10900K sysbench onednn Intel Core i9-10900K testing with a Gigabyte Z490 AORUS MASTER (F3 BIOS) and Gigabyte AMD Radeon RX 5500/5500M / Pro 5500M 8GB on Ubuntu 20.10 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2103134-PTS-10900KSY85&sro&grs .
10900K sysbench onednn Processor Motherboard Chipset Memory Disk Graphics Audio Monitor Network OS Kernel Desktop Display Server OpenGL Vulkan Compiler File-System Screen Resolution 1 2 3 4 Intel Core i9-10900K @ 5.30GHz (10 Cores / 20 Threads) Gigabyte Z490 AORUS MASTER (F3 BIOS) Intel Comet Lake PCH 16GB Samsung SSD 970 EVO 250GB Gigabyte AMD Radeon RX 5500/5500M / Pro 5500M 8GB (1900/875MHz) Realtek ALC1220 ASUS MG28U Intel + Intel Wi-Fi 6 AX201 Ubuntu 20.10 5.11.0-051100rc2daily20210106-generic (x86_64) 20210105 GNOME Shell 3.38.1 X Server 1.20.9 4.6 Mesa 20.2.1 (LLVM 11.0.0) 1.2.131 GCC 10.2.0 ext4 3840x2160 OpenBenchmarking.org Kernel Details - Transparent Huge Pages: madvise 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: intel_pstate powersave - CPU Microcode: 0xe0 - Thermald 2.3 Security Details - itlb_multihit: KVM: Mitigation of VMX disabled + 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 Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
10900K sysbench onednn onednn: Recurrent Neural Network Inference - u8s8f32 - CPU onednn: Recurrent Neural Network Inference - f32 - CPU onednn: IP Shapes 3D - u8s8f32 - CPU onednn: Convolution Batch Shapes Auto - u8s8f32 - CPU onednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPU onednn: Matrix Multiply Batch Shapes Transformer - f32 - CPU onednn: Recurrent Neural Network Training - f32 - CPU onednn: Matrix Multiply Batch Shapes Transformer - u8s8f32 - CPU onednn: Recurrent Neural Network Training - u8s8f32 - CPU onednn: Deconvolution Batch shapes_3d - u8s8f32 - CPU sysbench: RAM / Memory onednn: Deconvolution Batch shapes_3d - f32 - CPU onednn: IP Shapes 3D - f32 - CPU onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPU onednn: IP Shapes 1D - u8s8f32 - CPU onednn: Recurrent Neural Network Training - bf16bf16bf16 - CPU onednn: Convolution Batch Shapes Auto - f32 - CPU onednn: IP Shapes 1D - f32 - CPU sysbench: CPU onednn: Deconvolution Batch shapes_1d - f32 - CPU 1 2 3 4 1779.52 1745.36 2.45003 17.9280 1748.03 3.94807 2843.83 1.91115 2847.38 3.76517 34129.11 4.82270 12.2150 1.49138 1.17248 2834.49 21.3621 3.30456 26017.89 6.67724 1732.40 1729.05 2.42769 17.6531 1742.26 3.92455 2837.76 1.93276 2837.57 3.76150 34243.57 4.81353 12.1785 1.49445 1.17465 2836.92 21.3339 3.30545 26029.63 6.85465 1754.38 1742.27 2.47206 17.6899 1725.15 3.89827 2854.28 1.91319 2818.43 3.77453 34315.67 4.82441 12.1638 1.49196 1.17258 2836.90 21.3277 3.30228 26028.92 6.62239 1744.70 1764.34 2.45468 17.7791 1749.05 3.89701 2818.99 1.93351 2844.56 3.74976 34146.51 4.83458 12.1712 1.49279 1.17424 2832.25 21.3298 3.30052 26019.11 6.59741 OpenBenchmarking.org
oneDNN Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 400 800 1200 1600 2000 SE +/- 8.77, N = 3 SE +/- 13.35, N = 3 SE +/- 19.63, N = 3 SE +/- 6.63, N = 3 1779.52 1732.40 1754.38 1744.70 MIN: 1736.71 MIN: 1680.02 MIN: 1691.38 MIN: 1681.33 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU 1 2 3 4 400 800 1200 1600 2000 SE +/- 15.68, N = 3 SE +/- 21.18, N = 3 SE +/- 21.30, N = 3 SE +/- 6.63, N = 3 1745.36 1729.05 1742.27 1764.34 MIN: 1705.83 MIN: 1683.97 MIN: 1683.83 MIN: 1689.05 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 0.5562 1.1124 1.6686 2.2248 2.781 SE +/- 0.03980, N = 3 SE +/- 0.01028, N = 3 SE +/- 0.01612, N = 3 SE +/- 0.01102, N = 3 2.45003 2.42769 2.47206 2.45468 MIN: 2.34 MIN: 2.31 MIN: 2.33 MIN: 2.31 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 4 8 12 16 20 SE +/- 0.07, N = 3 SE +/- 0.19, N = 3 SE +/- 0.13, N = 3 SE +/- 0.08, N = 3 17.93 17.65 17.69 17.78 MIN: 17.28 MIN: 17.18 MIN: 17.18 MIN: 17.16 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU 1 2 3 4 400 800 1200 1600 2000 SE +/- 14.77, N = 3 SE +/- 26.98, N = 3 SE +/- 2.77, N = 3 SE +/- 27.03, N = 3 1748.03 1742.26 1725.15 1749.05 MIN: 1687.97 MIN: 1671.61 MIN: 1683.55 MIN: 