AWS Graviton Benchmarks

Benchmarks by Michael Larabel.

HTML result view exported from: https://openbenchmarking.org/result/2407235-NE-2407230NE75&grs&sor.

AWS Graviton BenchmarksProcessorMotherboardMemoryDiskNetworkOSKernelCompilerFile-SystemGraviton1 16 Cores a1.metalGraviton4 96 Cores r8g.metal-24xlARMv8 Cortex-A72 (16 Cores)Amazon EC2 a1.metal (1.0 BIOS)32GB429GB Amazon Elastic Block StoreAmazon ElasticUbuntu 24.046.8.0-1009-aws (aarch64)GCC 13.2.0ext4ARMv8 Neoverse-V2 (96 Cores)Amazon EC2 r8g.metal-24xl (1.0 BIOS)12 x 64GB DDR5-5600MT/sOpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseCompiler Details- --build=aarch64-linux-gnu --disable-libquadmath --disable-libquadmath-support --disable-werror --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-fix-cortex-a53-843419 --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-nls --enable-objc-gc=auto --enable-offload-defaulted --enable-offload-targets=nvptx-none=/build/gcc-13-dIwDw0/gcc-13-13.2.0/debian/tmp-nvptx/usr --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --program-prefix=aarch64-linux-gnu- --target=aarch64-linux-gnu --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-target-system-zlib=auto --without-cuda-driver -v Java Details- OpenJDK Runtime Environment (build 11.0.23+9-post-Ubuntu-1ubuntu1)Python Details- Python 3.12.3Security Details- Graviton1 16 Cores a1.metal: 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: Not affected + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Branch predictor hardening BHB + srbds: Not affected + tsx_async_abort: Not affected - Graviton4 96 Cores r8g.metal-24xl: 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: Mitigation of SSB disabled via prctl + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Not affected + srbds: Not affected + tsx_async_abort: Not affected

