GCC 13 Compiler Benchmarks AMD EPYC Genoa

AMD EPYC 9654 GCC 13 development compiler benchmarks by Michael Larabel for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2301045-NE-GCC13DEVE69&grr&sor.

GCC 13 Compiler Benchmarks AMD EPYC GenoaProcessorMotherboardChipsetMemoryDiskGraphicsMonitorNetworkOSKernelDesktopDisplay ServerVulkanCompilerFile-SystemScreen ResolutionZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-5122 x AMD EPYC 9654 96-Core @ 3.71GHz (192 Cores / 384 Threads)AMD Titanite_4G (RTI1002E BIOS)AMD Device 14a41520GB800GB INTEL SSDPF21Q800GBASPEEDVGA HDMIBroadcom NetXtreme BCM5720 PCIeUbuntu 23.045.19.0-21-generic (x86_64)GNOME Shell 43.1X Server 1.21.1.41.3.224GCC 13.0.0 20230103ext41920x1080OpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseEnvironment Details- Znver4: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"- Znver4 + Prefer AVX-512: CXXFLAGS="-O3 -march=native -mprefer-vector-width=512 -flto" CFLAGS="-O3 -march=native -mprefer-vector-width=512 -flto" - Znver3: CXXFLAGS="-O3 -march=znver3 -flto" CFLAGS="-O3 -march=znver3 -flto"- Znver3 + AVX-512: CXXFLAGS="-O3 -march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -flto" CFLAGS="-O3 -march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -flto" Compiler Details- --disable-multilib --enable-checking=releaseProcessor Details- Scaling Governor: amd-pstate performance (Boost: Enabled) - CPU Microcode: 0xa10110dPython Details- Python 3.10.9Security Details- itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + mmio_stale_data: Not affected + retbleed: 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 Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling PBRSB-eIBRS: Not affected + srbds: Not affected + tsx_async_abort: Not affected

GCC 13 Compiler Benchmarks AMD EPYC Genoajpegxl: JPEG - 100cp2k: Fayalite-FISTonednn: Recurrent Neural Network Inference - u8s8f32 - CPUgraphics-magick: Resizingjpegxl: PNG - 100securemark: SecureMark-TLSonednn: Recurrent Neural Network Training - bf16bf16bf16 - CPUgraphics-magick: Noise-Gaussianonednn: Recurrent Neural Network Training - u8s8f32 - CPUcompress-zstd: 19, Long Mode - Decompression Speedcompress-zstd: 19, Long Mode - Compression Speedgraphics-magick: HWB Color Spaceopenssl: SHA256onednn: Recurrent Neural Network Inference - bf16bf16bf16 - CPUlammps: 20k Atomsjpegxl: JPEG - 90jpegxl: PNG - 90ngspice: C2670ngspice: C7552compress-zstd: 19 - Decompression Speedcompress-zstd: 19 - Compression Speedgraphics-magick: Swirlpjsip: INVITEonednn: IP Shapes 1D - bf16bf16bf16 - CPUpjsip: OPTIONS, Statefulavifenc: 0stargate: 192000 - 1024simdjson: DistinctUserIDsimdjson: PartialTweetssimdjson: TopTweetopenssl: RSA4096openssl: RSA4096graphics-magick: Sharpengraphics-magick: Rotategraphics-magick: Enhancedsimdjson: Kostyajpegxl-decode: 1onednn: IP Shapes 1D - u8s8f32 - CPUsimdjson: LargeRandkripke: webp: Quality 100, Lossless, Highest Compressiononednn: IP Shapes 3D - bf16bf16bf16 - CPUstargate: 96000 - 1024cpuminer-opt: Deepcoinsvt-av1: Preset 4 - Bosphorus 4Kavifenc: 2cpuminer-opt: Quad SHA-256, Pyritepjsip: OPTIONS, Statelesscpuminer-opt: Garlicoincpuminer-opt: LBC, LBRY Creditscpuminer-opt: scryptcpuminer-opt: Triple SHA-256, Onecoincpuminer-opt: x25xcpuminer-opt: Skeincoincpuminer-opt: Magiquantlib: onednn: Deconvolution Batch shapes_1d - u8s8f32 - CPUgpaw: Carbon Nanotubesvt-av1: Preset 12 - Bosphorus 4Kjpegxl-decode: Allkvazaar: Bosphorus 4K - Ultra Fastgromacs: MPI CPU - water_GMX50_barecoremark: CoreMark Size 666 - Iterations Per Secondonednn: Deconvolution Batch shapes_1d - bf16bf16bf16 - CPUsvt-av1: Preset 13 - Bosphorus 4Kliquid-dsp: 384 - 256 - 57liquid-dsp: 256 - 256 - 57liquid-dsp: 128 - 256 - 57minibude: OpenMP - BM2minibude: OpenMP - BM2webp: Quality 100, Losslesssmhasher: MeowHash x86_64 AES-NIsmhasher: MeowHash x86_64 AES-NIonednn: Convolution Batch Shapes Auto - bf16bf16bf16 - CPUonednn: Matrix Multiply Batch Shapes Transformer - bf16bf16bf16 - CPUsvt-av1: Preset 8 - Bosphorus 4Kkvazaar: Bosphorus 4K - Mediumkvazaar: Bosphorus 4K - Very Fastonednn: IP Shapes 3D - u8s8f32 - CPUsmhasher: FarmHash32 x86_64 AVXsmhasher: FarmHash32 x86_64 AVXonednn: Deconvolution Batch shapes_3d - u8s8f32 - CPUastcenc: Mediumastcenc: Exhaustivewebp: Quality 100, Highest Compressionsmhasher: t1ha0_aes_avx2 x86_64smhasher: t1ha0_aes_avx2 x86_64onednn: Convolution Batch Shapes Auto - u8s8f32 - CPUastcenc: Thoroughavifenc: 6, Losslessminibude: OpenMP - BM1minibude: OpenMP - BM1avifenc: 10, Losslessmt-dgemm: Sustained Floating-Point Rateonednn: Deconvolution Batch shapes_3d - bf16bf16bf16 - CPUlammps: Rhodopsin Proteinavifenc: 6webp: Quality 100webp: DefaultZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-5120.731174.6652405.63880.832965482134.8510242020.243708.840.811802653267135872479.7255.6189.489.8395.06692.1023584.1102.12862520021.1359923761.1222.7837176.536.636.962939503.544490.1135967322344.1748.6414.81681.272615622800.574.266004.3312301591475.39234.10422510673357677241310654874790.1133066438042.8820147708440.793096.90.93716422.350210.386277.2078.6919.4937694653.8990162.33363196.1151124933333397895000006990866667264.1476603.6841.4544.97754272.140.4438910.50057993.75664.3570.710.88664226.49440565.070.274769493.224812.94123.2520.524102354.570.380486118.73924.398214.5025362.5423.54170.0529520.65810151.5852.34711.5418.950.751263.7072359.29870.822941222070.1310132123.353685.742.511672662301240702444.5255.7399.439.8195.59993.4533581.6105.32681508423.5341928861.0732.8672336.866.436.912924586.444301.8131465621504.1948.2114.29281.242546485330.584.333664.4082901622425.42333.90223239953367917213010858274782.7433012178217.7020093678467.243112.60.9541422.819196.804272.3174.9819.4377861097.8566402.29880208.6431127066666798098000006977633333265.5206638.0071.4744.95054297.040.4455320.48397994.34064.0570.810.88130326.50340563.960.274662420.699213.09363.2320.806102399.740.388483118.99324.462211.0175275.4263.57270.2982820.65006752.2872.31711.5418.970.741211.1262442.35890.822940572093.7510182108.463695.539.811342664643611932418.2855.7219.189.6195.17592.9433574.9104.42563514919.5885922661.6582.8200796.466.746.612938488.544435.3128564518374.2048.2214.61791.252638128470.584.524404.3564461649935.36033.9111378987336885495234970202959.1532552536116.9714140478490.733120.50.94701321.721222.902269.5978.0118.2317640546.2720742.31858219.5701117600000097356666676940833333264.2846607.0911.4744.95754281.940.4427480.51093995.11664.2772.010.90794126.49440565.720.279655459.692412.22213.6420.533102351.870.392765116.09124.437214.0785351.9443.58170.3780650.65634251.4552.33111.4818.990.711407.1252377.17860.812965752103.419752099.433684.743.910622668990894532531.4256.0369.279.8695.18693.5243594.8102.92826513222.9431923662.6842.8683736.616.686.732935372.144499.3132160522084.1647.0214.250211.252717357080.584.336584.4133791601575.37434.23022647473366157283710677434763.9133236807941.3820049908355.753114.90.93637422.823206.676266.5376.7619.0877871273.9327622.36166210.4261130166666798137000006999233333266.1556653.8641.4544.96854284.650.4312990.50484292.96463.8070.710.90230326.49440559.330.276965511.524813.03053.1120.527102403.920.392875117.69104.514214.6255365.6123.64770.1864670.64708751.8162.40611.3818.85OpenBenchmarking.org

