Benchmarks by Michael Larabel.
Graviton1 16 Cores a1.metal Processor: ARMv8 Cortex-A72 (16 Cores), Motherboard: Amazon EC2 a1.metal (1.0 BIOS), Memory: 32GB, Disk: 429GB Amazon Elastic Block Store, Network: Amazon Elastic
OS: Ubuntu 24.04, Kernel: 6.8.0-1009-aws (aarch64), Compiler: GCC 13.2.0, File-System: ext4
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vJava Notes: OpenJDK Runtime Environment (build 11.0.23+9-post-Ubuntu-1ubuntu1)Python Notes: Python 3.12.3Security Notes: 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 Changed Processor to ARMv8 Neoverse-V2 (96 Cores) .
Changed Motherboard to Amazon EC2 r8g.metal-24xl (1.0 BIOS) .
Changed Memory to 12 x 64GB DDR5-5600MT/s .
Security Change: 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 Benchmarks Processor Motherboard Memory Disk Network OS Kernel Compiler File-System Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl ARMv8 Cortex-A72 (16 Cores) Amazon EC2 a1.metal (1.0 BIOS) 32GB 429GB Amazon Elastic Block Store Amazon Elastic Ubuntu 24.04 6.8.0-1009-aws (aarch64) GCC 13.2.0 ext4 ARMv8 Neoverse-V2 (96 Cores) Amazon EC2 r8g.metal-24xl (1.0 BIOS) 12 x 64GB DDR5-5600MT/s OpenBenchmarking.org Kernel Details - Transparent Huge Pages: madvise Compiler 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 3.12.3 Security 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
Graviton1 16 Cores a1.metal vs. Graviton4 96 Cores r8g.metal-24xl Comparison Phoronix Test Suite Baseline +2125.7% +2125.7% +4251.4% +4251.4% +6377.1% +6377.1% +8502.8% +8502.8% 866.8% 811.3% 809.6% 8502.9% 706.4% 700% 666.6% 649.5% 6796.3% 525% 469.1% 439% 432.9% 413.4% 4646.7% 374.9% 374% 372.2% 355.7% 350.5% 3354.1% 299.2% 295.2% 271.6% 271.1% 254% 250% 249.2% 243% 232.5% 229.8% 224.2% 208.7% 206.8% 2895.4% 2895.1% 2779.2% 2740.2% 2737% 2652.4% 2628% 2168.4% 2112.8% 2023.1% 196.5% 193.6% 190.1% 188.4% 159.3% 150.8% 147.4% 140.6% 125.2% 1758.2% 1695.8% 1677.6% 1599.3% 1589% 1587.7% 1535% 1528% 1512.1% 1452.5% 1439.6% 1433.1% 1427.6% 1421.6% 1386.5% 1370% 1367.6% 1360.7% 1360.5% 1355.3% 1321.2% 1291.6% 1230.6% 1225.3% 1150% 1143.2% 1126.7% 1099.4% 1058.7% 993% 992.2% 970.4% R.R.W.R Read While Writing Writes Matrix Math Update Rand Noise-Gaussian 6 CPU Cache S.F.P.R pathlib Update Rand 6, Lossless Time To Compile Chess Benchmark 2 Read While Writing Resizing regex_compile V.F.P 2 - Compression Speed xml_etree HWB Color Space P.P.B.T.T P.B.S json_loads crypto_pyaes python_startup Q.1.L 12 - D.S 9 - D.S 3 - Compression Speed 3 - D.S 4K - 16 5K - 16 1080p - 16 100 - 1000 - Read Only 100 - 1000 - Read Only - Average Latency Rand Read Matrix 3D Math Rand Read MPI CPU - water_GMX50_bare SHA512 P.P.B.T.T 1 - Compression Speed 9 - Compression Speed 2 - D.S 1 - D.S Default 12 - Compression Speed Quality 100 Q.1.H.C Classroom - CPU-Only ChaCha20 BMW27 - CPU-Only Small ChaCha20-Poly1305 P.P.B.T.T 104 104 104 - 60 Memory Copying Time To Compile AES-256-GCM Vector Math HMAC-SHA512 Trigonometric Math Logarithmic Math Fishy Cat - CPU-Only Enhanced Compression Rating AES-128-GCM Floating Point Power Math CoreMark Size 666 - I.P.S SHA256 Fused Multiply-Add Time To Compile Sharpen MD5 R.R.W.R D.R P.P.B.T.T WPA PSK bcrypt Blowfish 100 - 1000 - Read Write - Average Latency 7.7% 100 - 1000 - Read Write 7.6% Speedb RocksDB Apache Cassandra Stress-NG RocksDB GraphicsMagick libavif avifenc Stress-NG ACES DGEMM PyPerformance Speedb libavif avifenc Timed PHP Compilation libavif avifenc Stockfish libavif avifenc Numpy Benchmark Speedb GraphicsMagick PyPerformance Stress-NG LZ4 Compression PyPerformance GraphicsMagick srsRAN Project PHPBench PyPerformance PyPerformance PyPerformance WebP Image Encode LZ4 Compression LZ4 Compression LZ4 Compression LZ4 Compression C-Ray C-Ray C-Ray PostgreSQL PostgreSQL Speedb Stress-NG RocksDB GROMACS OpenSSL srsRAN Project LZ4 Compression LZ4 Compression LZ4 Compression LZ4 Compression WebP Image Encode LZ4 Compression WebP Image Encode WebP Image Encode Blender OpenSSL Blender miniFE OpenSSL srsRAN Project High Performance Conjugate Gradient Stress-NG Timed Node.js Compilation OpenSSL Stress-NG John The Ripper Stress-NG Stress-NG Blender GraphicsMagick 7-Zip Compression OpenSSL Stress-NG Stress-NG Coremark OpenSSL Stress-NG Timed Godot Game Engine Compilation GraphicsMagick John The Ripper RocksDB 7-Zip Compression srsRAN Project John The Ripper John The Ripper John The Ripper PostgreSQL PostgreSQL Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl
