EPYC Whitepaper

2 x AMD EPYC 7601 32-Core testing with a Dell 02MJ3T (1.2.5 BIOS) and llvmpipe 504GB on Ubuntu 19.10 via the Phoronix Test Suite.

HTML result view exported from: https://openbenchmarking.org/result/1910236-HU-EPYCWHITE63.

EPYC WhitepaperProcessorMotherboardChipsetMemoryDiskGraphicsMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLCompilerFile-SystemScreen ResolutionEPYC 7601 2P2 x AMD EPYC 7601 32-Core (64 Cores / 128 Threads)Dell 02MJ3T (1.2.5 BIOS)AMD 17h516096MB280GB INTEL SSDPED1D280GA + 12 x 500GB Samsung SSD 860 + 120GB SSDSCKJB120G7Rllvmpipe 504GBVE2282 x Broadcom BCM57416 NetXtreme-E Dual-Media 10G RDMA + 2 x Broadcom NetXtreme BCM5720 2-port PCIeUbuntu 19.105.3.0-19-generic (x86_64)GNOME Shell 3.34.1X Server 1.20.5modesetting 1.20.53.3 Mesa 19.2.1 (LLVM 9.0 128 bits)GCC 9.2.1 20191008ext41600x1200OpenBenchmarking.org- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --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++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-offload-targets=nvptx-none,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 - OpenJDK Runtime Environment (build 11.0.5-ea+10-post-Ubuntu-0ubuntu1)- 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

EPYC Whitepaperstream: Copystream: Scalestream: Triadstream: Addhpcc: G-HPLhpcc: G-Fftehpcc: G-Fftehpcc: EP-DGEMMhpcc: G-Ptranshpcc: EP-STREAM Triadhpcc: G-Rand Accesshpcc: Rand Ring Latencyhpcc: Rand Ring Bandwidthhpcc: Max Ping Pong Bandwidthminife: Smallnamd: ATPase Simulation - 327,506 Atomsfftw: Float + SSE - 1D FFT Size 4096fftw: Float + SSE - 2D FFT Size 4096hmmer: Pfam Database Searchembree: Pathtracer - Crownembree: Pathtracer ISPC - Crownembree: Pathtracer - Asian Dragonembree: Pathtracer - Asian Dragon Objembree: Pathtracer ISPC - Asian Dragonembree: Pathtracer ISPC - Asian Dragon Objmt-dgemm: Sustained Floating-Point Rateaskap: tConvolve MT - Griddingaskap: tConvolve MT - Degriddingaskap: tConvolve MPI - Griddingaskap: tConvolve MPI - Degriddingaskap: tConvolve OpenMP - Griddingaskap: tConvolve OpenMP - Degriddingspec-jbb2015: SPECjbb2015-Composite max-jOPSspec-jbb2015: SPECjbb2015-Composite critical-jOPSEPYC 7601 2P158117.20157903.12178352.60178494.72196.9760027.2760027.2760016.787705.243692.817140.166722.704640.9124110937.7186735.320.4501627881133266.3727.6826.3428.1922.0827.5622.824.552535.336153.928154.5910755.57800.701771.656530231561OpenBenchmarking.org

Stream

Type: Copy

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: CopyEPYC 7601 2P30K60K90K120K150KSE +/- 1259.52, N = 5158117.201. (CC) gcc options: -O3 -march=native -fopenmp

Stream

Type: Scale

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: ScaleEPYC 7601 2P30K60K90K120K150KSE +/- 1398.05, N = 5157903.121. (CC) gcc options: -O3 -march=native -fopenmp

Stream

Type: Triad

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: TriadEPYC 7601 2P40K80K120K160K200KSE +/- 1850.63, N = 5178352.601. (CC) gcc options: -O3 -march=native -fopenmp

Stream

Type: Add

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: AddEPYC 7601 2P40K80K120K160K200KSE +/- 976.05, N = 5178494.721. (CC) gcc options: -O3 -march=native -fopenmp

HPC Challenge

Test / Class: G-HPL

OpenBenchmarking.orgGFLOPS, More Is BetterHPC Challenge 1.5.0Test / Class: G-HPLEPYC 7601 2P4080120160200196.981. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: G-Ffte