AWS Graviton Benchmarksspeedb: Read Rand Write Randrocksdb: Read While Writingcassandra: Writesstress-ng: Matrix Mathrocksdb: Update Randgraphics-magick: Noise-Gaussianavifenc: 6stress-ng: CPU Cachemt-dgemm: Sustained Floating-Point Ratepyperformance: pathlibspeedb: Update Randavifenc: 6, Losslessbuild-php: Time To Compileavifenc: 0avifenc: 2numpy: speedb: Read While Writinggraphics-magick: Resizingpyperformance: regex_compilestress-ng: Vector Floating Pointcompress-lz4: 2 - Compression Speedpyperformance: xml_etreegraphics-magick: HWB Color Spacesrsran: PUSCH Processor Benchmark, Throughput Threadphpbench: PHP Benchmark Suitepyperformance: json_loadspyperformance: crypto_pyaespyperformance: python_startupwebp: Quality 100, Losslesscompress-lz4: 12 - Decompression Speedcompress-lz4: 9 - Decompression Speedcompress-lz4: 3 - Compression Speedcompress-lz4: 3 - Decompression Speedc-ray: 4K - 16c-ray: 5K - 16c-ray: 1080p - 16pgbench: 100 - 1000 - Read Onlypgbench: 100 - 1000 - Read Only - Average Latencyspeedb: Rand Readstress-ng: Matrix 3D Mathrocksdb: Rand Readgromacs: MPI CPU - water_GMX50_bareopenssl: SHA512srsran: PDSCH Processor Benchmark, Throughput Threadcompress-lz4: 1 - Compression Speedcompress-lz4: 9 - Compression Speedcompress-lz4: 2 - Decompression Speedcompress-lz4: 1 - Decompression Speedwebp: Defaultcompress-lz4: 12 - Compression Speedwebp: Quality 100webp: Quality 100, Highest Compressionblender: Classroom - CPU-Onlyopenssl: ChaCha20blender: BMW27 - CPU-Onlyminife: Smallopenssl: ChaCha20-Poly1305srsran: PDSCH Processor Benchmark, Throughput Totalhpcg: 104 104 104 - 60stress-ng: Memory Copyingbuild-nodejs: Time To Compileopenssl: AES-256-GCMstress-ng: Vector Mathjohn-the-ripper: HMAC-SHA512stress-ng: Trigonometric Mathstress-ng: Logarithmic Mathblender: Fishy Cat - CPU-Onlygraphics-magick: Enhancedcompress-7zip: Compression Ratingopenssl: AES-128-GCMstress-ng: Floating Pointstress-ng: Power Mathcoremark: CoreMark Size 666 - Iterations Per Secondopenssl: SHA256stress-ng: Fused Multiply-Addbuild-godot: Time To Compilegraphics-magick: Sharpenjohn-the-ripper: MD5rocksdb: Read Rand Write Randcompress-7zip: Decompression Ratingsrsran: PUSCH Processor Benchmark, Throughput Totaljohn-the-ripper: WPA PSKjohn-the-ripper: bcryptjohn-the-ripper: Blowfishpgbench: 100 - 1000 - Read Write - Average Latencypgbench: 100 - 1000 - Read Writestockfish: Chess BenchmarkGraviton1 16 Cores a1.metalGraviton4 96 Cores r8g.metal-24xl5526211268650263487343.991646162621.288631797.130.87925011012816728.229479.642668.499381.112104.592433641614553466.2582.512308118.723421172.122725.90.401090.31102.037.301137.3566.9461008.693140.8221018709.81623286630838.94232528210.313250589187794.0237.6813.721153.11514.74.634.833.251.431289.618285450333599.394159.1265541467701307.33.768671997.894663.2762577349645728508.49103216672998.753302.68952.43203269730455741337852.81609.16192273.03371561517499334928995.121674.4692619042657445841244169.3789478428003215.89046322493057534291111560819239658631796.6513274232082.7774735225.1760.63583917.67293985.23789.999130.20380.243495.7811490911278101119727.57329.4158.230169.482922120.665.07.551.333595.53573.1115.143489.518.92733.6784.89128932640.34664093847122886.625274585476.92653203418283278.7697.8939.803326.13927.711.6111.957.823.2269.4014879168876033.7270677.811069945000322062.761.619532525.12289.267400130558450438910.4015823833345810.3750252.6664.0729447985144487114879712455.0223418.402732642.2688918560484677065585333.04126.346325236733370471014946601961.7862818564785667232.4334303118337918OpenBenchmarking.org

Speedb

Test: Read Random Write Random

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Read Random Write RandomGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1.1M2.2M3.3M4.4M5.5MSE +/- 4241.50, N = 3SE +/- 214.78, N = 353429115526211. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

RocksDB

Test: Read While Writing

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Read While WritingGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal2M4M6M8M10MSE +/- 18723.10, N = 3SE +/- 16876.08, N = 31156081912686501. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

Apache Cassandra

Test: Writes

OpenBenchmarking.orgOp/s, More Is BetterApache Cassandra 4.1.3Test: WritesGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal50K100K150K200K250KSE +/- 835.72, N = 3SE +/- 148.76, N = 323965826348

Stress-NG

Test: Matrix Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Matrix MathGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal140K280K420K560K700KSE +/- 9.55, N = 3SE +/- 74.42, N = 3631796.657343.991. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

RocksDB

Test: Update Random

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Update RandomGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal300K600K900K1200K1500KSE +/- 5313.65, N = 3SE +/- 443.99, N = 313274231646161. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

GraphicsMagick

Operation: Noise-Gaussian

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.43Operation: Noise-GaussianGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal50100150200250SE +/- 0.67, N = 3SE +/- 0.00, N = 3208261. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp

libavif avifenc

Encoder Speed: 6

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 1.0Encoder Speed: 6Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal510152025SE +/- 0.014, N = 3SE +/- 0.032, N = 32.77721.2881. (CXX) g++ options: -O3 -fPIC -lm

Stress-NG

Test: CPU Cache

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: CPU CacheGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1000K2000K3000K4000K5000KSE +/- 57877.80, N = 15SE +/- 3895.03, N = 34735225.17631797.131. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