AWS Graviton Benchmarks speedb: Read Rand Write Rand rocksdb: Read While Writing cassandra: Writes stress-ng: Matrix Math rocksdb: Update Rand graphics-magick: Noise-Gaussian avifenc: 6 stress-ng: CPU Cache mt-dgemm: Sustained Floating-Point Rate pyperformance: pathlib speedb: Update Rand avifenc: 6, Lossless build-php: Time To Compile avifenc: 0 avifenc: 2 numpy: speedb: Read While Writing graphics-magick: Resizing pyperformance: regex_compile stress-ng: Vector Floating Point compress-lz4: 2 - Compression Speed pyperformance: xml_etree graphics-magick: HWB Color Space srsran: PUSCH Processor Benchmark, Throughput Thread phpbench: PHP Benchmark Suite pyperformance: json_loads pyperformance: crypto_pyaes pyperformance: python_startup webp: Quality 100, Lossless compress-lz4: 12 - Decompression Speed compress-lz4: 9 - Decompression Speed compress-lz4: 3 - Compression Speed compress-lz4: 3 - Decompression Speed c-ray: 4K - 16 c-ray: 5K - 16 c-ray: 1080p - 16 pgbench: 100 - 1000 - Read Only pgbench: 100 - 1000 - Read Only - Average Latency speedb: Rand Read stress-ng: Matrix 3D Math rocksdb: Rand Read gromacs: MPI CPU - water_GMX50_bare openssl: SHA512 srsran: PDSCH Processor Benchmark, Throughput Thread compress-lz4: 1 - Compression Speed compress-lz4: 9 - Compression Speed compress-lz4: 2 - Decompression Speed compress-lz4: 1 - Decompression Speed webp: Default compress-lz4: 12 - Compression Speed webp: Quality 100 webp: Quality 100, Highest Compression blender: Classroom - CPU-Only openssl: ChaCha20 blender: BMW27 - CPU-Only minife: Small openssl: ChaCha20-Poly1305 srsran: PDSCH Processor Benchmark, Throughput Total hpcg: 104 104 104 - 60 stress-ng: Memory Copying build-nodejs: Time To Compile openssl: AES-256-GCM stress-ng: Vector Math john-the-ripper: HMAC-SHA512 stress-ng: Trigonometric Math stress-ng: Logarithmic Math blender: Fishy Cat - CPU-Only graphics-magick: Enhanced compress-7zip: Compression Rating openssl: AES-128-GCM stress-ng: Floating Point stress-ng: Power Math coremark: CoreMark Size 666 - Iterations Per Second openssl: SHA256 stress-ng: Fused Multiply-Add build-godot: Time To Compile graphics-magick: Sharpen john-the-ripper: MD5 rocksdb: Read Rand Write Rand compress-7zip: Decompression Rating srsran: PUSCH Processor Benchmark, Throughput Total john-the-ripper: WPA PSK john-the-ripper: bcrypt john-the-ripper: Blowfish pgbench: 100 - 1000 - Read Write - Average Latency pgbench: 100 - 1000 - Read Write stockfish: Chess Benchmark Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 552621 1268650 26348 7343.99 164616 26 21.288 631797.13 0.879250 110 128167 28.229 479.642 668.499 381.112 104.59 2433641 61 455 3466.25 82.51 230 81 18.7 234211 72.1 227 25.9 0.40 1090.3 1102.0 37.30 1137.3 566.946 1008.693 140.822 101870 9.816 23286630 838.94 23252821 0.313 2505891877 94.0 237.68 13.72 1153.1 1514.7 4.63 4.83 3.25 1.43 1289.61 8285450333 599.39 4159.12 6554146770 1307.3 3.76867 1997.89 4663.276 25773496457 28508.49 10321667 2998.75 3302.68 952.43 20 32697 30455741337 852.8 1609.16 192273.033715 6151749933 4928995.12 1674.469 26 190426 574458 41244 169.3 7894 7842 8003 215.890 4632 2493057 5342911 11560819 239658 631796.65 1327423 208 2.777 4735225.17 60.635839 17.6 729398 5.237 89.999 130.203 80.243 495.78 11490911 278 101 119727.57 329.41 58.2 301 69.4 829221 20.6 65.0 7.55 1.33 3595.5 3573.1 115.14 3489.5 18.927 33.678 4.891 2893264 0.346 640938471 22886.62 527458547 6.926 53203418283 278.7 697.89 39.80 3326.1 3927.7 11.61 11.95 7.82 3.22 69.40 148791688760 33.72 70677.8 110699450003 22062.7 61.6195 32525.12 289.267 400130558450 438910.40 158238333 45810.37 50252.66 64.07 294 479851 444871148797 12455.02 23418.40 2732642.268891 85604846770 65585333.04 126.346 325 2367333 7047101 494660 1961.7 86281 85647 85667 232.433 4303 118337918 OpenBenchmarking.org