1667.38 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU 1 2 3 4 0.8883 1.7766 2.6649 3.5532 4.4415 SE +/- 0.01417, N = 3 SE +/- 0.00768, N = 3 SE +/- 0.00147, N = 3 SE +/- 0.00827, N = 3 3.94807 3.92455 3.89827 3.89701 MIN: 3.79 MIN: 3.77 MIN: 3.76 MIN: 3.72 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU 1 2 3 4 600 1200 1800 2400 3000 SE +/- 10.02, N = 3 SE +/- 5.58, N = 3 SE +/- 23.35, N = 3 SE +/- 16.37, N = 3 2843.83 2837.76 2854.28 2818.99 MIN: 2768.24 MIN: 2798.07 MIN: 2775.53 MIN: 2765.73 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 0.435 0.87 1.305 1.74 2.175 SE +/- 0.00175, N = 3 SE +/- 0.01672, N = 3 SE +/- 0.00284, N = 3 SE +/- 0.01806, N = 3 1.91115 1.93276 1.91319 1.93351 MIN: 1.88 MIN: 1.88 MIN: 1.88 MIN: 1.88 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 600 1200 1800 2400 3000 SE +/- 3.12, N = 3 SE +/- 15.45, N = 3 SE +/- 6.65, N = 3 SE +/- 5.94, N = 3 2847.38 2837.57 2818.43 2844.56 MIN: 2765.94 MIN: 2767.4 MIN: 2766.26 MIN: 2768.09 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 0.8493 1.6986 2.5479 3.3972 4.2465 SE +/- 0.00296, N = 3 SE +/- 0.00158, N = 3 SE +/- 0.00173, N = 3 SE +/- 0.00653, N = 3 3.76517 3.76150 3.77453 3.74976 MIN: 3.74 MIN: 3.74 MIN: 3.75 MIN: 3.72 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
Sysbench Test: RAM / Memory OpenBenchmarking.org MiB/sec, More Is Better Sysbench 1.0.20 Test: RAM / Memory 1 2 3 4 7K 14K 21K 28K 35K SE +/- 68.06, N = 3 SE +/- 99.86, N = 3 SE +/- 108.17, N = 3 SE +/- 148.70, N = 3 34129.11 34243.57 34315.67 34146.51 1. (CC) gcc options: -pthread -O2 -funroll-loops -rdynamic -ldl -laio -lm
oneDNN Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU 1 2 3 4 1.0878 2.1756 3.2634 4.3512 5.439 SE +/- 0.00884, N = 3 SE +/- 0.00575, N = 3 SE +/- 0.00815, N = 3 SE +/- 0.00473, N = 3 4.82270 4.81353 4.82441 4.83458 MIN: 4.79 MIN: 4.71 MIN: 4.75 MIN: 4.73 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU 1 2 3 4 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 12.22 12.18 12.16 12.17 MIN: 12.1 MIN: 12.08 MIN: 12.05 MIN: 12.06 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 0.3363 0.6726 1.0089 1.3452 1.6815 SE +/- 0.00072, N = 3 SE +/- 0.00187, N = 3 SE +/- 0.00095, N = 3 SE +/- 0.00086, N = 3 1.49138 1.49445 1.49196 1.49279 MIN: 1.48 MIN: 1.48 MIN: 1.48 MIN: 1.48 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU 1 2 3 4 0.2643 0.5286 0.7929 1.0572 1.3215 SE +/- 0.00221, N = 3 SE +/- 0.00194, N = 3 SE +/- 0.00090, N = 3 SE +/- 0.00107, N = 3 1.17248 1.17465 1.17258 1.17424 MIN: 1.16 MIN: 1.16 MIN: 1.16 MIN: 1.16 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU 1 2 3 4 600 1200 1800 2400 3000 SE +/- 12.24, N = 3 SE +/- 0.84, N = 3 SE +/- 6.62, N = 3 SE +/- 11.43, N = 3 2834.49 2836.92 2836.90 2832.25 MIN: 2770.1 MIN: 2770.47 MIN: 2764.69 MIN: 2770.76 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU 1 2 3 4 5 10 15 20 25 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 21.36 21.33 21.33 21.33 MIN: 21.29 MIN: 21.24 MIN: 21.25 MIN: 21.24 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
oneDNN Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU 1 2 3 4 0.7437 1.4874 2.2311 2.9748 3.7185 SE +/- 0.00272, N = 3 SE +/- 0.00232, N = 3 SE +/- 0.00348, N = 3 SE +/- 0.01635, N = 3 3.30456 3.30545 3.30228 3.30052 MIN: 3.03 MIN: 3.03 MIN: 3.02 MIN: 3.03 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
Sysbench Test: CPU OpenBenchmarking.org Events Per Second, More Is Better Sysbench 1.0.20 Test: CPU 1 2 3 4 6K 12K 18K 24K 30K SE +/- 2.59, N = 3 SE +/- 5.92, N = 3 SE +/- 1.44, N = 3 SE +/- 1.33, N = 3 26017.89 26029.63 26028.92 26019.11 1. (CC) gcc options: -pthread -O2 -funroll-loops -rdynamic -ldl -laio -lm
oneDNN Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 2.1.2 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU 1 2 3 4 2 4 6 8 10 SE +/- 0.09219, N = 15 SE +/- 0.13332, N = 15 SE +/- 0.09531, N = 15 SE +/- 0.11275, N = 15 6.67724 6.85465 6.62239 6.59741 MIN: 3.45 MIN: 3.45 MIN: 3.47 MIN: 3.47 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
Phoronix Test Suite v10.8.5