ACES DGEMM

Sustained Floating-Point Rate

OpenBenchmarking.orgGFLOP/s, More Is BetterACES DGEMM 1.0Sustained Floating-Point RateGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1428425670SE +/- 0.164173, N = 3SE +/- 0.003641, N = 360.6358390.8792501. (CC) gcc options: -O3 -march=native -fopenmp

PyPerformance

Benchmark: pathlib

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.11Benchmark: pathlibGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal20406080100SE +/- 0.00, N = 3SE +/- 0.00, N = 317.6110.0

Speedb

Test: Update Random

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Update RandomGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal160K320K480K640K800KSE +/- 3844.86, N = 3SE +/- 302.33, N = 37293981281671. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

libavif avifenc

Encoder Speed: 6, Lossless

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 1.0Encoder Speed: 6, LosslessGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal714212835SE +/- 0.046, N = 3SE +/- 0.098, N = 35.23728.2291. (CXX) g++ options: -O3 -fPIC -lm

Timed PHP Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed PHP Compilation 8.3.4Time To CompileGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal100200300400500SE +/- 0.10, N = 3SE +/- 0.12, N = 390.00479.64

libavif avifenc

Encoder Speed: 0

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 1.0Encoder Speed: 0Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal140280420560700SE +/- 0.01, N = 3SE +/- 3.02, N = 3130.20668.501. (CXX) g++ options: -O3 -fPIC -lm

libavif avifenc

Encoder Speed: 2

OpenBenchmarking.orgSeconds, Fewer Is Betterlibavif avifenc 1.0Encoder Speed: 2Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal80160240320400SE +/- 0.02, N = 3SE +/- 0.85, N = 380.24381.111. (CXX) g++ options: -O3 -fPIC -lm

Numpy Benchmark

OpenBenchmarking.orgScore, More Is BetterNumpy BenchmarkGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal110220330440550SE +/- 0.76, N = 3SE +/- 0.23, N = 3495.78104.59

Speedb

Test: Read While Writing

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Read While WritingGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal2M4M6M8M10MSE +/- 87574.70, N = 15SE +/- 20302.77, N = 81149091124336411. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

GraphicsMagick

Operation: Resizing

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.43Operation: ResizingGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal60120180240300SE +/- 0.58, N = 3SE +/- 0.00, N = 3278611. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp

PyPerformance

Benchmark: regex_compile

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.11Benchmark: regex_compileGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal100200300400500SE +/- 0.00, N = 3SE +/- 0.33, N = 3101455

Stress-NG

Test: Vector Floating Point

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Vector Floating PointGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal30K60K90K120K150KSE +/- 17.26, N = 3SE +/- 0.38, N = 3119727.573466.251. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

LZ4 Compression

Compression Level: 2 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 2 - Compression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal70140210280350SE +/- 0.13, N = 3SE +/- 0.02, N = 3329.4182.511. (CC) gcc options: -O3 -pthread

PyPerformance

Benchmark: xml_etree

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.11Benchmark: xml_etreeGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal50100150200250SE +/- 0.00, N = 3SE +/- 0.58, N = 358.2230.0

GraphicsMagick

Operation: HWB Color Space

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.43Operation: HWB Color SpaceGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal70140210280350SE +/- 0.33, N = 3SE +/- 0.00, N = 3301811. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp

srsRAN Project

Test: PUSCH Processor Benchmark, Throughput Thread

OpenBenchmarking.orgMbps, More Is BettersrsRAN Project 23.10.1-20240325Test: PUSCH Processor Benchmark, Throughput ThreadGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1530456075SE +/- 0.00, N = 3SE +/- 0.23, N = 469.418.7MIN: 45.4MIN: 11.3 / MAX: 18.91. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl

PHPBench

PHP Benchmark Suite

OpenBenchmarking.orgScore, More Is BetterPHPBench 0.8.1PHP Benchmark SuiteGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal200K400K600K800K1000KSE +/- 273.39, N = 3SE +/- 481.21, N = 3829221234211

