epyc turin Benchmarks for a future article. 2 x AMD EPYC 9755 128-Core testing with a AMD VOLCANO (RVOT1001B BIOS) and ASPEED on Ubuntu 24.04 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2410191-NE-EPYCTURIN56&sro&grw .
epyc turin Processor Motherboard Chipset Memory Disk Graphics Network OS Kernel Compiler File-System Screen Resolution a b 2 x AMD EPYC 9755 128-Core @ 2.70GHz (256 Cores / 512 Threads) AMD VOLCANO (RVOT1001B BIOS) AMD Device 153a 1520GB 512GB SAMSUNG MZVL2512HCJQ-00B00 + 3201GB Micron_7450_MTFDKCB3T2TFS ASPEED Broadcom NetXtreme BCM5720 PCIe Ubuntu 24.04 6.12.0-rc3-phx (x86_64) GCC 13.2.0 + Clang 18.1.3 ext4 1024x768 OpenBenchmarking.org Kernel Details - Transparent Huge Pages: madvise Compiler Details - --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-cet --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --enable-libphobos-checking=release --enable-libstdcxx-backtrace --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-uJ7kn6/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-uJ7kn6/gcc-13-13.2.0/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-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 performance (Boost: Enabled) - CPU Microcode: 0xb002116 Security Details - gather_data_sampling: Not affected + itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + reg_file_data_sampling: Not affected + retbleed: Not affected + spec_rstack_overflow: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable; IBPB: disabled; STIBP: disabled; PBRSB-eIBRS: Not affected; BHI: Not affected + srbds: Not affected + tsx_async_abort: Not affected
epyc turin litert: Mobilenet Quant warpx: Uniform Plasma litert: Mobilenet Float litert: SqueezeNet litert: Inception V4 xnnpack: FP32MobileNetV1 litert: DeepLab V3 xnnpack: FP32MobileNetV3Large xnnpack: FP32MobileNetV3Small warpx: Plasma Acceleration xnnpack: FP16MobileNetV1 litert: Inception ResNet V2 xnnpack: FP32MobileNetV2 xnnpack: FP16MobileNetV2 xnnpack: FP16MobileNetV3Large xnnpack: FP16MobileNetV3Small litert: Quantized COCO SSD MobileNet v1 litert: NASNet Mobile xnnpack: QS8MobileNetV2 epoch: Cone onednn: IP Shapes 1D - CPU onednn: IP Shapes 3D - CPU onednn: Convolution Batch Shapes Auto - CPU onednn: Deconvolution Batch shapes_1d - CPU onednn: Deconvolution Batch shapes_3d - CPU onednn: Recurrent Neural Network Training - CPU onednn: Recurrent Neural Network Inference - CPU a b 43601.8 20.383024 75229.3 105380 422782 126508 93820.7 322181 306213 23.47043634 121353 824335 349733 243557 278473 308464 63357.4 1831650 202569 283.75 0.797341 0.652196 0.298302 26.1446 0.423216 761.155 514.987 46489.2 20.38324203 24645.9 46189.4 380938 30213 102742.9 269833 286163 23.42649518 99495 363512 283027 262462 304488 406223 71555.6 1254561 238886 280.78 0.799482 0.652110 0.301225 26.5673 0.421823 758.564 516.464 OpenBenchmarking.org
LiteRT Model: Mobilenet Quant OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: Mobilenet Quant a b 10K 20K 30K 40K 50K 43601.8 46489.2
WarpX Input: Uniform Plasma OpenBenchmarking.org Seconds, Fewer Is Better WarpX 24.10 Input: Uniform Plasma a b 5 10 15 20 25 SE +/- 0.19, N = 15 20.38 20.38 1. (CXX) g++ options: -O3 -lm
LiteRT Model: Mobilenet Float OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: Mobilenet Float a b 16K 32K 48K 64K 80K 75229.3 24645.9
LiteRT Model: SqueezeNet OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: SqueezeNet a b 20K 40K 60K 80K 100K SE +/- 759.10, N = 15 105380.0 46189.4
LiteRT Model: Inception V4 OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: Inception V4 a b 90K 180K 270K 360K 450K SE +/- 19162.36, N = 12 422782 380938
XNNPACK Model: FP32MobileNetV1 OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP32MobileNetV1 a b 30K 60K 90K 120K 150K 126508 30213 1. (CXX) g++ options: -O3 -lrt -lm
LiteRT Model: DeepLab V3 OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: DeepLab V3 a b 20K 40K 60K 80K 100K SE +/- 6031.32, N = 12 93820.7 102742.9
XNNPACK Model: FP32MobileNetV3Large OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP32MobileNetV3Large a b 70K 140K 210K 280K 350K 322181 269833 1. (CXX) g++ options: -O3 -lrt -lm
XNNPACK Model: FP32MobileNetV3Small OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP32MobileNetV3Small a b 70K 140K 210K 280K 350K 306213 286163 1. (CXX) g++ options: -O3 -lrt -lm