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Read Random Write Random Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 1.1M 2.2M 3.3M 4.4M 5.5M SE +/- 214.78, N = 3 SE +/- 4241.50, N = 3 552621 5342911 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Read While Writing Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 2M 4M 6M 8M 10M SE +/- 16876.08, N = 3 SE +/- 18723.10, N = 3 1268650 11560819 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Matrix Math Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 140K 280K 420K 560K 700K SE +/- 74.42, N = 3 SE +/- 9.55, N = 3 7343.99 631796.65 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Update Random Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 300K 600K 900K 1200K 1500K SE +/- 443.99, N = 3 SE +/- 5313.65, N = 3 164616 1327423 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Noise-Gaussian Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 50 100 150 200 250 SE +/- 0.00, N = 3 SE +/- 0.67, N = 3 26 208 1. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: CPU Cache Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 1000K 2000K 3000K 4000K 5000K SE +/- 3895.03, N = 3 SE +/- 57877.80, N = 15 631797.13 4735225.17 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
PyPerformance OpenBenchmarking.org Milliseconds, Fewer Is Better PyPerformance 1.11 Benchmark: pathlib Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 20 40 60 80 100 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 110.0 17.6
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Update Random Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 160K 320K 480K 640K 800K SE +/- 302.33, N = 3 SE +/- 3844.86, N = 3 128167 729398 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
OpenBenchmarking.org Seconds, Fewer Is Better libavif avifenc 1.0 Encoder Speed: 2 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 80 160 240 320 400 SE +/- 0.85, N = 3 SE +/- 0.02, N = 3 381.11 80.24 1. (CXX) g++ options: -O3 -fPIC -lm
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Read While Writing Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 2M 4M 6M 8M 10M SE +/- 20302.77, N = 8 SE +/- 87574.70, N = 15 2433641 11490911 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Resizing Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 60 120 180 240 300 SE +/- 0.00, N = 3 SE +/- 0.58, N = 3 61 278 1. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp
PyPerformance OpenBenchmarking.org Milliseconds, Fewer Is Better PyPerformance 1.11 Benchmark: regex_compile Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 100 200 300 400 500 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 455 101
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Vector Floating Point Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 30K 60K 90K 120K 150K SE +/- 0.38, N = 3 SE +/- 17.26, N = 3 3466.25 119727.57 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
LZ4 Compression OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 2 - Compression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 70 140 210 280 350 SE +/- 0.02, N = 3 SE +/- 0.13, N = 3 82.51 329.41 1. (CC) gcc options: -O3 -pthread
PyPerformance OpenBenchmarking.org Milliseconds, Fewer Is Better PyPerformance 1.11 Benchmark: xml_etree Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 50 100 150 200 250 SE +/- 0.58, N = 3 SE +/- 0.00, N = 3 230.0 58.2
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: HWB Color Space Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 70 140 210 280 350 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 81 301 1. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PUSCH Processor Benchmark, Throughput Thread Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 15 30 45 60 75 SE +/- 0.23, N = 4 SE +/- 0.00, N = 3 18.7 69.4 MIN: 11.3 / MAX: 18.9 MIN: 45.4 1. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl
PHPBench PHPBench is a benchmark suite for PHP. It performs a large number of simple tests in order to bench various aspects of the PHP interpreter. PHPBench can be used to compare hardware, operating systems, PHP versions, PHP accelerators and caches, compiler options, etc. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Score, More Is Better PHPBench 0.8.1 PHP Benchmark Suite Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 200K 400K 600K 800K 1000K SE +/- 481.21, N = 3 SE +/- 273.39, N = 3 234211 829221
PyPerformance OpenBenchmarking.org Milliseconds, Fewer Is Better PyPerformance 1.11 Benchmark: json_loads Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 16 32 48 64 80 SE +/- 0.10, N = 3 SE +/- 0.00, N = 3 72.1 20.6
