED on Ubuntu 21.10 via the Phoronix Test Suite.
Compare your own system(s) to this result file with the
Phoronix Test Suite by running the command:
phoronix-test-suite benchmark 2204097-NE-EPYC75F3N46 epyc-75f3-new - Phoronix Test Suite epyc-75f3-new 2 x AMD EPYC 75F3 32-Core testing with a ASRockRack ROME2D16-2T (P3.30 BIOS) and ASPEED on Ubuntu 21.10 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2204097-NE-EPYC75F3N46&grs&export=pdf&rdt&rro .
epyc-75f3-new Processor Motherboard Chipset Memory Disk Graphics Audio Network OS Kernel Desktop Display Server Vulkan Compiler File-System Screen Resolution A AA B C 2 x AMD EPYC 75F3 32-Core @ 2.95GHz (64 Cores / 128 Threads) ASRockRack ROME2D16-2T (P3.30 BIOS) AMD Starship/Matisse 128GB 1000GB Western Digital WD_BLACK SN850 1TB ASPEED AMD Starship/Matisse 2 x Intel 10G X550T Ubuntu 21.10 5.17.0-051700rc4daily20220219-generic (x86_64) GNOME Shell 40.5 X Server 1.1.182 GCC 11.2.0 ext4 1024x768 OpenBenchmarking.org Kernel Details - Transparent Huge Pages: madvise Compiler Details - --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-cet --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-link-serialization=2 --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none=/build/gcc-11-ZPT0kp/gcc-11-11.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-11-ZPT0kp/gcc-11-11.2.0/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-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v Processor Details - Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001114 Python Details - Python 3.9.7 Security Details - itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: 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 Full AMD retpoline IBPB: conditional IBRS_FW STIBP: always-on RSB filling + srbds: Not affected + tsx_async_abort: Not affected
epyc-75f3-new onnx: super-resolution-10 - CPU - Standard onnx: GPT-2 - CPU - Standard onednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU perf-bench: Epoll Wait onnx: yolov4 - CPU - Standard onnx: bertsquad-12 - CPU - Standard onednn: Recurrent Neural Network Training - bf16bf16bf16 - CPU onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPU onnx: fcn-resnet101-11 - CPU - Standard onednn: Deconvolution Batch shapes_3d - f32 - CPU avifenc: 6 onnx: ArcFace ResNet-100 - CPU - Standard onnx: GPT-2 - CPU - Parallel avifenc: 2 avifenc: 10, Lossless avifenc: 6, Lossless perf-bench: Futex Lock-Pi onednn: IP Shapes 3D - f32 - CPU perf-bench: Sched Pipe onednn: Convolution Batch Shapes Auto - f32 - CPU perf-bench: Memcpy 1MB perf-bench: Memset 1MB onnx: super-resolution-10 - CPU - Parallel avifenc: 0 onednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPU perf-bench: Futex Hash onednn: Deconvolution Batch shapes_1d - f32 - CPU onnx: ArcFace ResNet-100 - CPU - Parallel onnx: fcn-resnet101-11 - CPU - Parallel perf-bench: Syscall Basic onnx: yolov4 - CPU - Parallel onnx: bertsquad-12 - CPU - Parallel onednn: Matrix Multiply Batch Shapes Transformer - f32 - CPU onednn: Recurrent Neural Network Inference - u8s8f32 - CPU onednn: Recurrent Neural Network Training - u8s8f32 - CPU onednn: Recurrent Neural Network Inference - f32 - CPU onednn: Recurrent Neural Network Training - f32 - CPU onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPU onednn: Convolution Batch Shapes Auto - u8s8f32 - CPU onednn: IP Shapes 3D - u8s8f32 - CPU onednn: IP Shapes 1D - u8s8f32 - CPU onednn: IP Shapes 1D - f32 - CPU A AA B C 2352 0.878553 2.12767 4.278 38.784 5.010 7.215 77 2.63912 345990 0.705380 42.868389 62.455072 69.460 2976792 11.2361 17051621 11.51214 1624.246 