PyPerformance

Benchmark: json_loads

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.11Benchmark: json_loadsGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1632486480SE +/- 0.00, N = 3SE +/- 0.10, N = 320.672.1

PyPerformance

Benchmark: crypto_pyaes

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.11Benchmark: crypto_pyaesGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal50100150200250SE +/- 0.03, N = 3SE +/- 0.33, N = 365.0227.0

PyPerformance

Benchmark: python_startup

OpenBenchmarking.orgMilliseconds, Fewer Is BetterPyPerformance 1.11Benchmark: python_startupGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal612182430SE +/- 0.00, N = 3SE +/- 0.00, N = 37.5525.90

WebP Image Encode

Encode Settings: Quality 100, Lossless

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.4Encode Settings: Quality 100, LosslessGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal0.29930.59860.89791.19721.4965SE +/- 0.00, N = 3SE +/- 0.00, N = 31.330.401. (CC) gcc options: -fvisibility=hidden -O2 -lm

LZ4 Compression

Compression Level: 12 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 12 - Decompression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal8001600240032004000SE +/- 1.89, N = 3SE +/- 2.61, N = 33595.51090.31. (CC) gcc options: -O3 -pthread

LZ4 Compression

Compression Level: 9 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 9 - Decompression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal8001600240032004000SE +/- 0.40, N = 3SE +/- 0.06, N = 33573.11102.01. (CC) gcc options: -O3 -pthread

LZ4 Compression

Compression Level: 3 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 3 - Compression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal306090120150SE +/- 0.08, N = 3SE +/- 0.01, N = 3115.1437.301. (CC) gcc options: -O3 -pthread

LZ4 Compression

Compression Level: 3 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 3 - Decompression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal7001400210028003500SE +/- 0.15, N = 3SE +/- 0.06, N = 33489.51137.31. (CC) gcc options: -O3 -pthread

C-Ray

Resolution: 4K - Rays Per Pixel: 16

OpenBenchmarking.orgSeconds, Fewer Is BetterC-Ray 2.0Resolution: 4K - Rays Per Pixel: 16Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal120240360480600SE +/- 0.02, N = 3SE +/- 1.65, N = 318.93566.951. (CC) gcc options: -lpthread -lm

C-Ray

Resolution: 5K - Rays Per Pixel: 16

OpenBenchmarking.orgSeconds, Fewer Is BetterC-Ray 2.0Resolution: 5K - Rays Per Pixel: 16Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal2004006008001000SE +/- 0.03, N = 3SE +/- 2.98, N = 333.681008.691. (CC) gcc options: -lpthread -lm

C-Ray

Resolution: 1080p - Rays Per Pixel: 16

OpenBenchmarking.orgSeconds, Fewer Is BetterC-Ray 2.0Resolution: 1080p - Rays Per Pixel: 16Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal306090120150SE +/- 0.007, N = 3SE +/- 0.248, N = 34.891140.8221. (CC) gcc options: -lpthread -lm

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Only

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL 16Scaling Factor: 100 - Clients: 1000 - Mode: Read OnlyGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal600K1200K1800K2400K3000KSE +/- 15759.90, N = 3SE +/- 107.42, N = 328932641018701. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Only - Average Latency

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL 16Scaling Factor: 100 - Clients: 1000 - Mode: Read Only - Average LatencyGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal3691215SE +/- 0.002, N = 3SE +/- 0.010, N = 30.3469.8161. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm

Speedb

Test: Random Read

OpenBenchmarking.orgOp/s, More Is BetterSpeedb 2.7Test: Random ReadGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal140M280M420M560M700MSE +/- 229391.69, N = 3SE +/- 174965.25, N = 15640938471232866301. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

Stress-NG

Test: Matrix 3D Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Matrix 3D MathGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal5K10K15K20K25KSE +/- 50.98, N = 3SE +/- 3.67, N = 322886.62838.941. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