OpenBenchmarking.org Milliseconds, Fewer Is Better PyPerformance 1.11 Benchmark: crypto_pyaes Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 50 100 150 200 250 SE +/- 0.33, N = 3 SE +/- 0.03, N = 3 227.0 65.0
OpenBenchmarking.org Milliseconds, Fewer Is Better PyPerformance 1.11 Benchmark: python_startup Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 6 12 18 24 30 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 25.90 7.55
WebP Image Encode OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.4 Encode Settings: Quality 100, Lossless Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 0.2993 0.5986 0.8979 1.1972 1.4965 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.40 1.33 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
LZ4 Compression OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 12 - Decompression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 800 1600 2400 3200 4000 SE +/- 2.61, N = 3 SE +/- 1.89, N = 3 1090.3 3595.5 1. (CC) gcc options: -O3 -pthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 9 - Decompression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 800 1600 2400 3200 4000 SE +/- 0.06, N = 3 SE +/- 0.40, N = 3 1102.0 3573.1 1. (CC) gcc options: -O3 -pthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 3 - Compression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 30 60 90 120 150 SE +/- 0.01, N = 3 SE +/- 0.08, N = 3 37.30 115.14 1. (CC) gcc options: -O3 -pthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 3 - Decompression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 700 1400 2100 2800 3500 SE +/- 0.06, N = 3 SE +/- 0.15, N = 3 1137.3 3489.5 1. (CC) gcc options: -O3 -pthread
C-Ray OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 2.0 Resolution: 4K - Rays Per Pixel: 16 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 120 240 360 480 600 SE +/- 1.65, N = 3 SE +/- 0.02, N = 3 566.95 18.93 1. (CC) gcc options: -lpthread -lm
OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 2.0 Resolution: 5K - Rays Per Pixel: 16 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 200 400 600 800 1000 SE +/- 2.98, N = 3 SE +/- 0.03, N = 3 1008.69 33.68 1. (CC) gcc options: -lpthread -lm
OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 2.0 Resolution: 1080p - Rays Per Pixel: 16 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 30 60 90 120 150 SE +/- 0.248, N = 3 SE +/- 0.007, N = 3 140.822 4.891 1. (CC) gcc options: -lpthread -lm
PostgreSQL This is a benchmark of PostgreSQL using the integrated pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org TPS, More Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 1000 - Mode: Read Only Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 600K 1200K 1800K 2400K 3000K SE +/- 107.42, N = 3 SE +/- 15759.90, N = 3 101870 2893264 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 1000 - Mode: Read Only - Average Latency Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 3 6 9 12 15 SE +/- 0.010, N = 3 SE +/- 0.002, N = 3 9.816 0.346 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
Speedb Speedb is a next-generation key value storage engine that is RocksDB compatible and aiming for stability, efficiency, and performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Speedb 2.7 Test: Random Read Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 140M 280M 420M 560M 700M SE +/- 174965.25, N = 15 SE +/- 229391.69, N = 3 23286630 640938471 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Matrix 3D Math Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 5K 10K 15K 20K 25K SE +/- 3.67, N = 3 SE +/- 50.98, N = 3 838.94 22886.62 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Random Read Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 110M 220M 330M 440M 550M SE +/- 12274.71, N = 3 SE +/- 53499.62, N = 3 23252821 527458547 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
GROMACS The GROMACS (GROningen MAchine for Chemical Simulations) molecular dynamics package testing with the water_GMX50 data. This test profile allows selecting between CPU and GPU-based GROMACS builds. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2024 Implementation: MPI CPU - Input: water_GMX50_bare Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 2 4 6 8 10 SE +/- 0.000, N = 3 SE +/- 0.001, N = 3 0.313 6.926 1. (CXX) g++ options: -O3 -lm