3819.45 1693.03 3874.18 1.68391 0.897288 0.747084 4.05569 2.94158 4626 13234 22.9061 2299 355 540 4276.7 0.904662 170 2.10407 4.093 1194 1308 38.728 5.229 7.389 77 2.54811 351368 0.708812 42.829829 63.415414 4320 69.411 1711.61 2976712 11.1395 1225 123 17052202 434 594 17.5534 2016.94 4416.22 1803.51 4530.24 1.07515 1.24701 0.815632 4.25865 3.7792 7762 13106 27.3436 2572 291 643 4464.12 0.803426 158 1.94632 3.986 1244 1225 37.566 5.197 7.514 80 2.56643 346133 0.684799 44.147916 63.789341 4242 70.327 1734.92 2979134 11.2174 1235 123 16941879 436 592 12.4965 1801.29 4623.03 1810.37 4336.25 1.853 0.928893 0.757826 4.77292 3.18091 4488 9094 29.1013 2062 316 539 3862.48 0.829217 152 1.99639 4.215 1279 1242 39.573 4.971 7.185 77 2.54404 339461 0.689583 42.926485 63.185354 4233 70.646 1721.12 2945494 11.1942 1229 124 16924594 436 592 21.6718 1696.6 4390.23 1713.39 4224.87 1.23753 0.78443 0.74301 2.55861 3.85444 OpenBenchmarking.org
ONNX Runtime Model: super-resolution-10 - Device: CPU - Executor: Standard OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: super-resolution-10 - Device: CPU - Executor: Standard C B AA 1700 3400 5100 6800 8500 4488 7762 4626 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
ONNX Runtime Model: GPT-2 - Device: CPU - Executor: Standard OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: GPT-2 - Device: CPU - Executor: Standard C B AA 3K 6K 9K 12K 15K 9094 13106 13234 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU C B AA 7 14 21 28 35 29.10 27.34 22.91 MIN: 20.22 MIN: 19.51 MIN: 16.94 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
perf-bench Benchmark: Epoll Wait OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Epoll Wait C B AA A 600 1200 1800 2400 3000 SE +/- 23.81, N = 12 2062 2572 2299 2352 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
ONNX Runtime Model: yolov4 - Device: CPU - Executor: Standard OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: yolov4 - Device: CPU - Executor: Standard C B AA 80 160 240 320 400 316 291 355 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
ONNX Runtime Model: bertsquad-12 - Device: CPU - Executor: Standard OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: bertsquad-12 - Device: CPU - Executor: Standard C B AA 140 280 420 560 700 539 643 540 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
oneDNN Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU C B AA 1000 2000 3000 4000 5000 3862.48 4464.12 4276.70 MIN: 3242.51 MIN: 4162.92 MIN: 3959.8 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU C B AA A 0.2035 0.407 0.6105 0.814 1.0175 SE +/- 0.011448, N = 12 0.829217 0.803426 0.904662 0.878553 MIN: 0.6 MIN: 0.57 MIN: 0.68 MIN: 0.65 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
ONNX Runtime Model: fcn-resnet101-11 - Device: CPU - Executor: Standard OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: fcn-resnet101-11 - Device: CPU - Executor: Standard C B AA 40 80 120 160 200 152 158 170 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU C B AA A 0.4787 0.9574 1.4361 1.9148 2.3935 SE +/- 0.02508, N = 3 1.99639 1.94632 2.10407 2.12767 MIN: 1.73 MIN: 1.74 MIN: 1.79 MIN: 1.72 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
libavif avifenc Encoder Speed: 6 OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 0.10 Encoder Speed: 6 C B AA A 0.9626 1.9252 2.8878 3.8504 4.813 SE +/- 0.011, N = 3 4.215 3.986 4.093 4.278 1. (CXX) g++ options: -O3 -fPIC -lm
ONNX Runtime Model: ArcFace ResNet-100 - Device: CPU - Executor: Standard OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: ArcFace ResNet-100 - Device: CPU - Executor: Standard C B AA 300 600 900 1200 1500 1279 1244 1194 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