WarpX Input: Plasma Acceleration OpenBenchmarking.org Seconds, Fewer Is Better WarpX 24.10 Input: Plasma Acceleration a b 6 12 18 24 30 SE +/- 0.13, N = 3 23.47 23.43 1. (CXX) g++ options: -O3 -lm
XNNPACK Model: FP16MobileNetV1 OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP16MobileNetV1 a b 30K 60K 90K 120K 150K 121353 99495 1. (CXX) g++ options: -O3 -lrt -lm
LiteRT Model: Inception ResNet V2 OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: Inception ResNet V2 a b 200K 400K 600K 800K 1000K 824335 363512
XNNPACK Model: FP32MobileNetV2 OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP32MobileNetV2 a b 70K 140K 210K 280K 350K 349733 283027 1. (CXX) g++ options: -O3 -lrt -lm
XNNPACK Model: FP16MobileNetV2 OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP16MobileNetV2 a b 60K 120K 180K 240K 300K 243557 262462 1. (CXX) g++ options: -O3 -lrt -lm
XNNPACK Model: FP16MobileNetV3Large OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP16MobileNetV3Large a b 70K 140K 210K 280K 350K 278473 304488 1. (CXX) g++ options: -O3 -lrt -lm
XNNPACK Model: FP16MobileNetV3Small OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: FP16MobileNetV3Small a b 90K 180K 270K 360K 450K 308464 406223 1. (CXX) g++ options: -O3 -lrt -lm
LiteRT Model: Quantized COCO SSD MobileNet v1 OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: Quantized COCO SSD MobileNet v1 a b 15K 30K 45K 60K 75K 63357.4 71555.6
LiteRT Model: NASNet Mobile OpenBenchmarking.org Microseconds, Fewer Is Better LiteRT 2024-10-15 Model: NASNet Mobile a b 400K 800K 1200K 1600K 2000K SE +/- 119244.97, N = 12 1831650 1254561
XNNPACK Model: QS8MobileNetV2 OpenBenchmarking.org us, Fewer Is Better XNNPACK b7b048 Model: QS8MobileNetV2 a b 50K 100K 150K 200K 250K 202569 238886 1. (CXX) g++ options: -O3 -lrt -lm
Epoch Epoch3D Deck: Cone OpenBenchmarking.org Seconds, Fewer Is Better Epoch 4.19.4 Epoch3D Deck: Cone a b 60 120 180 240 300 SE +/- 0.98, N = 3 283.75 280.78 1. (F9X) gfortran options: -O3 -std=f2003 -Jobj -lsdf -lmpi_usempif08 -lmpi_mpifh -lmpi -lopen-rte -lopen-pal -lhwloc -levent_core -levent_pthreads -lm -lz
oneDNN Harness: IP Shapes 1D - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: IP Shapes 1D - Engine: CPU a b 0.1799 0.3598 0.5397 0.7196 0.8995 SE +/- 0.000618, N = 3 0.797341 0.799482 MIN: 0.76 MIN: 0.76 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
oneDNN Harness: IP Shapes 3D - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: IP Shapes 3D - Engine: CPU a b 0.1467 0.2934 0.4401 0.5868 0.7335 SE +/- 0.003624, N = 3 0.652196 0.652110 MIN: 0.61 MIN: 0.6 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
oneDNN Harness: Convolution Batch Shapes Auto - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: Convolution Batch Shapes Auto - Engine: CPU a b 0.0678 0.1356 0.2034 0.2712 0.339 SE +/- 0.001827, N = 3 0.298302 0.301225 MIN: 0.29 MIN: 0.28 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
oneDNN Harness: Deconvolution Batch shapes_1d - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: Deconvolution Batch shapes_1d - Engine: CPU a b 6 12 18 24 30 SE +/- 0.13, N = 3 26.14 26.57 MIN: 24.29 MIN: 24.16 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
oneDNN Harness: Deconvolution Batch shapes_3d - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: Deconvolution Batch shapes_3d - Engine: CPU a b 0.0952 0.1904 0.2856 0.3808 0.476 SE +/- 0.000942, N = 3 0.423216 0.421823 MIN: 0.41 MIN: 0.4 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
oneDNN Harness: Recurrent Neural Network Training - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: Recurrent Neural Network Training - Engine: CPU a b 160 320 480 640 800 SE +/- 0.94, N = 3 761.16 758.56 MIN: 753.78 MIN: 750.89 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
oneDNN Harness: Recurrent Neural Network Inference - Engine: CPU OpenBenchmarking.org ms, Fewer Is Better oneDNN 3.6 Harness: Recurrent Neural Network Inference - Engine: CPU a b 110 220 330 440 550 SE +/- 1.40, N = 3 514.99 516.46 MIN: 510.29 MIN: 510.39 1. (CXX) g++ options: -O3 -march=native -fopenmp -msse4.1 -fPIC -fcf-protection=full -pie -ldl -lpthread
Phoronix Test Suite v10.8.5