OpenSSL OpenBenchmarking.org byte/s, More Is Better OpenSSL Algorithm: SHA512 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 11000M 22000M 33000M 44000M 55000M SE +/- 4619831.10, N = 3 SE +/- 4634061.90, N = 3 2505891877 53203418283 1. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PDSCH Processor Benchmark, Throughput Thread Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 60 120 180 240 300 SE +/- 0.21, N = 3 SE +/- 3.00, N = 3 94.0 278.7 1. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl
LZ4 Compression OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 1 - Compression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 150 300 450 600 750 SE +/- 3.13, N = 3 SE +/- 0.04, N = 3 237.68 697.89 1. (CC) gcc options: -O3 -pthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 9 - Compression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 9 18 27 36 45 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 13.72 39.80 1. (CC) gcc options: -O3 -pthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 2 - Decompression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 700 1400 2100 2800 3500 SE +/- 0.27, N = 3 SE +/- 0.13, N = 3 1153.1 3326.1 1. (CC) gcc options: -O3 -pthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 1 - Decompression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 800 1600 2400 3200 4000 SE +/- 0.30, N = 3 SE +/- 0.19, N = 3 1514.7 3927.7 1. (CC) gcc options: -O3 -pthread
WebP Image Encode OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.4 Encode Settings: Default Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 4.63 11.61 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
LZ4 Compression OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.10 Compression Level: 12 - Compression Speed Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 3 6 9 12 15 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 4.83 11.95 1. (CC) gcc options: -O3 -pthread
WebP Image Encode OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.4 Encode Settings: Quality 100 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 2 4 6 8 10 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 3.25 7.82 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
OpenBenchmarking.org MP/s, More Is Better WebP Image Encode 1.4 Encode Settings: Quality 100, Highest Compression Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 0.7245 1.449 2.1735 2.898 3.6225 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 1.43 3.22 1. (CC) gcc options: -fvisibility=hidden -O2 -lm
Blender Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing is supported. This system/blender test profile makes use of the system-supplied Blender. Use pts/blender if wishing to stick to a fixed version of Blender. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.0.2 Blend File: Classroom - Compute: CPU-Only Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 300 600 900 1200 1500 SE +/- 0.28, N = 3 SE +/- 0.03, N = 3 1289.61 69.40
OpenSSL OpenBenchmarking.org byte/s, More Is Better OpenSSL Algorithm: ChaCha20 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 30000M 60000M 90000M 120000M 150000M SE +/- 4780384.74, N = 3 SE +/- 430328.22, N = 3 8285450333 148791688760 1. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)
Blender Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing is supported. This system/blender test profile makes use of the system-supplied Blender. Use pts/blender if wishing to stick to a fixed version of Blender. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.0.2 Blend File: BMW27 - Compute: CPU-Only Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 130 260 390 520 650 SE +/- 0.63, N = 3 SE +/- 0.03, N = 3 599.39 33.72
OpenSSL OpenBenchmarking.org byte/s, More Is Better OpenSSL Algorithm: ChaCha20-Poly1305 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 20000M 40000M 60000M 80000M 100000M SE +/- 4948671.03, N = 3 SE +/- 512363.59, N = 3 6554146770 110699450003 1. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PDSCH Processor Benchmark, Throughput Total Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 5K 10K 15K 20K 25K SE +/- 10.32, N = 3 SE +/- 15.31, N = 3 1307.3 22062.7 1. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl
High Performance Conjugate Gradient HPCG is the High Performance Conjugate Gradient and is a new scientific benchmark from Sandia National Lans focused for super-computer testing with modern real-world workloads compared to HPCC. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org GFLOP/s, More Is Better High Performance Conjugate Gradient 3.1 X Y Z: 104 104 104 - RT: 60 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 14 28 42 56 70 SE +/- 0.00496, N = 3 SE +/- 0.01489, N = 3 3.76867 61.61950 1. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -lmpi_cxx -lmpi
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Memory Copying Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 7K 14K 21K 28K 35K SE +/- 0.09, N = 3 SE +/- 13.60, N = 3 1997.89 32525.12 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
OpenSSL OpenBenchmarking.org byte/s, More Is Better OpenSSL Algorithm: AES-256-GCM Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 90000M 180000M 270000M 360000M 450000M SE +/- 17186337.18, N = 3 SE +/- 15611507.33, N = 3 25773496457 400130558450 1. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Vector Math Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 90K 180K 270K 360K 450K SE +/- 0.54, N = 3 SE +/- 85.60, N = 3 28508.49 438910.40 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Trigonometric Math Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 10K 20K 30K 40K 50K SE +/- 0.04, N = 3 SE +/- 11.48, N = 3 2998.75 45810.37 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Logarithmic Math Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 11K 22K 33K 44K 55K SE +/- 0.28, N = 3 SE +/- 1.22, N = 3 3302.68 50252.66 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
Blender Blender is an open-source 3D creation software project. This test is of Blender's Cycles benchmark with various sample files. GPU computing is supported. This system/blender test profile makes use of the system-supplied Blender. Use pts/blender if wishing to stick to a fixed version of Blender. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better Blender 4.0.2 Blend File: Fishy Cat - Compute: CPU-Only Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 200 400 600 800 1000 SE +/- 0.38, N = 3 SE +/- 0.26, N = 3 952.43 64.07
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Enhanced Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 60 120 180 240 300 SE +/- 0.00, N = 3 SE +/- 0.67, N = 3 20 294 1. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp
7-Zip Compression OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 24.05 Test: Compression Rating Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 100K 200K 300K 400K 500K SE +/- 28.98, N = 3 SE +/- 2385.17, N = 3 32697 479851 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
OpenSSL OpenBenchmarking.org byte/s, More Is Better OpenSSL Algorithm: AES-128-GCM Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 100000M 200000M 300000M 400000M 500000M SE +/- 35254075.37, N = 3 SE +/- 2173107.96, N = 3 30455741337 444871148797 1. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Floating Point Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 3K 6K 9K 12K 15K SE +/- 1.60, N = 3 SE +/- 2.60, N = 3 852.80 12455.02 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Power Math Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 5K 10K 15K 20K 25K SE +/- 0.08, N = 3 SE +/- 0.98, N = 3 1609.16 23418.40 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
OpenSSL OpenBenchmarking.org byte/s, More Is Better OpenSSL Algorithm: SHA256 Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 20000M 40000M 60000M 80000M 100000M SE +/- 62991235.00, N = 3 SE +/- 7443314.27, N = 3 6151749933 85604846770 1. OpenSSL 3.0.13 30 Jan 2024 (Library: OpenSSL 3.0.13 30 Jan 2024)
Stress-NG OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.17.08 Test: Fused Multiply-Add Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 14M 28M 42M 56M 70M SE +/- 2472.21, N = 3 SE +/- 1063.93, N = 3 4928995.12 65585333.04 1. (CXX) g++ options: -O2 -std=gnu99 -lc -lm
