m6g.metal Graviton2 vs. Ampere Altra

Ampere Altra ARMv8 Neoverse-N1 tests against Amazon Graviton2 m6g.metal for possible future article.

HTML result view exported from: https://openbenchmarking.org/result/2012190-HA-ALTRAEC2377.

m6g.metal Graviton2 vs. Ampere AltraProcessorMotherboardMemoryDiskNetworkChipsetGraphicsMonitorOSKernelCompilerFile-SystemScreen Resolutionm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160cARMv8 Neoverse-N1 (64 Cores)Amazon EC2 m6g.metal v1.0252GB94GB Amazon Elastic Block StoreAmazon ElasticUbuntu 20.045.4.0-1029-aws (aarch64)GCC 9.3.0ext4Ampere Altra ARMv8 Neoverse-N1 @ 3.30GHz (64 Cores)WIWYNN Mt.Jade (1.1.20201019 BIOS)Ampere Computing LLC Device e10016 x 32 GB DDR4-3200MT/s Samsung M393A4K40DB3-CWE3841GB Micron_9300_MTFDHAL3T8TDP + 960GB SAMSUNG MZ1LB960HAJQ-00007ASPEEDMellanox MT28908 + Intel I2105.4.0-58-generic (aarch64)1024x768Ampere Altra ARMv8 Neoverse-N1 @ 3.30GHz (80 Cores)VE2281920x1080Ampere Altra ARMv8 Neoverse-N1 @ 3.30GHz (160 Cores)OpenBenchmarking.orgCompiler 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++,gm2 --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-nls --enable-objc-gc=auto --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 -v Python Details- Python 3.8.5Security Details- itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: 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 Processor Details- Ampere Altra Q80-33 1P 64c, Ampere Altra Q80-33 1P 80c, Ampere Altra Q80-33 2P 160c: Scaling Governor: cppc_cpufreq performance

m6g.metal Graviton2 vs. Ampere Altrastream: Copystream: Triadstream: Addhpcg: byte: Dhrystone 2compress-lz4: 3 - Compression Speedcompress-lz4: 9 - Compression Speedjohn-the-ripper: Blowfishcoremark: CoreMark Size 666 - Iterations Per Secondc-ray: Total Time - 4K, 16 Rays Per Pixelm-queens: Time To Solvegromacs: Water Benchmarkastcenc: Exhaustivestress-ng: Cryptostress-ng: CPU Stressstress-ng: Matrix Mathstress-ng: Vector Mathstress-ng: Context Switchingtnn: CPU - MobileNet v2tnn: CPU - SqueezeNet v1.1m6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c165514.7170932.8170481.721.436425752636.437.3435.72442181205893.12310815.60419.4392.76170.7311355.887228.57256330.06353821.3419337654.56355.256332.793162886.4167747.4166546.222.464633950617.148.1644.37508031551293.00223312.04714.8572.83560.8215037.109415.04311734.28465202.1218513502.97300.419255.599162465.9165070.5166228.622.710133953700.047.0743.84487111854977.7105059.45412.0353.87149.1218691.0711784.05420159.19577961.0722953240.38291.322251.554239685.8272185.3266737.944.910833963046.248.1746.01907733256887.0019614.9076.5516.61128.6736316.9623501.47819104.511143937.0745404883.10311.925267.864OpenBenchmarking.org

Stream

Type: Copy

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: Copym6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c50K100K150K200K250KSE +/- 83.68, N = 5SE +/- 1497.95, N = 5SE +/- 1282.06, N = 25SE +/- 4388.23, N = 25165514.7162886.4162465.9239685.81. (CC) gcc options: -O3 -march=native -fopenmp

Stream

Type: Triad

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: Triadm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c60K120K180K240K300KSE +/- 69.34, N = 5SE +/- 292.37, N = 5SE +/- 3156.35, N = 5SE +/- 15170.38, N = 5170932.8167747.4165070.5272185.31. (CC) gcc options: -O3 -march=native -fopenmp

Stream

Type: Add

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17Type: Addm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c60K120K180K240K300KSE +/- 33.82, N = 5SE +/- 329.58, N = 5SE +/- 2620.82, N = 5SE +/- 15401.72, N = 5170481.7166546.2166228.6266737.91. (CC) gcc options: -O3 -march=native -fopenmp

High Performance Conjugate Gradient

OpenBenchmarking.orgGFLOP/s, More Is BetterHigh Performance Conjugate Gradient 3.1m6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c1020304050SE +/- 0.00, N = 3SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.05, N = 321.4422.4622.7144.911. (CXX) g++ options: -O3 -ffast-math -ftree-vectorize -pthread -lmpi_cxx -lmpi

BYTE Unix Benchmark

Computational Test: Dhrystone 2

OpenBenchmarking.orgLPS, More Is BetterBYTE Unix Benchmark 3.6Computational Test: Dhrystone 2m6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c7M14M21M28M35MSE +/- 3044.65, N = 3SE +/- 34772.76, N = 3SE +/- 7604.20, N = 3SE +/- 8257.21, N = 325752636.433950617.133953700.033963046.2

LZ4 Compression

Compression Level: 3 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Compression Speedm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c1122334455SE +/- 0.02, N = 3SE +/- 0.11, N = 3SE +/- 0.68, N = 3SE +/- 0.08, N = 337.3448.1647.0748.171. (CC) gcc options: -O3

LZ4 Compression

Compression Level: 9 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Compression Speedm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c1020304050SE +/- 0.01, N = 3SE +/- 0.44, N = 15SE +/- 0.03, N = 3SE +/- 0.09, N = 335.7244.3743.8446.011. (CC) gcc options: -O3