JPEG XL libjxl

Input: JPEG - Quality: 100

OpenBenchmarking.orgMP/s, More Is BetterJPEG XL libjxl 0.7Input: JPEG - Quality: 100Znver4 + Prefer AVX-512Znver3Znver4Znver3 + AVX-5120.16880.33760.50640.67520.844SE +/- 0.01, N = 3SE +/- 0.01, N = 6SE +/- 0.01, N = 9SE +/- 0.01, N = 90.750.740.730.71-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -fno-rtti -funwind-tables -O2 -fPIE -pie -latomic

CP2K Molecular Dynamics

Input: Fayalite-FIST

OpenBenchmarking.orgSeconds, Fewer Is BetterCP2K Molecular Dynamics 8.2Input: Fayalite-FISTZnver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-512300600900120015001174.671211.131263.711407.13

oneDNN

Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Recurrent Neural Network Inference - Data Type: u8s8f32 - Engine: CPUZnver4 + Prefer AVX-512Znver3 + AVX-512Znver4Znver35001000150020002500SE +/- 31.29, N = 15SE +/- 26.39, N = 15SE +/- 20.65, N = 15SE +/- 32.77, N = 122359.292377.172405.632442.35MIN: 2066.84-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 2080.2MIN: 2158.4-march=znver3 - MIN: 2123.491. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

GraphicsMagick

Operation: Resizing

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: ResizingZnver3Znver4Znver4 + Prefer AVX-512Znver3 + AVX-51220406080100SE +/- 1.27, N = 15SE +/- 1.00, N = 15SE +/- 0.94, N = 15SE +/- 0.90, N = 1589888786-march=znver3-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

JPEG XL libjxl

Input: PNG - Quality: 100

OpenBenchmarking.orgMP/s, More Is BetterJPEG XL libjxl 0.7Input: PNG - Quality: 100Znver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-5120.18680.37360.56040.74720.934SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.830.820.820.81-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -fno-rtti -funwind-tables -O2 -fPIE -pie -latomic

SecureMark

Benchmark: SecureMark-TLS

OpenBenchmarking.orgmarks, More Is BetterSecureMark 1.0.4Benchmark: SecureMark-TLSZnver3 + AVX-512Znver4Znver4 + Prefer AVX-512Znver360K120K180K240K300KSE +/- 380.50, N = 3SE +/- 383.28, N = 3SE +/- 640.61, N = 3SE +/- 1276.92, N = 32965752965482941222940571. (CC) gcc options: -pedantic -O3

oneDNN

Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Recurrent Neural Network Training - Data Type: bf16bf16bf16 - Engine: CPUZnver4 + Prefer AVX-512Znver3Znver3 + AVX-512Znver45001000150020002500SE +/- 16.85, N = 3SE +/- 27.68, N = 15SE +/- 15.62, N = 15SE +/- 9.82, N = 32070.132093.752103.412134.85MIN: 1938.32-march=znver3 - MIN: 1826.26-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 1883.21MIN: 2008.71. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

GraphicsMagick

Operation: Noise-Gaussian

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: Noise-GaussianZnver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-5122004006008001000SE +/- 6.60, N = 15SE +/- 5.13, N = 3SE +/- 11.39, N = 15SE +/- 11.43, N = 15102410181013975-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

oneDNN

Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Recurrent Neural Network Training - Data Type: u8s8f32 - Engine: CPUZnver4Znver3 + AVX-512Znver3Znver4 + Prefer AVX-5125001000150020002500SE +/- 16.34, N = 3SE +/- 18.84, N = 15SE +/- 26.49, N = 3SE +/- 26.64, N = 122020.242099.432108.462123.35MIN: 1878.54-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 1917.53-march=znver3 - MIN: 1945.83MIN: 1873.091. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Zstd Compression

Compression Level: 19, Long Mode - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 19, Long Mode - Decompression SpeedZnver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-5128001600240032004000SE +/- 12.59, N = 15SE +/- 11.99, N = 15SE +/- 13.90, N = 15SE +/- 12.89, N = 153708.83695.53685.73684.7-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -O3 -flto -pthread -lz -llzma

Zstd Compression

Compression Level: 19, Long Mode - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 19, Long Mode - Compression SpeedZnver3 + AVX-512Znver4 + Prefer AVX-512Znver4Znver31020304050SE +/- 0.79, N = 15SE +/- 0.59, N = 15SE +/- 0.40, N = 15SE +/- 0.40, N = 1543.942.540.839.8-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CC) gcc options: -O3 -flto -pthread -lz -llzma

GraphicsMagick

Operation: HWB Color Space

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: HWB Color SpaceZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-51230060090012001500SE +/- 4.84, N = 3SE +/- 15.31, N = 15SE +/- 11.42, N = 15SE +/- 1.86, N = 31180116711341062-march=native-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