GraphicsMagick This is a test of GraphicsMagick with its OpenMP implementation that performs various imaging tests on a sample high resolution (currently 15400 x 6940) JPEG image. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.43 Operation: Sharpen Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 70 140 210 280 350 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 26 325 1. (CC) gcc options: -fopenmp -O2 -ljpeg -lXext -lSM -lICE -lX11 -lxml2 -lz -lm -lpthread -lgomp
RocksDB This is a benchmark of Meta/Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better RocksDB 9.0 Test: Read Random Write Random Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 1.5M 3M 4.5M 6M 7.5M SE +/- 4162.42, N = 3 SE +/- 15765.85, N = 3 574458 7047101 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti
7-Zip Compression OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 24.05 Test: Decompression Rating Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 110K 220K 330K 440K 550K SE +/- 50.81, N = 3 SE +/- 158.70, N = 3 41244 494660 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
srsRAN Project srsRAN Project is a complete ORAN-native 5G RAN solution created by Software Radio Systems (SRS). The srsRAN Project radio suite was formerly known as srsLTE and can be used for building your own software-defined radio (SDR) 4G/5G mobile network. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Mbps, More Is Better srsRAN Project 23.10.1-20240325 Test: PUSCH Processor Benchmark, Throughput Total Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 400 800 1200 1600 2000 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 169.3 1961.7 MIN: 1132.6 / MAX: 1961.8 1. (CXX) g++ options: -O3 -fno-trapping-math -fno-math-errno -ldl
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: bcrypt Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 20K 40K 60K 80K 100K SE +/- 94.53, N = 15 SE +/- 22.45, N = 3 7842 85647 1. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt
OpenBenchmarking.org Real C/S, More Is Better John The Ripper 2023.03.14 Test: Blowfish Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 20K 40K 60K 80K 100K SE +/- 106.90, N = 15 SE +/- 3.00, N = 3 8003 85667 1. (CC) gcc options: -lssl -lcrypto -fopenmp -lm -lrt -lz -ldl -lcrypt
PostgreSQL This is a benchmark of PostgreSQL using the integrated pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 1000 - Mode: Read Write - Average Latency Graviton4 96 Cores r8g.metal-24xl Graviton1 16 Cores a1.metal 50 100 150 200 250 SE +/- 2.16, N = 3 SE +/- 1.47, N = 3 232.43 215.89 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL 16 Scaling Factor: 100 - Clients: 1000 - Mode: Read Write Graviton4 96 Cores r8g.metal-24xl Graviton1 16 Cores a1.metal 1000 2000 3000 4000 5000 SE +/- 40.13, N = 3 SE +/- 31.31, N = 3 4303 4632 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lm
Stockfish This is a test of Stockfish, an advanced open-source C++11 chess benchmark that can scale up to 1024 CPU threads. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Nodes Per Second, More Is Better Stockfish 16.1 Chess Benchmark Graviton1 16 Cores a1.metal Graviton4 96 Cores r8g.metal-24xl 30M 60M 90M 120M 150M SE +/- 39945.87, N = 9 SE +/- 2918876.82, N = 15 2493057 118337918 1. (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
Graviton1 16 Cores a1.metal Processor: ARMv8 Cortex-A72 (16 Cores), Motherboard: Amazon EC2 a1.metal (1.0 BIOS), Memory: 32GB, Disk: 429GB Amazon Elastic Block Store, Network: Amazon Elastic
OS: Ubuntu 24.04, Kernel: 6.8.0-1009-aws (aarch64), Compiler: GCC 13.2.0, File-System: ext4
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vJava Notes: OpenJDK Runtime Environment (build 11.0.23+9-post-Ubuntu-1ubuntu1)Python Notes: Python 3.12.3Security Notes: 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
Testing initiated at 22 July 2024 16:02 by user ubuntu.
Graviton4 96 Cores r8g.metal-24xl Processor: ARMv8 Neoverse-V2 (96 Cores), Motherboard: Amazon EC2 r8g.metal-24xl (1.0 BIOS), Memory: 12 x 64GB DDR5-5600MT/s, Disk: 429GB Amazon Elastic Block Store, Network: Amazon Elastic
OS: Ubuntu 24.04, Kernel: 6.8.0-1009-aws (aarch64), Compiler: GCC 13.2.0, File-System: ext4
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vJava Notes: OpenJDK Runtime Environment (build 11.0.23+9-post-Ubuntu-1ubuntu1)Python Notes: Python 3.12.3Security Notes: 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
Testing initiated at 23 July 2024 14:12 by user ubuntu.