OpenBenchmarking.orgGFLOPS, More Is BetterHPC Challenge 1.5.0Test / Class: G-FfteEPYC 7601 2P61218243027.281. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: G-Ffte

OpenBenchmarking.orgGFLOP/s, More Is BetterHPC Challenge 1.5.0Test / Class: G-FfteEPYC 7601 2P61218243027.281. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: EP-DGEMM

OpenBenchmarking.orgGFLOPS, More Is BetterHPC Challenge 1.5.0Test / Class: EP-DGEMMEPYC 7601 2P4812162016.791. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: G-Ptrans

OpenBenchmarking.orgGB/s, More Is BetterHPC Challenge 1.5.0Test / Class: G-PtransEPYC 7601 2P1.17982.35963.53944.71925.8995.243691. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: EP-STREAM Triad

OpenBenchmarking.orgGB/s, More Is BetterHPC Challenge 1.5.0Test / Class: EP-STREAM TriadEPYC 7601 2P0.63391.26781.90172.53563.16952.817141. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: G-Random Access

OpenBenchmarking.orgGUP/s, More Is BetterHPC Challenge 1.5.0Test / Class: G-Random AccessEPYC 7601 2P0.03750.0750.11250.150.18750.166721. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: Random Ring Latency

OpenBenchmarking.orgusecs, Fewer Is BetterHPC Challenge 1.5.0Test / Class: Random Ring LatencyEPYC 7601 2P0.60851.2171.82552.4343.04252.704641. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: Random Ring Bandwidth

OpenBenchmarking.orgGB/s, More Is BetterHPC Challenge 1.5.0Test / Class: Random Ring BandwidthEPYC 7601 2P0.20530.41060.61590.82121.02650.912411. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

HPC Challenge

Test / Class: Max Ping Pong Bandwidth

OpenBenchmarking.orgMB/s, More Is BetterHPC Challenge 1.5.0Test / Class: Max Ping Pong BandwidthEPYC 7601 2P2K4K6K8K10K10937.721. (CC) gcc options: -lblas -lm -pthread -lmpi -fomit-frame-pointer -funroll-loops2. OpenBLAS + Open MPI 3.1.3

miniFE

Problem Size: Small

OpenBenchmarking.orgCG Mflops, More Is BetterminiFE 2.2Problem Size: SmallEPYC 7601 2P14002800420056007000SE +/- 65.81, N = 36735.321. (CXX) g++ options: -O3 -fopenmp -pthread -lmpi_cxx -lmpi

NAMD

ATPase Simulation - 327,506 Atoms

OpenBenchmarking.orgdays/ns, Fewer Is BetterNAMD 2.13b1ATPase Simulation - 327,506 AtomsEPYC 7601 2P0.10130.20260.30390.40520.5065SE +/- 0.00039, N = 120.45016

FFTW

Build: Float + SSE - Size: 1D FFT Size 4096

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 1D FFT Size 4096EPYC 7601 2P6K12K18K24K30KSE +/- 257.07, N = 3278811. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

FFTW

Build: Float + SSE - Size: 2D FFT Size 4096

OpenBenchmarking.orgMflops, More Is BetterFFTW 3.3.6Build: Float + SSE - Size: 2D FFT Size 4096EPYC 7601 2P3K6K9K12K15KSE +/- 150.93, N = 3133261. (CC) gcc options: -pthread -O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns -ffast-math -lm

Timed HMMer Search

Pfam Database Search

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed HMMer Search 2.3.2Pfam Database SearchEPYC 7601 2P246810SE +/- 0.06, N = 36.371. (CC) gcc options: -O2 -pthread -lhmmer -lsquid -lm

Embree

Binary: Pathtracer - Model: Crown

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer - Model: CrownEPYC 7601 2P714212835SE +/- 0.07, N = 327.68MIN: 27.18 / MAX: 28.74

Embree

Binary: Pathtracer ISPC - Model: Crown

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer ISPC - Model: CrownEPYC 7601 2P612182430SE +/- 0.09, N = 326.34MIN: 25.81 / MAX: 27.61