OpenSSL

Algorithm: SHA256

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.0Algorithm: SHA256Znver3 + AVX-512Znver3Znver4 + Prefer AVX-512Znver460000M120000M180000M240000M300000MSE +/- 143443542.20, N = 3SE +/- 18811489.95, N = 3SE +/- 172236238.05, N = 3SE +/- 150345864.28, N = 3266899089453266464361193266230124070265326713587-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver3-march=native-march=native1. (CC) gcc options: -pthread -m64 -O3 -flto -lssl -lcrypto -ldl

oneDNN

Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Recurrent Neural Network Inference - Data Type: bf16bf16bf16 - Engine: CPUZnver3Znver4 + Prefer AVX-512Znver4Znver3 + AVX-5125001000150020002500SE +/- 24.11, N = 15SE +/- 25.52, N = 3SE +/- 24.81, N = 3SE +/- 16.53, N = 32418.282444.522479.722531.42-march=znver3 - MIN: 2149.67MIN: 2258.45MIN: 2315.89-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 2281.871. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

LAMMPS Molecular Dynamics Simulator

Model: 20k Atoms

OpenBenchmarking.orgns/day, More Is BetterLAMMPS Molecular Dynamics Simulator 23Jun2022Model: 20k AtomsZnver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver41326395265SE +/- 0.08, N = 3SE +/- 0.21, N = 3SE +/- 0.04, N = 3SE +/- 0.15, N = 356.0455.7455.7255.62-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver3-march=native1. (CXX) g++ options: -O3 -flto -lm -ldl

JPEG XL libjxl

Input: JPEG - Quality: 90

OpenBenchmarking.orgMP/s, More Is BetterJPEG XL libjxl 0.7Input: JPEG - Quality: 90Znver4Znver4 + Prefer AVX-512Znver3 + AVX-512Znver33691215SE +/- 0.05, N = 3SE +/- 0.03, N = 3SE +/- 0.13, N = 3SE +/- 0.03, N = 39.489.439.279.18-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto -fno-rtti -funwind-tables -O2 -fPIE -pie -latomic

JPEG XL libjxl

Input: PNG - Quality: 90

OpenBenchmarking.orgMP/s, More Is BetterJPEG XL libjxl 0.7Input: PNG - Quality: 90Znver3 + AVX-512Znver4Znver4 + Prefer AVX-512Znver33691215SE +/- 0.07, N = 3SE +/- 0.03, N = 3SE +/- 0.04, N = 3SE +/- 0.01, N = 39.869.839.819.61-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CXX) g++ options: -O3 -flto -fno-rtti -funwind-tables -O2 -fPIE -pie -latomic

Ngspice

Circuit: C2670

OpenBenchmarking.orgSeconds, Fewer Is BetterNgspice 34Circuit: C2670Znver4Znver3Znver3 + AVX-512Znver4 + Prefer AVX-51220406080100SE +/- 0.76, N = 3SE +/- 0.09, N = 3SE +/- 0.24, N = 3SE +/- 0.17, N = 395.0795.1895.1995.60-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -O3 -flto -fopenmp -lm -lfftw3 -lXaw -lXmu -lXt -lXext -lX11 -lSM -lICE

Ngspice

Circuit: C7552

OpenBenchmarking.orgSeconds, Fewer Is BetterNgspice 34Circuit: C7552Znver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-51220406080100SE +/- 1.00, N = 3SE +/- 0.28, N = 3SE +/- 0.14, N = 3SE +/- 0.28, N = 392.1092.9493.4593.52-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -O3 -flto -fopenmp -lm -lfftw3 -lXaw -lXmu -lXt -lXext -lX11 -lSM -lICE

Zstd Compression

Compression Level: 19 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 19 - Decompression SpeedZnver3 + AVX-512Znver4Znver4 + Prefer AVX-512Znver38001600240032004000SE +/- 32.49, N = 3SE +/- 17.00, N = 6SE +/- 36.60, N = 3SE +/- 12.56, N = 153594.83584.13581.63574.9-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CC) gcc options: -O3 -flto -pthread -lz -llzma

Zstd Compression

Compression Level: 19 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 19 - Compression SpeedZnver4 + Prefer AVX-512Znver3Znver3 + AVX-512Znver420406080100SE +/- 1.30, N = 3SE +/- 0.91, N = 15SE +/- 1.27, N = 3SE +/- 1.03, N = 6105.3104.4102.9102.1-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -O3 -flto -pthread -lz -llzma

GraphicsMagick

Operation: Swirl

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: SwirlZnver4Znver3 + AVX-512Znver4 + Prefer AVX-512Znver36001200180024003000SE +/- 26.19, N = 7SE +/- 28.99, N = 3SE +/- 8.09, N = 3SE +/- 36.67, N = 32862282626812563-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

PJSIP

Method: INVITE

OpenBenchmarking.orgResponses Per Second, More Is BetterPJSIP 2.11Method: INVITEZnver4Znver3Znver3 + AVX-512Znver4 + Prefer AVX-51211002200330044005500SE +/- 19.91, N = 3SE +/- 15.00, N = 3SE +/- 24.85, N = 3SE +/- 51.36, N = 55200514951325084-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -lavformat -lavcodec -lswscale -lavutil -lstdc++ -lopus -lssl -lcrypto -lm -lrt -lpthread -lasound -O3 -flto

oneDNN

Harness: IP Shapes 1D - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: IP Shapes 1D - Data Type: bf16bf16bf16 - Engine: CPUZnver3Znver4Znver3 + AVX-512Znver4 + Prefer AVX-512612182430SE +/- 0.82, N = 15SE +/- 0.95, N = 15SE +/- 0.76, N = 15SE +/- 0.56, N = 1219.5921.1422.9423.53-march=znver3 - MIN: 9.21MIN: 9.61-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 9.35MIN: 10.061. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

PJSIP

Method: OPTIONS, Stateful

OpenBenchmarking.orgResponses Per Second, More Is BetterPJSIP 2.11Method: OPTIONS, StatefulZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver32K4K6K8K10KSE +/- 32.42, N = 3SE +/- 23.25, N = 3SE +/- 36.35, N = 3SE +/- 72.95, N = 39288923792369226-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CC) gcc options: -lavformat -lavcodec -lswscale -lavutil -lstdc++ -lopus -lssl -lcrypto -lm -lrt -lpthread -lasound -O3 -flto

libavif avifenc

Encoder Speed: 0

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.11Encoder Speed: 0Znver4 + Prefer AVX-512Znver4Znver3Znver3 + AVX-5121428425670SE +/- 0.07, N = 3SE +/- 0.13, N = 3SE +/- 0.30, N = 3SE +/- 0.02, N = 361.0761.1261.6662.68-march=native-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -fPIC -flto -lm