RocksDB

Test: Random Read

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Random ReadGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal110M220M330M440M550MSE +/- 53499.62, N = 3SE +/- 12274.71, N = 3527458547232528211. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

GROMACS

Implementation: MPI CPU - Input: water_GMX50_bare

OpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2024Implementation: MPI CPU - Input: water_GMX50_bareGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal246810SE +/- 0.001, N = 3SE +/- 0.000, N = 36.9260.3131. (CXX) g++ options: -O3 -lm

OpenSSL

Algorithm: SHA512

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: SHA512Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal11000M22000M33000M44000M55000MSE +/- 4634061.90, N = 3SE +/- 4619831.10, N = 35320341828325058918771. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)

srsRAN Project

Test: PDSCH Processor Benchmark, Throughput Thread

OpenBenchmarking.orgMbps, More Is BettersrsRAN Project 23.10.1-20240325Test: PDSCH Processor Benchmark, Throughput ThreadGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal60120180240300SE +/- 3.00, N = 3SE +/- 0.21, N = 3278.794.01. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl

LZ4 Compression

Compression Level: 1 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 1 - Compression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal150300450600750SE +/- 0.04, N = 3SE +/- 3.13, N = 3697.89237.681. (CC) gcc options: -O3 -pthread

LZ4 Compression

Compression Level: 9 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 9 - Compression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal918273645SE +/- 0.01, N = 3SE +/- 0.01, N = 339.8013.721. (CC) gcc options: -O3 -pthread

LZ4 Compression

Compression Level: 2 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 2 - Decompression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal7001400210028003500SE +/- 0.13, N = 3SE +/- 0.27, N = 33326.11153.11. (CC) gcc options: -O3 -pthread

LZ4 Compression

Compression Level: 1 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 1 - Decompression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal8001600240032004000SE +/- 0.19, N = 3SE +/- 0.30, N = 33927.71514.71. (CC) gcc options: -O3 -pthread

WebP Image Encode

Encode Settings: Default

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.4Encode Settings: DefaultGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal3691215SE +/- 0.02, N = 3SE +/- 0.01, N = 311.614.631. (CC) gcc options: -fvisibility=hidden -O2 -lm

LZ4 Compression

Compression Level: 12 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.10Compression Level: 12 - Compression SpeedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal3691215SE +/- 0.00, N = 3SE +/- 0.00, N = 311.954.831. (CC) gcc options: -O3 -pthread

WebP Image Encode

Encode Settings: Quality 100

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.4Encode Settings: Quality 100Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal246810SE +/- 0.01, N = 3SE +/- 0.01, N = 37.823.251. (CC) gcc options: -fvisibility=hidden -O2 -lm

WebP Image Encode

Encode Settings: Quality 100, Highest Compression

OpenBenchmarking.orgMP/s, More Is BetterWebP Image Encode 1.4Encode Settings: Quality 100, Highest CompressionGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal0.72451.4492.17352.8983.6225SE +/- 0.00, N = 3SE +/- 0.01, N = 33.221.431. (CC) gcc options: -fvisibility=hidden -O2 -lm

Blender

Blend File: Classroom - Compute: CPU-Only

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.0.2Blend File: Classroom - Compute: CPU-OnlyGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal30060090012001500SE +/- 0.03, N = 3SE +/- 0.28, N = 369.401289.61

OpenSSL

Algorithm: ChaCha20

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: ChaCha20Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal30000M60000M90000M120000M150000MSE +/- 430328.22, N = 3SE +/- 4780384.74, N = 314879168876082854503331. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)

Blender

Blend File: BMW27 - Compute: CPU-Only

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.0.2Blend File: BMW27 - Compute: CPU-OnlyGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal130260390520650SE +/- 0.03, N = 3SE +/- 0.63, N = 333.72599.39

miniFE

Problem Size: Small

OpenBenchmarking.orgCG Mflops, More Is BetterminiFE 2.2Problem Size: SmallGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal15K30K45K60K75KSE +/- 233.77, N = 3SE +/- 1.93, N = 370677.804159.121. (CXX) g++ options: -O3 -fopenmp -lmpi_cxx -lmpi

