EPYC 7F32 Last AMD EPYC 7F32 8-Core testing with a Supermicro H11DSi-NT v2.00 (2.1 BIOS) and llvmpipe on Ubuntu 20.04 via the Phoronix Test Suite. Run 1: Processor: AMD EPYC 7F32 8-Core @ 3.70GHz (8 Cores / 16 Threads), Motherboard: Supermicro H11DSi-NT v2.00 (2.1 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 280GB INTEL SSDPE21D280GA, Graphics: llvmpipe, Monitor: VE228 OS: Ubuntu 20.04, Kernel: 5.8.0-050800rc6daily20200721-generic (x86_64) 20200720, Desktop: GNOME Shell 3.36.1, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 3.3 Mesa 20.0.4 (LLVM 9.0.1 128 bits), Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080 Run 2: Processor: AMD EPYC 7F32 8-Core @ 3.70GHz (8 Cores / 16 Threads), Motherboard: Supermicro H11DSi-NT v2.00 (2.1 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 280GB INTEL SSDPE21D280GA, Graphics: llvmpipe, Monitor: VE228 OS: Ubuntu 20.04, Kernel: 5.8.0-050800rc6daily20200721-generic (x86_64) 20200720, Desktop: GNOME Shell 3.36.1, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 3.3 Mesa 20.0.4 (LLVM 9.0.1 128 bits), Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080 Run 3: Processor: AMD EPYC 7F32 8-Core @ 3.70GHz (8 Cores / 16 Threads), Motherboard: Supermicro H11DSi-NT v2.00 (2.1 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 280GB INTEL SSDPE21D280GA, Graphics: llvmpipe, Monitor: VE228 OS: Ubuntu 20.04, Kernel: 5.8.0-050800rc6daily20200721-generic (x86_64) 20200720, Desktop: GNOME Shell 3.36.1, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 3.3 Mesa 20.0.4 (LLVM 9.0.1 128 bits), Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080 Run 4: Processor: AMD EPYC 7F32 8-Core @ 3.70GHz (8 Cores / 16 Threads), Motherboard: Supermicro H11DSi-NT v2.00 (2.1 BIOS), Chipset: AMD Starship/Matisse, Memory: 64GB, Disk: 280GB INTEL SSDPE21D280GA, Graphics: llvmpipe, Monitor: VE228 OS: Ubuntu 20.04, Kernel: 5.8.0-050800rc6daily20200721-generic (x86_64) 20200720, Desktop: GNOME Shell 3.36.1, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 3.3 Mesa 20.0.4 (LLVM 9.0.1 128 bits), Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 1920x1080 Unpacking The Linux Kernel linux-4.15.tar.xz Seconds < Lower Is Better Run 1 . 5.974 |================================================================ Run 2 . 5.972 |================================================================ Run 3 . 6.016 |================================================================ Run 4 . 5.852 |============================================================== CLOMP 1.2 Static OMP Speedup Speedup > Higher Is Better Run 1 . 29.8 |================================================================= Run 2 . 29.6 |================================================================= Run 3 . 29.0 |=============================================================== Run 4 . 29.7 |================================================================= oneDNN 2.0 Harness: IP Shapes 1D - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 3.54939 |============================================================== Run 2 . 3.55093 |============================================================== Run 3 . 3.55967 |============================================================== Run 4 . 3.55938 |============================================================== oneDNN 2.0 Harness: IP Shapes 3D - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 6.22116 |============================================================== Run 2 . 6.15155 |============================================================= Run 3 . 6.12560 |============================================================= Run 4 . 6.17937 |============================================================== oneDNN 2.0 Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 2.79140 |============================================================== Run 2 . 2.78976 |============================================================== Run 3 . 2.79003 |============================================================== Run 4 . 2.79287 |============================================================== oneDNN 2.0 Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 0.788963 |============================================================ Run 2 . 0.795590 |============================================================= Run 3 . 0.797020 |============================================================= Run 4 . 0.788043 |============================================================ oneDNN 2.0 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 4.98064 |============================================================== Run 2 . 4.92337 |============================================================= Run 3 . 4.95449 |============================================================== Run 4 . 4.92898 |============================================================= oneDNN 2.0 Harness: Deconvolution Batch shapes_1d - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 4.94409 |============================================================== Run 2 . 4.96823 |============================================================== Run 3 . 4.95793 |============================================================== Run 4 . 4.93690 |============================================================== oneDNN 2.0 Harness: Deconvolution Batch shapes_3d - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 6.42420 |============================================================== Run 2 . 6.42196 |============================================================== Run 3 . 6.42311 |============================================================== Run 4 . 6.41776 |============================================================== oneDNN 2.0 Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 9.66757 |============================================================== Run 2 . 9.62355 |============================================================== Run 3 . 9.46619 |============================================================= Run 4 . 9.59529 |============================================================== oneDNN 2.0 Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 7.20000 |============================================================== Run 2 . 