Stargate Digital Audio Workstation

Sample Rate: 192000 - Buffer Size: 1024

OpenBenchmarking.orgRender Ratio, More Is BetterStargate Digital Audio Workstation 22.11.5Sample Rate: 192000 - Buffer Size: 1024Znver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver40.64541.29081.93622.58163.227SE +/- 0.000776, N = 3SE +/- 0.003508, N = 3SE +/- 0.014903, N = 3SE +/- 0.009628, N = 32.8683732.8672332.8200792.7837171. (CXX) g++ options: -lpthread -lsndfile -lm -O3 -march=native -ffast-math -funroll-loops -fstrength-reduce -fstrict-aliasing -finline-functions

simdjson

Throughput Test: DistinctUserID

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 2.0Throughput Test: DistinctUserIDZnver4 + Prefer AVX-512Znver3 + AVX-512Znver4Znver3246810SE +/- 0.07, N = 3SE +/- 0.04, N = 3SE +/- 0.04, N = 3SE +/- 0.02, N = 36.866.616.536.46-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CXX) g++ options: -O3 -flto

simdjson

Throughput Test: PartialTweets

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 2.0Throughput Test: PartialTweetsZnver3Znver3 + AVX-512Znver4Znver4 + Prefer AVX-512246810SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.03, N = 3SE +/- 0.05, N = 36.746.686.636.43-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native1. (CXX) g++ options: -O3 -flto

simdjson

Throughput Test: TopTweet

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 2.0Throughput Test: TopTweetZnver4Znver4 + Prefer AVX-512Znver3 + AVX-512Znver3246810SE +/- 0.05, N = 3SE +/- 0.07, N = 3SE +/- 0.02, N = 3SE +/- 0.05, N = 36.966.916.736.61-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto

OpenSSL

Algorithm: RSA4096

OpenBenchmarking.orgverify/s, More Is BetterOpenSSL 3.0Algorithm: RSA4096Znver4Znver3Znver3 + AVX-512Znver4 + Prefer AVX-512600K1200K1800K2400K3000KSE +/- 140.73, N = 3SE +/- 228.34, N = 3SE +/- 1641.03, N = 3SE +/- 10617.65, N = 32939503.52938488.52935372.12924586.4-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -pthread -m64 -O3 -flto -lssl -lcrypto -ldl

OpenSSL

Algorithm: RSA4096

OpenBenchmarking.orgsign/s, More Is BetterOpenSSL 3.0Algorithm: RSA4096Znver3 + AVX-512Znver4Znver3Znver4 + Prefer AVX-51210K20K30K40K50KSE +/- 0.44, N = 3SE +/- 8.03, N = 3SE +/- 22.49, N = 3SE +/- 151.57, N = 344499.344490.144435.344301.8-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver3-march=native1. (CC) gcc options: -pthread -m64 -O3 -flto -lssl -lcrypto -ldl

GraphicsMagick

Operation: Sharpen

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: SharpenZnver4Znver3 + AVX-512Znver4 + Prefer AVX-512Znver330060090012001500SE +/- 6.56, N = 3SE +/- 1.15, N = 3SE +/- 13.17, N = 3SE +/- 10.73, N = 31359132113141285-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

GraphicsMagick

Operation: Rotate

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: RotateZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-512150300450600750SE +/- 1.73, N = 3SE +/- 1.00, N = 3SE +/- 1.20, N = 3SE +/- 3.84, N = 3673656645605-march=native-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

GraphicsMagick

Operation: Enhanced

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.38Operation: EnhancedZnver4Znver3 + AVX-512Znver4 + Prefer AVX-512Znver35001000150020002500SE +/- 2.00, N = 3SE +/- 14.95, N = 3SE +/- 5.00, N = 3SE +/- 19.19, N = 32234220821501837-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CC) gcc options: -fopenmp -O3 -flto -ljbig -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lz -lm -lpthread

simdjson

Throughput Test: Kostya

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 2.0Throughput Test: KostyaZnver3Znver4 + Prefer AVX-512Znver4Znver3 + AVX-5120.9451.892.8353.784.725SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 34.204.194.174.16-march=znver3-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto

JPEG XL Decoding libjxl

CPU Threads: 1

OpenBenchmarking.orgMP/s, More Is BetterJPEG XL Decoding libjxl 0.7CPU Threads: 1Znver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-5121122334455SE +/- 0.12, N = 3SE +/- 0.12, N = 3SE +/- 0.12, N = 3SE +/- 0.25, N = 348.6448.2248.2147.02

oneDNN

Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: IP Shapes 1D - Data Type: u8s8f32 - Engine: CPUZnver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver448121620SE +/- 0.42, N = 12SE +/- 0.24, N = 15SE +/- 0.34, N = 12SE +/- 0.11, N = 314.2514.2914.6214.82-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 6.05MIN: 8.38-march=znver3 - MIN: 8.56MIN: 9.811. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

simdjson

Throughput Test: LargeRandom

OpenBenchmarking.orgGB/s, More Is Bettersimdjson 2.0Throughput Test: LargeRandomZnver4Znver3 + AVX-512Znver3Znver4 + Prefer AVX-5120.28580.57160.85741.14321.429SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.271.251.251.24-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver3-march=native1. (CXX) g++ options: -O3 -flto

Kripke

OpenBenchmarking.orgThroughput FoM, More Is BetterKripke 1.2.4Znver3 + AVX-512Znver3Znver4Znver4 + Prefer AVX-51260M120M180M240M300MSE +/- 3107958.27, N = 12SE +/- 2950522.72, N = 15SE +/- 3293132.01, N = 15SE +/- 2000945.99, N = 3271735708263812847261562280254648533-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver3-march=native-march=native1. (CXX) g++ options: -O3 -flto -fopenmp

WebP Image Encode

Encode Settings: Quality 100, Lossless, Highest Compression

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.2.4Encode Settings: Quality 100, Lossless, Highest CompressionZnver3 + AVX-512Znver3Znver4 + Prefer AVX-512Znver40.13050.2610.39150.5220.6525SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 30.580.580.580.57-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -ljpeg -ltiff-march=znver3 -lgif -ltiff-march=native -ljpeg -ltiff-march=native -ljpeg1. (CC) gcc options: -fvisibility=hidden -O3 -flto -lm -lpng16

oneDNN

Harness: IP Shapes 3D - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: IP Shapes 3D - Data Type: bf16bf16bf16 - Engine: CPUZnver4Znver4 + Prefer AVX-512Znver3 + AVX-512Znver31.0182.0363.0544.0725.09SE +/- 0.04728, N = 15SE +/- 0.04436, N = 15SE +/- 0.03566, N = 9SE +/- 0.04256, N = 154.266004.333664.336584.52440MIN: 2.83MIN: 2.77-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 2.98-march=znver3 - MIN: 3.031. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