ONNX Runtime Model: GPT-2 - Device: CPU - Executor: Parallel OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: GPT-2 - Device: CPU - Executor: Parallel C B AA 300 600 900 1200 1500 1242 1225 1308 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
libavif avifenc Encoder Speed: 2 OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 0.10 Encoder Speed: 2 C B AA A 9 18 27 36 45 SE +/- 0.11, N = 3 39.57 37.57 38.73 38.78 1. (CXX) g++ options: -O3 -fPIC -lm
libavif avifenc Encoder Speed: 10, Lossless OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 0.10 Encoder Speed: 10, Lossless C B AA A 1.1765 2.353 3.5295 4.706 5.8825 SE +/- 0.030, N = 3 4.971 5.197 5.229 5.010 1. (CXX) g++ options: -O3 -fPIC -lm
libavif avifenc Encoder Speed: 6, Lossless OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 0.10 Encoder Speed: 6, Lossless C B AA A 2 4 6 8 10 SE +/- 0.066, N = 7 7.185 7.514 7.389 7.215 1. (CXX) g++ options: -O3 -fPIC -lm
perf-bench Benchmark: Futex Lock-Pi OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Futex Lock-Pi C B AA A 20 40 60 80 100 SE +/- 0.58, N = 3 77 80 77 77 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
oneDNN Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU C B AA A 0.5938 1.1876 1.7814 2.3752 2.969 SE +/- 0.01578, N = 3 2.54404 2.56643 2.54811 2.63912 MIN: 2.05 MIN: 2.05 MIN: 2.06 MIN: 2.08 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
perf-bench Benchmark: Sched Pipe OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Sched Pipe C B AA A 80K 160K 240K 320K 400K SE +/- 3279.19, N = 3 339461 346133 351368 345990 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
oneDNN Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU C B AA A 0.1595 0.319 0.4785 0.638 0.7975 SE +/- 0.006925, N = 3 0.689583 0.684799 0.708812 0.705380 MIN: 0.64 MIN: 0.64 MIN: 0.67 MIN: 0.64 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
perf-bench Benchmark: Memcpy 1MB OpenBenchmarking.org GB/sec, More Is Better perf-bench Benchmark: Memcpy 1MB C B AA A 10 20 30 40 50 SE +/- 0.17, N = 3 42.93 44.15 42.83 42.87 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
perf-bench Benchmark: Memset 1MB OpenBenchmarking.org GB/sec, More Is Better perf-bench Benchmark: Memset 1MB C B AA A 14 28 42 56 70 SE +/- 0.65, N = 5 63.19 63.79 63.42 62.46 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
ONNX Runtime Model: super-resolution-10 - Device: CPU - Executor: Parallel OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: super-resolution-10 - Device: CPU - Executor: Parallel C B AA 900 1800 2700 3600 4500 4233 4242 4320 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
libavif avifenc Encoder Speed: 0 OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 0.10 Encoder Speed: 0 C B AA A 16 32 48 64 80 SE +/- 0.35, N = 3 70.65 70.33 69.41 69.46 1. (CXX) g++ options: -O3 -fPIC -lm
oneDNN Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU C B AA 400 800 1200 1600 2000 1721.12 1734.92 1711.61 MIN: 1603.97 MIN: 1463.54 MIN: 1495.43 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
perf-bench Benchmark: Futex Hash OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Futex Hash C B AA A 600K 1200K 1800K 2400K 3000K SE +/- 6620.62, N = 3 2945494 2979134 2976712 2976792 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU C B AA A 3 6 9 12 15 SE +/- 0.09, N = 3 11.19 11.22 11.14 11.24 MIN: 8.84 MIN: 9 MIN: 9 MIN: 8.6 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