Embree

Binary: Pathtracer - Model: Asian Dragon

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer - Model: Asian DragonEPYC 7601 2P714212835SE +/- 0.02, N = 328.19MIN: 27.9 / MAX: 28.87

Embree

Binary: Pathtracer - Model: Asian Dragon Obj

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer - Model: Asian Dragon ObjEPYC 7601 2P510152025SE +/- 0.25, N = 322.08MIN: 21.52 / MAX: 22.97

Embree

Binary: Pathtracer ISPC - Model: Asian Dragon

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer ISPC - Model: Asian DragonEPYC 7601 2P612182430SE +/- 0.06, N = 327.56MIN: 27.14 / MAX: 28.29

Embree

Binary: Pathtracer ISPC - Model: Asian Dragon Obj

OpenBenchmarking.orgFrames Per Second, More Is BetterEmbree 3.6.1Binary: Pathtracer ISPC - Model: Asian Dragon ObjEPYC 7601 2P510152025SE +/- 0.29, N = 322.82MIN: 22.09 / MAX: 23.73

ACES DGEMM

Sustained Floating-Point Rate

OpenBenchmarking.orgGFLOP/s, More Is BetterACES DGEMM 1.0Sustained Floating-Point RateEPYC 7601 2P1.02382.04763.07144.09525.119SE +/- 0.06, N = 44.551. (CC) gcc options: -O3 -march=native -fopenmp

ASKAP

Test: tConvolve MT - Gridding

OpenBenchmarking.orgMillion Grid Points Per Second, More Is BetterASKAP 2018-11-10Test: tConvolve MT - GriddingEPYC 7601 2P5001000150020002500SE +/- 124.64, N = 122535.331. (CXX) g++ options: -lpthread

ASKAP

Test: tConvolve MT - Degridding

OpenBenchmarking.orgMillion Grid Points Per Second, More Is BetterASKAP 2018-11-10Test: tConvolve MT - DegriddingEPYC 7601 2P13002600390052006500SE +/- 117.38, N = 126153.921. (CXX) g++ options: -lpthread

ASKAP

Test: tConvolve MPI - Gridding

OpenBenchmarking.orgMillion Grid Points Per Second, More Is BetterASKAP 2018-11-10Test: tConvolve MPI - GriddingEPYC 7601 2P2K4K6K8K10KSE +/- 1.30, N = 38154.591. (CXX) g++ options: -lpthread

ASKAP

Test: tConvolve MPI - Degridding

OpenBenchmarking.orgMillion Grid Points Per Second, More Is BetterASKAP 2018-11-10Test: tConvolve MPI - DegriddingEPYC 7601 2P2K4K6K8K10KSE +/- 5.97, N = 310755.571. (CXX) g++ options: -lpthread

ASKAP

Test: tConvolve OpenMP - Gridding

OpenBenchmarking.orgMillion Grid Points Per Second, More Is BetterASKAP 2018-11-10Test: tConvolve OpenMP - GriddingEPYC 7601 2P2004006008001000SE +/- 11.35, N = 15800.701. (CXX) g++ options: -lpthread

ASKAP

Test: tConvolve OpenMP - Degridding

OpenBenchmarking.orgMillion Grid Points Per Second, More Is BetterASKAP 2018-11-10Test: tConvolve OpenMP - DegriddingEPYC 7601 2P400800120016002000SE +/- 147.52, N = 151771.651. (CXX) g++ options: -lpthread

SPECjbb 2015

SPECjbb2015-Composite max-jOPS

OpenBenchmarking.orgjOPS, More Is BetterSPECjbb 2015SPECjbb2015-Composite max-jOPSEPYC 7601 2P14K28K42K56K70K65302

SPECjbb 2015

SPECjbb2015-Composite critical-jOPS

OpenBenchmarking.orgjOPS, More Is BetterSPECjbb 2015SPECjbb2015-Composite critical-jOPSEPYC 7601 2P7K14K21K28K35K31561

Phoronix Test Suite v10.8.4