Stargate Digital Audio Workstation

Sample Rate: 96000 - Buffer Size: 1024

OpenBenchmarking.orgRender Ratio, More Is BetterStargate Digital Audio Workstation 22.11.5Sample Rate: 96000 - Buffer Size: 1024Znver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver40.9931.9862.9793.9724.965SE +/- 0.011844, N = 3SE +/- 0.013873, N = 3SE +/- 0.016378, N = 3SE +/- 0.006484, N = 34.4133794.4082904.3564464.3312301. (CXX) g++ options: -lpthread -lsndfile -lm -O3 -march=native -ffast-math -funroll-loops -fstrength-reduce -fstrict-aliasing -finline-functions

Cpuminer-Opt

Algorithm: Deepcoin

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: DeepcoinZnver3Znver4 + Prefer AVX-512Znver3 + AVX-512Znver440K80K120K160K200KSE +/- 218.28, N = 3SE +/- 1703.12, N = 5SE +/- 880.18, N = 3SE +/- 81.72, N = 3164993162242160157159147-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

SVT-AV1

Encoder Mode: Preset 4 - Input: Bosphorus 4K

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 1.4Encoder Mode: Preset 4 - Input: Bosphorus 4KZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver31.22022.44043.66064.88086.101SE +/- 0.043, N = 3SE +/- 0.015, N = 3SE +/- 0.037, N = 3SE +/- 0.012, N = 35.4235.3925.3745.360-march=znver3 -mavx512cd -mavx512vl -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -march=native -flto -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

libavif avifenc

Encoder Speed: 2

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.11Encoder Speed: 2Znver4 + Prefer AVX-512Znver3Znver4Znver3 + AVX-512816243240SE +/- 0.19, N = 3SE +/- 0.25, N = 3SE +/- 0.26, N = 3SE +/- 0.15, N = 333.9033.9134.1034.23-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -fPIC -flto -lm

Cpuminer-Opt

Algorithm: Quad SHA-256, Pyrite

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: Quad SHA-256, PyriteZnver4 + Prefer AVX-512Znver3 + AVX-512Znver4Znver3500K1000K1500K2000K2500KSE +/- 25368.27, N = 4SE +/- 22716.96, N = 3SE +/- 10306.20, N = 3SE +/- 5899.09, N = 32323995226474722510671378987-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

PJSIP

Method: OPTIONS, Stateless

OpenBenchmarking.orgResponses Per Second, More Is BetterPJSIP 2.11Method: OPTIONS, StatelessZnver3Znver4 + Prefer AVX-512Znver3 + AVX-512Znver470K140K210K280K350KSE +/- 4433.83, N = 12SE +/- 5240.59, N = 15SE +/- 3613.75, N = 15SE +/- 2531.21, N = 3336885336791336615335767-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -lavformat -lavcodec -lswscale -lavutil -lstdc++ -lopus -lssl -lcrypto -lm -lrt -lpthread -lasound -O3 -flto

Cpuminer-Opt

Algorithm: Garlicoin

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: GarlicoinZnver3 + AVX-512Znver4Znver4 + Prefer AVX-512Znver316K32K48K64K80KSE +/- 742.93, N = 3SE +/- 461.75, N = 3SE +/- 295.35, N = 3SE +/- 66.92, N = 372837724137213049523-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

Cpuminer-Opt

Algorithm: LBC, LBRY Credits

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: LBC, LBRY CreditsZnver4 + Prefer AVX-512Znver3 + AVX-512Znver4Znver3200K400K600K800K1000KSE +/- 12123.53, N = 3SE +/- 1800.04, N = 3SE +/- 829.54, N = 3SE +/- 3568.28, N = 3108582710677431065487497020-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

Cpuminer-Opt

Algorithm: scrypt

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: scryptZnver4Znver4 + Prefer AVX-512Znver3 + AVX-512Znver310002000300040005000SE +/- 0.45, N = 3SE +/- 1.22, N = 3SE +/- 1.99, N = 3SE +/- 0.45, N = 34790.114782.744763.912959.15-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

Cpuminer-Opt

Algorithm: Triple SHA-256, Onecoin

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: Triple SHA-256, OnecoinZnver3 + AVX-512Znver4Znver4 + Prefer AVX-512Znver3700K1400K2100K2800K3500KSE +/- 26057.45, N = 3SE +/- 26496.25, N = 3SE +/- 22336.69, N = 3SE +/- 4313.22, N = 33323680330664333012173255253-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

Cpuminer-Opt

Algorithm: x25x

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: x25xZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver32K4K6K8K10KSE +/- 90.39, N = 3SE +/- 74.42, N = 3SE +/- 15.17, N = 3SE +/- 17.26, N = 38217.708042.887941.386116.97-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

Cpuminer-Opt

Algorithm: Skeincoin

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: SkeincoinZnver4Znver4 + Prefer AVX-512Znver3 + AVX-512Znver3400K800K1200K1600K2000KSE +/- 10323.18, N = 3SE +/- 7108.55, N = 3SE +/- 24878.62, N = 3SE +/- 5094.00, N = 32014770200936720049901414047-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

Cpuminer-Opt

Algorithm: Magi

OpenBenchmarking.orgkH/s, More Is BetterCpuminer-Opt 3.20.3Algorithm: MagiZnver3Znver4 + Prefer AVX-512Znver4Znver3 + AVX-5122K4K6K8K10KSE +/- 51.27, N = 3SE +/- 64.85, N = 3SE +/- 47.28, N = 3SE +/- 50.94, N = 38490.738467.248440.798355.75-march=znver3-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -lcurl -lz -lpthread -lssl -lcrypto -lgmp

QuantLib

OpenBenchmarking.orgMFLOPS, More Is BetterQuantLib 1.21Znver3Znver3 + AVX-512Znver4 + Prefer AVX-512Znver47001400210028003500SE +/- 4.52, N = 3SE +/- 2.74, N = 3SE +/- 1.71, N = 3SE +/- 1.99, N = 33120.53114.93112.63096.91. (CXX) g++ options: -O3 -march=native -rdynamic

oneDNN

Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Deconvolution Batch shapes_1d - Data Type: u8s8f32 - Engine: CPUZnver3 + AVX-512Znver4Znver3Znver4 + Prefer AVX-5120.21470.42940.64410.85881.0735SE +/- 0.011238, N = 4SE +/- 0.012915, N = 3SE +/- 0.009578, N = 5SE +/- 0.009537, N = 30.9363740.9371640.9470130.954140-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.79MIN: 0.77-march=znver3 - MIN: 0.79MIN: 0.781. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

GPAW

Input: Carbon Nanotube

OpenBenchmarking.orgSeconds, Fewer Is BetterGPAW 22.1Input: Carbon NanotubeZnver3Znver4Znver4 + Prefer AVX-512Znver3 + AVX-512510152025SE +/- 0.26, N = 4SE +/- 0.15, N = 3SE +/- 0.28, N = 3SE +/- 0.07, N = 321.7222.3522.8222.82-march=znver3-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -shared -fwrapv -O2 -O3 -flto -lxc -lblas -lmpi