John The Ripper

Test: Blowfish

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 1.9.0-jumbo-1Test: Blowfishm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c20K40K60K80K100KSE +/- 9.00, N = 3SE +/- 1823.77, N = 15SE +/- 1790.62, N = 15SE +/- 4405.69, N = 15442185080348711907731. (CC) gcc options: -lssl -lcrypto -fopenmp -pthread -lm -lz -ldl -lcrypt

Coremark

CoreMark Size 666 - Iterations Per Second

OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per Secondm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c700K1400K2100K2800K3500KSE +/- 20239.38, N = 15SE +/- 12939.15, N = 3SE +/- 13320.07, N = 3SE +/- 54021.40, N = 151205893.121551293.001854977.713256887.001. (CC) gcc options: -O2 -lrt" -lrt

C-Ray

Total Time - 4K, 16 Rays Per Pixel

OpenBenchmarking.orgSeconds, Fewer Is BetterC-Ray 1.1Total Time - 4K, 16 Rays Per Pixelm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c48121620SE +/- 0.020, N = 3SE +/- 0.033, N = 3SE +/- 0.021, N = 3SE +/- 0.018, N = 315.60412.0479.4544.9071. (CC) gcc options: -lm -lpthread -O3

m-queens

Time To Solve

OpenBenchmarking.orgSeconds, Fewer Is Betterm-queens 1.2Time To Solvem6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c510152025SE +/- 0.008, N = 3SE +/- 0.034, N = 3SE +/- 0.132, N = 3SE +/- 0.070, N = 519.43914.85712.0356.5511. (CXX) g++ options: -fopenmp -O2 -march=native

GROMACS

Water Benchmark

OpenBenchmarking.orgNs Per Day, More Is BetterGROMACS 2020.3Water Benchmarkm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c246810SE +/- 0.001, N = 3SE +/- 0.058, N = 15SE +/- 0.016, N = 3SE +/- 0.014, N = 32.7612.8353.8716.6111. (CXX) g++ options: -O3 -pthread -lrt -lpthread -lm

ASTC Encoder

Preset: Exhaustive

OpenBenchmarking.orgSeconds, Fewer Is BetterASTC Encoder 2.0Preset: Exhaustivem6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c1632486480SE +/- 0.01, N = 3SE +/- 0.17, N = 3SE +/- 0.21, N = 3SE +/- 0.45, N = 1370.7360.8249.1228.671. (CXX) g++ options: -std=c++14 -fvisibility=hidden -O3 -flto -lpthread

Stress-NG

Test: Crypto

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Cryptom6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c8K16K24K32K40KSE +/- 22.19, N = 3SE +/- 32.09, N = 3SE +/- 75.96, N = 3SE +/- 360.65, N = 311355.8815037.1018691.0736316.961. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

Stress-NG

Test: CPU Stress

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: CPU Stressm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c5K10K15K20K25KSE +/- 0.93, N = 3SE +/- 29.91, N = 3SE +/- 9.44, N = 3SE +/- 120.41, N = 37228.579415.0411784.0523501.471. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

Stress-NG

Test: Matrix Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Matrix Mathm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c200K400K600K800K1000KSE +/- 9.77, N = 3SE +/- 1405.87, N = 3SE +/- 335.41, N = 3SE +/- 6853.85, N = 3256330.06311734.28420159.19819104.511. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

Stress-NG

Test: Vector Math

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Vector Mathm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c200K400K600K800K1000KSE +/- 2.19, N = 3SE +/- 419.80, N = 3SE +/- 2494.00, N = 3SE +/- 1813.61, N = 3353821.34465202.12577961.071143937.071. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

Stress-NG

Test: Context Switching

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Context Switchingm6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c10M20M30M40M50MSE +/- 252656.21, N = 3SE +/- 176208.12, N = 15SE +/- 217249.98, N = 6SE +/- 353501.41, N = 319337654.5618513502.9722953240.3845404883.101. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

TNN

Target: CPU - Model: MobileNet v2

OpenBenchmarking.orgms, Fewer Is BetterTNN 0.2.3Target: CPU - Model: MobileNet v2m6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c80160240320400SE +/- 0.14, N = 3SE +/- 3.40, N = 15SE +/- 2.60, N = 3SE +/- 1.51, N = 3355.26300.42291.32311.93MIN: 354.82 / MAX: 355.87MIN: 285.58 / MAX: 343.17MIN: 285.33 / MAX: 397.43MIN: 293.86 / MAX: 414.931. (CXX) g++ options: -fopenmp -pthread -fvisibility=hidden -O3 -rdynamic -ldl

TNN

Target: CPU - Model: SqueezeNet v1.1

OpenBenchmarking.orgms, Fewer Is BetterTNN 0.2.3Target: CPU - Model: SqueezeNet v1.1m6g.metal Graviton2 64cAmpere Altra Q80-33 1P 64cAmpere Altra Q80-33 1P 80cAmpere Altra Q80-33 2P 160c70140210280350SE +/- 0.27, N = 3SE +/- 1.68, N = 3SE +/- 0.46, N = 3SE +/- 3.88, N = 15332.79255.60251.55267.86MIN: 332.17 / MAX: 333.54MIN: 251.62 / MAX: 265.03MIN: 249.95 / MAX: 252.68MIN: 249.7 / MAX: 411.031. (CXX) g++ options: -fopenmp -pthread -fvisibility=hidden -O3 -rdynamic -ldl


Phoronix Test Suite v10.8.4