7.21888 |============================================================== Run 3 . 7.22277 |============================================================== Run 4 . 7.22008 |============================================================== oneDNN 2.0 Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 5.58298 |============================================================== Run 2 . 5.57986 |============================================================== Run 3 . 5.57860 |============================================================== Run 4 . 5.58763 |============================================================== oneDNN 2.0 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 3604.40 |============================================================== Run 2 . 3436.43 |=========================================================== Run 3 . 3431.50 |=========================================================== Run 4 . 3417.31 |=========================================================== oneDNN 2.0 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 1708.29 |============================================================== Run 2 . 1708.22 |============================================================== Run 3 . 1711.14 |============================================================== Run 4 . 1711.87 |============================================================== oneDNN 2.0 Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 3429.54 |============================================================== Run 2 . 3438.99 |============================================================== Run 3 . 3436.78 |============================================================== Run 4 . 3427.26 |============================================================== oneDNN 2.0 Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 1711.39 |============================================================== Run 2 . 1718.14 |============================================================== Run 3 . 1716.53 |============================================================== Run 4 . 1711.28 |============================================================== oneDNN 2.0 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU ms < Lower Is Better Run 1 . 1.16439 |============================================================== Run 2 . 1.15868 |============================================================== Run 3 . 1.15947 |============================================================== Run 4 . 1.16157 |============================================================== oneDNN 2.0 Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better Run 1 . 3417.91 |============================================================== Run 2 . 3441.93 |============================================================== Run 3 . 3441.20 |============================================================== Run 4 . 3421.95 |============================================================== oneDNN 2.0 Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU ms < Lower Is Better Run 1 . 1704.97 |============================================================== Run 2 . 1715.65 |============================================================== Run 3 . 1712.85 |============================================================== Run 4 . 1714.22 |============================================================== oneDNN 2.0 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: u8s8f32 - Engine: CPU ms < Lower Is Better Run 1 . 3.59741 |============================================================== Run 2 . 3.59581 |============================================================== Run 3 . 3.59103 |============================================================== Run 4 . 3.59799 |============================================================== Build2 0.13 Time To Compile Seconds < Lower Is Better Run 1 . 112.08 |=============================================================== Run 2 . 112.78 |=============================================================== Run 3 . 112.67 |=============================================================== Run 4 . 112.46 |=============================================================== Timed Eigen Compilation 3.3.9 Time To Compile Seconds < Lower Is Better Run 1 . 82.88 |================================================================ Run 2 . 82.95 |================================================================ Run 3 . 83.07 |================================================================ Run 4 . 82.99 |================================================================ Monkey Audio Encoding 3.99.6 WAV To APE Seconds < Lower Is Better Run 1 . 12.51 |================================================================ Run 2 . 12.51 |================================================================ Run 3 . 12.49 |================================================================ Run 4 . 12.49 |================================================================ Ogg Audio Encoding 1.3.4 WAV To Ogg Seconds < Lower Is Better Run 1 . 20.57 |================================================================ Run 2 . 20.57 |================================================================ Run 3 . 20.58 |================================================================ Run 4 . 20.56 |================================================================ Opus Codec Encoding 1.3.1 WAV To Opus Encode Seconds < Lower Is Better Run 1 . 7.972 |================================================================ Run 2 . 7.973 |================================================================ Run 3 . 7.969 |================================================================ Run 4 . 7.966 |================================================================ NCNN 20201218 Target: CPU - Model: mobilenet ms < Lower Is Better Run 1 . 19.94 |================================================================ Run 2 . 19.81 |================================================================ Run 3 . 19.89 |================================================================ Run 4 . 19.88 |================================================================ NCNN 20201218 Target: CPU-v2-v2 - Model: mobilenet-v2 ms < Lower Is Better Run 1 . 