SVT-AV1

Encoder Mode: Preset 12 - Input: Bosphorus 4K

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 1.4Encoder Mode: Preset 12 - Input: Bosphorus 4KZnver3Znver4Znver3 + AVX-512Znver4 + Prefer AVX-51250100150200250SE +/- 5.39, N = 15SE +/- 4.58, N = 15SE +/- 3.02, N = 15SE +/- 1.78, N = 3222.90210.39206.68196.80-march=znver3-march=znver3 -mavx512cd -mavx512vl -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

JPEG XL Decoding libjxl

CPU Threads: All

OpenBenchmarking.orgMP/s, More Is BetterJPEG XL Decoding libjxl 0.7CPU Threads: AllZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-51260120180240300SE +/- 1.41, N = 3SE +/- 0.30, N = 3SE +/- 1.29, N = 3SE +/- 1.24, N = 3277.20272.31269.59266.53

Kvazaar

Video Input: Bosphorus 4K - Video Preset: Ultra Fast

OpenBenchmarking.orgFrames Per Second, More Is BetterKvazaar 2.1Video Input: Bosphorus 4K - Video Preset: Ultra FastZnver4Znver3Znver3 + AVX-512Znver4 + Prefer AVX-51220406080100SE +/- 1.01, N = 3SE +/- 0.62, N = 3SE +/- 0.64, N = 15SE +/- 0.73, N = 1578.6978.0176.7674.98-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O3 -flto -lpthread -lm -lrt

GROMACS

Implementation: MPI CPU - Input: water_GMX50_bare

OpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2022.1Implementation: MPI CPU - Input: water_GMX50_bareZnver4Znver4 + Prefer AVX-512Znver3 + AVX-512Znver3510152025SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 319.4919.4419.0918.23-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto

Coremark

CoreMark Size 666 - Iterations Per Second

OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per SecondZnver3 + AVX-512Znver4 + Prefer AVX-512Znver4Znver32M4M6M8M10MSE +/- 2823.14, N = 3SE +/- 83462.96, N = 3SE +/- 11460.97, N = 3SE +/- 62802.17, N = 37871273.937861097.867694653.907640546.27-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CC) gcc options: -O2 -O3 -flto -lrt" -lrt

oneDNN

Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Deconvolution Batch shapes_1d - Data Type: bf16bf16bf16 - Engine: CPUZnver4 + Prefer AVX-512Znver3Znver4Znver3 + AVX-5120.53141.06281.59422.12562.657SE +/- 0.00909, N = 3SE +/- 0.00532, N = 3SE +/- 0.00839, N = 3SE +/- 0.00584, N = 32.298802.318582.333632.36166MIN: 1.92-march=znver3 - MIN: 1.91MIN: 1.95-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 1.921. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

SVT-AV1

Encoder Mode: Preset 13 - Input: Bosphorus 4K

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 1.4Encoder Mode: Preset 13 - Input: Bosphorus 4KZnver3Znver3 + AVX-512Znver4 + Prefer AVX-512Znver450100150200250SE +/- 4.56, N = 12SE +/- 2.81, N = 15SE +/- 3.20, N = 15SE +/- 3.33, N = 15219.57210.43208.64196.12-march=znver3-march=znver3 -mavx512cd -mavx512vl -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

Liquid-DSP

Threads: 384 - Buffer Length: 256 - Filter Length: 57

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 384 - Buffer Length: 256 - Filter Length: 57Znver3 + AVX-512Znver4 + Prefer AVX-512Znver4Znver32000M4000M6000M8000M10000MSE +/- 2403700.85, N = 3SE +/- 7218802.61, N = 3SE +/- 7356025.50, N = 3SE +/- 10066445.91, N = 311301666667112706666671124933333311176000000-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CC) gcc options: -O3 -flto -pthread -lm -lc -lliquid

Liquid-DSP

Threads: 256 - Buffer Length: 256 - Filter Length: 57

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 256 - Buffer Length: 256 - Filter Length: 57Znver3 + AVX-512Znver4 + Prefer AVX-512Znver4Znver32000M4000M6000M8000M10000MSE +/- 4697162.26, N = 3SE +/- 8533658.85, N = 3SE +/- 5651843.36, N = 3SE +/- 24004606.04, N = 39813700000980980000097895000009735666667-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CC) gcc options: -O3 -flto -pthread -lm -lc -lliquid

Liquid-DSP

Threads: 128 - Buffer Length: 256 - Filter Length: 57

OpenBenchmarking.orgsamples/s, More Is BetterLiquid-DSP 2021.01.31Threads: 128 - Buffer Length: 256 - Filter Length: 57Znver3 + AVX-512Znver4Znver4 + Prefer AVX-512Znver31500M3000M4500M6000M7500MSE +/- 4658445.14, N = 3SE +/- 3192874.01, N = 3SE +/- 6145549.43, N = 3SE +/- 6548367.06, N = 36999233333699086666769776333336940833333-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=native-march=znver31. (CC) gcc options: -O3 -flto -pthread -lm -lc -lliquid

miniBUDE

Implementation: OpenMP - Input Deck: BM2

OpenBenchmarking.orgBillion Interactions/s, More Is BetterminiBUDE 20210901Implementation: OpenMP - Input Deck: BM2Znver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver460120180240300SE +/- 1.06, N = 3SE +/- 0.73, N = 3SE +/- 0.73, N = 3SE +/- 1.75, N = 3266.16265.52264.28264.151. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm

miniBUDE

Implementation: OpenMP - Input Deck: BM2

OpenBenchmarking.orgGFInst/s, More Is BetterminiBUDE 20210901Implementation: OpenMP - Input Deck: BM2Znver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver414002800420056007000SE +/- 26.43, N = 3SE +/- 18.27, N = 3SE +/- 18.35, N = 3SE +/- 43.78, N = 36653.866638.016607.096603.681. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm

WebP Image Encode

Encode Settings: Quality 100, Lossless

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.2.4Encode Settings: Quality 100, LosslessZnver3Znver4 + Prefer AVX-512Znver3 + AVX-512Znver40.33080.66160.99241.32321.654SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 31.471.471.451.45-march=znver3 -lgif -ltiff-march=native -ljpeg -ltiff-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -ljpeg -ltiff-march=native -ljpeg1. (CC) gcc options: -fvisibility=hidden -O3 -flto -lpng16 -lm

SMHasher

Hash: MeowHash x86_64 AES-NI

OpenBenchmarking.orgcycles/hash, Fewer Is BetterSMHasher 2022-08-22Hash: MeowHash x86_64 AES-NIZnver4 + Prefer AVX-512Znver3Znver3 + AVX-512Znver41020304050SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 344.9544.9644.9744.98-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -march=native -flto -flto=auto -fno-fat-lto-objects