ONNX Runtime Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallel OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: ArcFace ResNet-100 - Device: CPU - Executor: Parallel C B AA 300 600 900 1200 1500 1229 1235 1225 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
ONNX Runtime Model: fcn-resnet101-11 - Device: CPU - Executor: Parallel OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: fcn-resnet101-11 - Device: CPU - Executor: Parallel C B AA 30 60 90 120 150 124 123 123 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
perf-bench Benchmark: Syscall Basic OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Syscall Basic C B AA A 4M 8M 12M 16M 20M SE +/- 7517.61, N = 3 16924594 16941879 17052202 17051621 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -lpthread -lrt -lm -ldl -lelf -lcrypto -lslang -lz -lnuma
ONNX Runtime Model: yolov4 - Device: CPU - Executor: Parallel OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: yolov4 - Device: CPU - Executor: Parallel C B AA 90 180 270 360 450 436 436 434 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
ONNX Runtime Model: bertsquad-12 - Device: CPU - Executor: Parallel OpenBenchmarking.org Inferences Per Minute, More Is Better ONNX Runtime 1.11 Model: bertsquad-12 - Device: CPU - Executor: Parallel C B AA 130 260 390 520 650 592 592 594 1. (CXX) g++ options: -ffunction-sections -fdata-sections -march=native -mtune=native -O3 -flto -fno-fat-lto-objects -ldl -lrt
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU C B AA A 5 10 15 20 25 SE +/- 1.25, N = 15 21.67 12.50 17.55 11.51 MIN: 15.66 MIN: 8.99 MIN: 12.56 MIN: 5.15 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU C B AA A 400 800 1200 1600 2000 SE +/- 88.05, N = 12 1696.60 1801.29 2016.94 1624.25 MIN: 1543.98 MIN: 1534.31 MIN: 1858.23 MIN: 827.28 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU C B AA A 1000 2000 3000 4000 5000 SE +/- 242.60, N = 12 4390.23 4623.03 4416.22 3819.45 MIN: 3742.53 MIN: 4242.04 MIN: 4047.45 MIN: 1342.54 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU C B AA A 400 800 1200 1600 2000 SE +/- 64.86, N = 12 1713.39 1810.37 1803.51 1693.03 MIN: 1504.13 MIN: 1535.85 MIN: 1650.67 MIN: 1036.24 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU C B AA A 1000 2000 3000 4000 5000 SE +/- 239.78, N = 14 4224.87 4336.25 4530.24 3874.18 MIN: 3842.95 MIN: 3938.66 MIN: 3953.07 MIN: 1544.77 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU C B AA A 0.4169 0.8338 1.2507 1.6676 2.0845 SE +/- 0.09292, N = 15 1.23753 1.85300 1.07515 1.68391 MIN: 1.09 MIN: 1.67 MIN: 0.92 MIN: 0.87 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU C B AA A 0.2806 0.5612 0.8418 1.1224 1.403 SE +/- 0.048482, N = 15 0.784430 0.928893 1.247010 0.897288 MIN: 0.54 MIN: 0.54 MIN: 0.56 MIN: 0.51 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU C B AA A 0.1835 0.367 0.5505 0.734 0.9175 SE +/- 0.011604, N = 15 0.743010 0.757826 0.815632 0.747084 MIN: 0.61 MIN: 0.62 MIN: 0.62 MIN: 0.57 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU C B AA A 1.0739 2.1478 3.2217 4.2956 5.3695 SE +/- 0.14420, N = 15 2.55861 4.77292 4.25865 4.05569 MIN: 1.85 MIN: 2.83 MIN: 2.63 MIN: 2.07 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
oneDNN Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.6 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU C B AA A 0.8672 1.7344 2.6016 3.4688 4.336 SE +/- 0.19680, N = 15 3.85444 3.18091 3.77920 2.94158 MIN: 2.65 MIN: 2.24 MIN: 2.55 MIN: 1.53 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -std=c++11 -pie -ldl -lpthread
Phoronix Test Suite v10.8.4