6.87 |================================================================= Run 2 . 6.84 |================================================================= Run 3 . 6.85 |================================================================= Run 4 . 6.83 |================================================================= NCNN 20201218 Target: CPU-v3-v3 - Model: mobilenet-v3 ms < Lower Is Better Run 1 . 6.66 |================================================================= Run 2 . 6.64 |================================================================= Run 3 . 6.62 |================================================================= Run 4 . 6.60 |================================================================ NCNN 20201218 Target: CPU - Model: shufflenet-v2 ms < Lower Is Better Run 1 . 9.61 |================================================================= Run 2 . 9.63 |================================================================= Run 3 . 9.60 |================================================================= Run 4 . 9.60 |================================================================= NCNN 20201218 Target: CPU - Model: mnasnet ms < Lower Is Better Run 1 . 6.14 |================================================================= Run 2 . 5.94 |=============================================================== Run 3 . 5.97 |=============================================================== Run 4 . 6.02 |================================================================ NCNN 20201218 Target: CPU - Model: efficientnet-b0 ms < Lower Is Better Run 1 . 10.58 |================================================================ Run 2 . 10.55 |================================================================ Run 3 . 10.46 |=============================================================== Run 4 . 10.48 |=============================================================== NCNN 20201218 Target: CPU - Model: blazeface ms < Lower Is Better Run 1 . 3.31 |================================================================ Run 2 . 3.34 |================================================================= Run 3 . 3.30 |================================================================ Run 4 . 3.31 |================================================================ NCNN 20201218 Target: CPU - Model: googlenet ms < Lower Is Better Run 1 . 15.65 |============================================================= Run 2 . 15.52 |============================================================= Run 3 . 15.48 |============================================================ Run 4 . 16.38 |================================================================ NCNN 20201218 Target: CPU - Model: vgg16 ms < Lower Is Better Run 1 . 32.45 |================================================================ Run 2 . 32.41 |================================================================ Run 3 . 32.32 |================================================================ Run 4 . 32.31 |================================================================ NCNN 20201218 Target: CPU - Model: resnet18 ms < Lower Is Better Run 1 . 11.52 |================================================================ Run 2 . 11.57 |================================================================ Run 3 . 11.48 |================================================================ Run 4 . 11.57 |================================================================ NCNN 20201218 Target: CPU - Model: alexnet ms < Lower Is Better Run 1 . 7.58 |================================================================= Run 2 . 7.60 |================================================================= Run 3 . 7.60 |================================================================= Run 4 . 7.59 |================================================================= NCNN 20201218 Target: CPU - Model: resnet50 ms < Lower Is Better Run 1 . 23.02 |================================================================ Run 2 . 22.96 |================================================================ Run 3 . 22.89 |================================================================ Run 4 . 23.04 |================================================================ NCNN 20201218 Target: CPU - Model: yolov4-tiny ms < Lower Is Better Run 1 . 27.06 |================================================================ Run 2 . 26.71 |=============================================================== Run 3 . 26.40 |============================================================== Run 4 . 27.10 |================================================================ NCNN 20201218 Target: CPU - Model: squeezenet_ssd ms < Lower Is Better Run 1 . 24.86 |=============================================================== Run 2 . 24.48 |============================================================== Run 3 . 24.46 |============================================================== Run 4 . 25.31 |================================================================ NCNN 20201218 Target: CPU - Model: regnety_400m ms < Lower Is Better Run 1 . 32.07 |=============================================================== Run 2 . 32.19 |================================================================ Run 3 . 32.35 |================================================================ Run 4 . 32.03 |=============================================================== WavPack Audio Encoding 5.3 WAV To WavPack Seconds < Lower Is Better Run 1 . 13.73 |================================================================ Run 2 . 13.73 |================================================================ Run 3 . 13.73 |================================================================ Run 4 . 13.73 |================================================================ Unpacking Firefox 84.0 Extracting: firefox-84.0.source.tar.xz Seconds < Lower Is Better Run 1 . 20.37 |================================================================ Run 2 . 20.27 |================================================================ Run 3 . 20.22 |================================================================ Run 4 . 20.21 |===============================================================