SMHasher

Hash: MeowHash x86_64 AES-NI

OpenBenchmarking.orgMiB/sec, More Is BetterSMHasher 2022-08-22Hash: MeowHash x86_64 AES-NIZnver4 + Prefer AVX-512Znver3 + AVX-512Znver3Znver412K24K36K48K60KSE +/- 9.47, N = 3SE +/- 10.52, N = 3SE +/- 9.98, N = 3SE +/- 7.42, N = 354297.0454284.6554281.9454272.14-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -march=native -flto -flto=auto -fno-fat-lto-objects

oneDNN

Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Convolution Batch Shapes Auto - Data Type: bf16bf16bf16 - Engine: CPUZnver3 + AVX-512Znver3Znver4Znver4 + Prefer AVX-5120.10020.20040.30060.40080.501SE +/- 0.004441, N = 15SE +/- 0.002906, N = 3SE +/- 0.001884, N = 3SE +/- 0.000431, N = 30.4312990.4427480.4438910.445532-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.33-march=znver3 - MIN: 0.34MIN: 0.37MIN: 0.341. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

oneDNN

Harness: Matrix Multiply Batch Shapes Transformer - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Matrix Multiply Batch Shapes Transformer - Data Type: bf16bf16bf16 - Engine: CPUZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver30.1150.230.3450.460.575SE +/- 0.003087, N = 3SE +/- 0.004211, N = 3SE +/- 0.006958, N = 3SE +/- 0.006281, N = 30.4839790.5005790.5048420.510939MIN: 0.39MIN: 0.39-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.39-march=znver3 - MIN: 0.391. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

SVT-AV1

Encoder Mode: Preset 8 - Input: Bosphorus 4K

OpenBenchmarking.orgFrames Per Second, More Is BetterSVT-AV1 1.4Encoder Mode: Preset 8 - Input: Bosphorus 4KZnver3Znver4 + Prefer AVX-512Znver4Znver3 + AVX-51220406080100SE +/- 0.46, N = 3SE +/- 0.32, N = 3SE +/- 0.33, N = 3SE +/- 0.52, N = 395.1294.3493.7692.96-march=znver3-march=znver3 -mavx512cd -mavx512vl -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -march=native -mno-avx -mavx2 -mavx512f -mavx512bw -mavx512dq

Kvazaar

Video Input: Bosphorus 4K - Video Preset: Medium

OpenBenchmarking.orgFrames Per Second, More Is BetterKvazaar 2.1Video Input: Bosphorus 4K - Video Preset: MediumZnver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-5121428425670SE +/- 0.52, N = 3SE +/- 0.36, N = 3SE +/- 0.46, N = 3SE +/- 0.10, N = 364.3564.2764.0563.80-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O3 -flto -lpthread -lm -lrt

Kvazaar

Video Input: Bosphorus 4K - Video Preset: Very Fast

OpenBenchmarking.orgFrames Per Second, More Is BetterKvazaar 2.1Video Input: Bosphorus 4K - Video Preset: Very FastZnver3Znver4 + Prefer AVX-512Znver3 + AVX-512Znver41632486480SE +/- 0.28, N = 3SE +/- 0.92, N = 3SE +/- 0.87, N = 4SE +/- 0.83, N = 372.0170.8170.7170.71-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native1. (CC) gcc options: -pthread -ftree-vectorize -fvisibility=hidden -O3 -flto -lpthread -lm -lrt

oneDNN

Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: IP Shapes 3D - Data Type: u8s8f32 - Engine: CPUZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver30.20430.40860.61290.81721.0215SE +/- 0.004696, N = 3SE +/- 0.004516, N = 3SE +/- 0.002136, N = 3SE +/- 0.011617, N = 30.8813030.8866420.9023030.907941MIN: 0.74MIN: 0.76-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.75-march=znver3 - MIN: 0.751. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

SMHasher

Hash: FarmHash32 x86_64 AVX

OpenBenchmarking.orgcycles/hash, Fewer Is BetterSMHasher 2022-08-22Hash: FarmHash32 x86_64 AVXZnver4Znver3Znver3 + AVX-512Znver4 + Prefer AVX-512612182430SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 326.4926.4926.4926.50-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -march=native -flto -flto=auto -fno-fat-lto-objects

SMHasher

Hash: FarmHash32 x86_64 AVX

OpenBenchmarking.orgMiB/sec, More Is BetterSMHasher 2022-08-22Hash: FarmHash32 x86_64 AVXZnver3Znver4Znver4 + Prefer AVX-512Znver3 + AVX-5129K18K27K36K45KSE +/- 0.13, N = 3SE +/- 0.95, N = 3SE +/- 1.44, N = 3SE +/- 1.84, N = 340565.7240565.0740563.9640559.33-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -flto -march=native -flto=auto -fno-fat-lto-objects

oneDNN

Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Deconvolution Batch shapes_3d - Data Type: u8s8f32 - Engine: CPUZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver30.06290.12580.18870.25160.3145SE +/- 0.001864, N = 3SE +/- 0.001957, N = 12SE +/- 0.003530, N = 3SE +/- 0.002168, N = 150.2746620.2747690.2769650.279655MIN: 0.24MIN: 0.24-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.24-march=znver3 - MIN: 0.231. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

ASTC Encoder

Preset: Medium

OpenBenchmarking.orgMT/s, More Is BetterASTC Encoder 4.0Preset: MediumZnver3 + AVX-512Znver4Znver3Znver4 + Prefer AVX-512110220330440550SE +/- 5.68, N = 3SE +/- 3.36, N = 13SE +/- 6.17, N = 3SE +/- 0.56, N = 3511.52493.22459.69420.70-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver3-march=native1. (CXX) g++ options: -O3 -flto -pthread

ASTC Encoder

Preset: Exhaustive

OpenBenchmarking.orgMT/s, More Is BetterASTC Encoder 4.0Preset: ExhaustiveZnver4 + Prefer AVX-512Znver3 + AVX-512Znver4Znver33691215SE +/- 0.05, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.03, N = 313.0913.0312.9412.22-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CXX) g++ options: -O3 -flto -pthread

WebP Image Encode

Encode Settings: Quality 100, Highest Compression

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.2.4Encode Settings: Quality 100, Highest CompressionZnver3Znver4Znver4 + Prefer AVX-512Znver3 + AVX-5120.8191.6382.4573.2764.095SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 33.643.253.233.11-march=znver3 -lgif -ltiff-march=native -ljpeg-march=native -ljpeg -ltiff-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -ljpeg -ltiff1. (CC) gcc options: -fvisibility=hidden -O3 -flto -lpng16 -lm

SMHasher

Hash: t1ha0_aes_avx2 x86_64

OpenBenchmarking.orgcycles/hash, Fewer Is BetterSMHasher 2022-08-22Hash: t1ha0_aes_avx2 x86_64Znver4Znver3 + AVX-512Znver3Znver4 + Prefer AVX-512510152025SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 320.5220.5320.5320.81-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -march=native -flto -flto=auto -fno-fat-lto-objects