OpenSSL

Algorithm: ChaCha20-Poly1305

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: ChaCha20-Poly1305Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal20000M40000M60000M80000M100000MSE +/- 512363.59, N = 3SE +/- 4948671.03, N = 311069945000365541467701. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)

srsRAN Project

Test: PDSCH Processor Benchmark, Throughput Total

OpenBenchmarking.orgMbps, More Is BettersrsRAN Project 23.10.1-20240325Test: PDSCH Processor Benchmark, Throughput TotalGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal5K10K15K20K25KSE +/- 15.31, N = 3SE +/- 10.32, N = 322062.71307.31. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl

High Performance Conjugate Gradient

X Y Z: 104 104 104 - RT: 60

OpenBenchmarking.orgGFLOP/s, More Is BetterHigh Performance Conjugate Gradient 3.1X Y Z: 104 104 104 - RT: 60Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1428425670SE +/- 0.01489, N = 3SE +/- 0.00496, N = 361.619503.768671. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -lmpi_cxx -lmpi

Stress-NG

Test: Memory Copying

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Memory CopyingGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal7K14K21K28K35KSE +/- 13.60, N = 3SE +/- 0.09, N = 332525.121997.891. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

Timed Node.js Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Node.js Compilation 21.7.2Time To CompileGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal10002000300040005000SE +/- 0.69, N = 3SE +/- 1.44, N = 3289.274663.28

OpenSSL

Algorithm: AES-256-GCM

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: AES-256-GCMGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal90000M180000M270000M360000M450000MSE +/- 15611507.33, N = 3SE +/- 17186337.18, N = 3400130558450257734964571. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)

Stress-NG

Test: Vector Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Vector MathGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal90K180K270K360K450KSE +/- 85.60, N = 3SE +/- 0.54, N = 3438910.4028508.491. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

John The Ripper

Test: HMAC-SHA512

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 2023.03.14Test: HMAC-SHA512Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal30M60M90M120M150MSE +/- 162829.50, N = 3SE +/- 9769.57, N = 3158238333103216671. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt

Stress-NG

Test: Trigonometric Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Trigonometric MathGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal10K20K30K40K50KSE +/- 11.48, N = 3SE +/- 0.04, N = 345810.372998.751. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

Stress-NG

Test: Logarithmic Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Logarithmic MathGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal11K22K33K44K55KSE +/- 1.22, N = 3SE +/- 0.28, N = 350252.663302.681. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

Blender

Blend File: Fishy Cat - Compute: CPU-Only

OpenBenchmarking.orgSeconds, Fewer Is BetterBlender 4.0.2Blend File: Fishy Cat - Compute: CPU-OnlyGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal2004006008001000SE +/- 0.26, N = 3SE +/- 0.38, N = 364.07952.43

GraphicsMagick

Operation: Enhanced

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.43Operation: EnhancedGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal60120180240300SE +/- 0.67, N = 3SE +/- 0.00, N = 3294201. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp

7-Zip Compression

Test: Compression Rating

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 24.05Test: Compression RatingGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal100K200K300K400K500KSE +/- 2385.17, N = 3SE +/- 28.98, N = 3479851326971. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

OpenSSL

Algorithm: AES-128-GCM

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: AES-128-GCMGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal100000M200000M300000M400000M500000MSE +/- 2173107.96, N = 3SE +/- 35254075.37, N = 3444871148797304557413371. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)

Stress-NG

Test: Floating Point

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Floating PointGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal3K6K9K12K15KSE +/- 2.60, N = 3SE +/- 1.60, N = 312455.02852.801. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

Stress-NG

Test: Power Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Power MathGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal5K10K15K20K25KSE +/- 0.98, N = 3SE +/- 0.08, N = 323418.401609.161. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

Coremark

CoreMark Size 666 - Iterations Per Second

OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per SecondGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal600K1200K1800K2400K3000KSE +/- 64.82, N = 3SE +/- 24.05, N = 32732642.27192273.031. (CC) gcc options: -O2 -lrt" -lrt