SMHasher

Hash: t1ha0_aes_avx2 x86_64

OpenBenchmarking.orgMiB/sec, More Is BetterSMHasher 2022-08-22Hash: t1ha0_aes_avx2 x86_64Znver3 + AVX-512Znver4 + Prefer AVX-512Znver4Znver320K40K60K80K100KSE +/- 19.58, N = 3SE +/- 11.55, N = 3SE +/- 20.00, N = 3SE +/- 8.85, N = 3102403.92102399.74102354.57102351.87-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto -march=native -flto=auto -fno-fat-lto-objects

oneDNN

Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Convolution Batch Shapes Auto - Data Type: u8s8f32 - Engine: CPUZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-5120.08840.17680.26520.35360.442SE +/- 0.000892, N = 3SE +/- 0.003460, N = 3SE +/- 0.000630, N = 3SE +/- 0.004423, N = 30.3804860.3884830.3927650.392875MIN: 0.28MIN: 0.28-march=znver3 - MIN: 0.28-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.281. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

ASTC Encoder

Preset: Thorough

OpenBenchmarking.orgMT/s, More Is BetterASTC Encoder 4.0Preset: ThoroughZnver4 + Prefer AVX-512Znver4Znver3 + AVX-512Znver3306090120150SE +/- 0.08, N = 3SE +/- 0.12, N = 3SE +/- 0.06, N = 3SE +/- 0.10, N = 3118.99118.74117.69116.09-march=native-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=znver31. (CXX) g++ options: -O3 -flto -pthread

libavif avifenc

Encoder Speed: 6, Lossless

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.11Encoder Speed: 6, LosslessZnver4Znver3Znver4 + Prefer AVX-512Znver3 + AVX-5121.01572.03143.04714.06285.0785SE +/- 0.052, N = 4SE +/- 0.027, N = 3SE +/- 0.064, N = 3SE +/- 0.010, N = 34.3984.4374.4624.514-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -fPIC -flto -lm

miniBUDE

Implementation: OpenMP - Input Deck: BM1

OpenBenchmarking.orgBillion Interactions/s, More Is BetterminiBUDE 20210901Implementation: OpenMP - Input Deck: BM1Znver3 + AVX-512Znver4Znver3Znver4 + Prefer AVX-51250100150200250SE +/- 2.13, N = 3SE +/- 2.07, N = 3SE +/- 2.52, N = 3SE +/- 2.51, N = 4214.63214.50214.08211.021. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm

miniBUDE

Implementation: OpenMP - Input Deck: BM1

OpenBenchmarking.orgGFInst/s, More Is BetterminiBUDE 20210901Implementation: OpenMP - Input Deck: BM1Znver3 + AVX-512Znver4Znver3Znver4 + Prefer AVX-51212002400360048006000SE +/- 53.26, N = 3SE +/- 51.63, N = 3SE +/- 62.89, N = 3SE +/- 62.65, N = 45365.615362.545351.945275.431. (CC) gcc options: -std=c99 -Ofast -ffast-math -fopenmp -march=native -lm

libavif avifenc

Encoder Speed: 10, Lossless

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.11Encoder Speed: 10, LosslessZnver4Znver4 + Prefer AVX-512Znver3Znver3 + AVX-5120.82061.64122.46183.28244.103SE +/- 0.012, N = 3SE +/- 0.017, N = 3SE +/- 0.016, N = 3SE +/- 0.038, N = 33.5413.5723.5813.647-march=native-march=native-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -fPIC -flto -lm

ACES DGEMM

Sustained Floating-Point Rate

OpenBenchmarking.orgGFLOP/s, More Is BetterACES DGEMM 1.0Sustained Floating-Point RateZnver3Znver4 + Prefer AVX-512Znver3 + AVX-512Znver41632486480SE +/- 0.06, N = 3SE +/- 0.07, N = 3SE +/- 0.28, N = 3SE +/- 0.10, N = 370.3870.3070.1970.05-march=znver3-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CC) gcc options: -O3 -march=native -fopenmp -flto

oneDNN

Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPU

OpenBenchmarking.orgms, Fewer Is BetteroneDNN 3.0Harness: Deconvolution Batch shapes_3d - Data Type: bf16bf16bf16 - Engine: CPUZnver3 + AVX-512Znver4 + Prefer AVX-512Znver3Znver40.14810.29620.44430.59240.7405SE +/- 0.003927, N = 3SE +/- 0.001081, N = 3SE +/- 0.006925, N = 4SE +/- 0.005902, N = 30.6470870.6500670.6563420.658101-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi - MIN: 0.56MIN: 0.53-march=znver3 - MIN: 0.54MIN: 0.531. (CXX) g++ options: -O3 -march=native -flto -fopenmp -msse4.1 -fPIC -pie -ldl -lpthread

LAMMPS Molecular Dynamics Simulator

Model: Rhodopsin Protein

OpenBenchmarking.orgns/day, More Is BetterLAMMPS Molecular Dynamics Simulator 23Jun2022Model: Rhodopsin ProteinZnver4 + Prefer AVX-512Znver3 + AVX-512Znver4Znver31224364860SE +/- 0.48, N = 3SE +/- 0.40, N = 10SE +/- 0.21, N = 3SE +/- 0.43, N = 352.2951.8251.5951.46-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi-march=native-march=znver31. (CXX) g++ options: -O3 -flto -lm -ldl

libavif avifenc

Encoder Speed: 6

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 0.11Encoder Speed: 6Znver4 + Prefer AVX-512Znver3Znver4Znver3 + AVX-5120.54141.08281.62422.16562.707SE +/- 0.002, N = 3SE +/- 0.010, N = 3SE +/- 0.014, N = 3SE +/- 0.005, N = 32.3172.3312.3472.406-march=native-march=znver3-march=native-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi1. (CXX) g++ options: -O3 -fPIC -flto -lm

WebP Image Encode

Encode Settings: Quality 100

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.2.4Encode Settings: Quality 100Znver4 + Prefer AVX-512Znver4Znver3Znver3 + AVX-5123691215SE +/- 0.01, N = 3SE +/- 0.03, N = 3SE +/- 0.01, N = 3SE +/- 0.01, N = 311.5411.5411.4811.38-march=native -ljpeg -ltiff-march=native -ljpeg-march=znver3 -lgif -ltiff-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -ljpeg -ltiff1. (CC) gcc options: -fvisibility=hidden -O3 -flto -lm -lpng16

WebP Image Encode

Encode Settings: Default

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.2.4Encode Settings: DefaultZnver3Znver4 + Prefer AVX-512Znver4Znver3 + AVX-512510152025SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.07, N = 318.9918.9718.9518.85-march=znver3 -lgif -ltiff-march=native -ljpeg -ltiff-march=native -ljpeg-march=znver3 -mavx512f -mavx512cd -mavx512vl -mavx512bw -mavx512dq -mavx512ifma -mavx512vbmi -ljpeg -ltiff1. (CC) gcc options: -fvisibility=hidden -O3 -flto -lpng16 -lm

Phoronix Test Suite v10.8.5