OpenSSL

Algorithm: SHA256

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSLAlgorithm: SHA256Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal20000M40000M60000M80000M100000MSE +/- 7443314.27, N = 3SE +/- 62991235.00, N = 38560484677061517499331. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)

Stress-NG

Test: Fused Multiply-Add

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.17.08Test: Fused Multiply-AddGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal14M28M42M56M70MSE +/- 1063.93, N = 3SE +/- 2472.21, N = 365585333.044928995.121. (CXX) g++ options: -O2 -std=gnu99 -lc -lm

Timed Godot Game Engine Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Godot Game Engine Compilation 4.0Time To CompileGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal400800120016002000SE +/- 1.14, N = 3SE +/- 0.31, N = 3126.351674.47

GraphicsMagick

Operation: Sharpen

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.43Operation: SharpenGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal70140210280350SE +/- 0.33, N = 3SE +/- 0.00, N = 3325261. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp

John The Ripper

Test: MD5

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 2023.03.14Test: MD5Graviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal500K1000K1500K2000K2500KSE +/- 1201.85, N = 3SE +/- 287.69, N = 323673331904261. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt

RocksDB

Test: Read Random Write Random

OpenBenchmarking.orgOp/s, More Is BetterRocksDB 9.0Test: Read Random Write RandomGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal1.5M3M4.5M6M7.5MSE +/- 15765.85, N = 3SE +/- 4162.42, N = 370471015744581. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti

7-Zip Compression

Test: Decompression Rating

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 24.05Test: Decompression RatingGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal110K220K330K440K550KSE +/- 158.70, N = 3SE +/- 50.81, N = 3494660412441. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

srsRAN Project

Test: PUSCH Processor Benchmark, Throughput Total

OpenBenchmarking.orgMbps, More Is BettersrsRAN Project 23.10.1-20240325Test: PUSCH Processor Benchmark, Throughput TotalGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal400800120016002000SE +/- 0.07, N = 3SE +/- 0.03, N = 31961.7169.3MIN: 1132.6 / MAX: 1961.81. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl

John The Ripper

Test: WPA PSK

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 2023.03.14Test: WPA PSKGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal20K40K60K80K100KSE +/- 0.00, N = 3SE +/- 1.33, N = 38628178941. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt

John The Ripper

Test: bcrypt

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 2023.03.14Test: bcryptGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal20K40K60K80K100KSE +/- 22.45, N = 3SE +/- 94.53, N = 158564778421. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt

John The Ripper

Test: Blowfish

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 2023.03.14Test: BlowfishGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal20K40K60K80K100KSE +/- 3.00, N = 3SE +/- 106.90, N = 158566780031. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Write - Average Latency

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL 16Scaling Factor: 100 - Clients: 1000 - Mode: Read Write - Average LatencyGraviton1 16 Cores a1.metalGraviton4 96 Cores r8g.metal-24xl50100150200250SE +/- 1.47, N = 3SE +/- 2.16, N = 3215.89232.431. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm

PostgreSQL

Scaling Factor: 100 - Clients: 1000 - Mode: Read Write

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL 16Scaling Factor: 100 - Clients: 1000 - Mode: Read WriteGraviton1 16 Cores a1.metalGraviton4 96 Cores r8g.metal-24xl10002000300040005000SE +/- 31.31, N = 3SE +/- 40.13, N = 3463243031. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm

Stockfish

Chess Benchmark

OpenBenchmarking.orgNodes Per Second, More Is BetterStockfish 16.1Chess BenchmarkGraviton4 96 Cores r8g.metal-24xlGraviton1 16 Cores a1.metal30M60M90M120M150MSE +/- 2918876.82, N = 15SE +/- 39945.87, N = 911833791824930571. (CXX) g++ options: -lgcov -lpthread -fno-exceptions -std=c++17 -fno-peel-loops -fno-tracer -pedantic -O3 -funroll-loops -flto -flto-partition=one -flto=jobserver


Phoronix Test Suite v10.8.5