Graviton2 EC2 AMD EPYC 7002 Rome Comparison Benchmarks

Benchmarks for a future article.

HTML result view exported from: https://openbenchmarking.org/result/2006047-PTS-EPYC2EC201&obr_sgm=y&obr_sor=y&obr_hgv=EPYC+c5a.16xlarge,EPYC+c5a.8xlarge.

Graviton2 EC2 AMD EPYC 7002 Rome Comparison BenchmarksProcessorMotherboardChipsetMemoryDiskNetworkOSKernelCompilerFile-SystemSystem LayerEPYC m5a.16xlargeEPYC m5a.8xlargeXeon m5.16xlargeXeon m5.8xlargeArm m6g.16xlargeArm m6g.8xlargeEPYC c5a.16xlargeEPYC c5a.8xlargeAMD EPYC 7571 (32 Cores / 64 Threads)Amazon EC2 m5a.16xlarge (1.0 BIOS)Intel 440FX 82441FX PMC248GB94GB Amazon Elastic Block StoreAmazon ElasticAmazon Linux 24.14.173-137.229.amzn2.x86_64 (x86_64)GCC 7.3.1 20180712xfsKVMAMD EPYC 7571 (16 Cores / 32 Threads)Amazon EC2 m5a.8xlarge (1.0 BIOS)124GB2 x Intel Xeon Platinum 8175M (32 Cores / 64 Threads)Amazon EC2 m5.16xlarge (1.0 BIOS)248GBIntel Xeon Platinum 8175M (16 Cores / 32 Threads)Amazon EC2 m5.8xlarge (1.0 BIOS)124GBARMv8 (64 Cores)Amazon EC2 m6g.16xlarge (1.0 BIOS)Amazon Device 0200248GB4.14.173-137.229.amzn2.aarch64 (aarch64) 20200401ARMv8 (32 Cores)Amazon EC2 m6g.8xlarge (1.0 BIOS)124GBAMD EPYC 7R32 (32 Cores / 64 Threads)Amazon EC2 c5a.16xlarge (1.0 BIOS)Intel 440FX 82441FX PMC126GB4.14.177-139.254.amzn2.x86_64 (x86_64)KVMAMD EPYC 7R32 (16 Cores / 32 Threads)Amazon EC2 c5a.8xlarge (1.0 BIOS)64GBOpenBenchmarking.orgKernel Details- nvme_core.io_timeout=4294967295Environment Details- EPYC m5a.16xlarge: COST_PERF_PER_HOUR=0.6385- EPYC m5a.8xlarge: COST_PERF_PER_HOUR=0.3193- Xeon m5.16xlarge: COST_PERF_PER_HOUR=0.6385- Xeon m5.8xlarge: COST_PERF_PER_HOUR=0.3193- Arm m6g.16xlarge: COST_PERF_PER_HOUR=0.4482- Arm m6g.8xlarge: COST_PERF_PER_HOUR=0.2241Compiler Details- EPYC m5a.16xlarge: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libcilkrts --enable-libitm --enable-libmpx --enable-libquadmath --enable-libsanitizer --enable-multilib --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic - EPYC m5a.8xlarge: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libcilkrts --enable-libitm --enable-libmpx --enable-libquadmath --enable-libsanitizer --enable-multilib --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic - Xeon m5.16xlarge: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libcilkrts --enable-libitm --enable-libmpx --enable-libquadmath --enable-libsanitizer --enable-multilib --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic - Xeon m5.8xlarge: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libcilkrts --enable-libitm --enable-libmpx --enable-libquadmath --enable-libsanitizer --enable-multilib --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic - Arm m6g.16xlarge: --build=aarch64-redhat-linux --disable-libcilkrts --disable-libmpx --disable-libquadmath --disable-libunwind-exceptions --disable-multilib --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libitm --enable-libsanitizer --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu - Arm m6g.8xlarge: --build=aarch64-redhat-linux --disable-libcilkrts --disable-libmpx --disable-libquadmath --disable-libunwind-exceptions --disable-multilib --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libitm --enable-libsanitizer --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu - EPYC c5a.16xlarge: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libcilkrts --enable-libitm --enable-libmpx --enable-libquadmath --enable-libsanitizer --enable-multilib --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic - EPYC c5a.8xlarge: --build=x86_64-redhat-linux --disable-libunwind-exceptions --enable-__cxa_atexit --enable-bootstrap --enable-checking=release --enable-gnu-indirect-function --enable-gnu-unique-object --enable-initfini-array --enable-languages=c,c++,objc,obj-c++,fortran,ada,go,lto --enable-libatomic --enable-libcilkrts --enable-libitm --enable-libmpx --enable-libquadmath --enable-libsanitizer --enable-multilib --enable-plugin --enable-shared --enable-threads=posix --mandir=/usr/share/man --with-arch_32=x86-64 --with-gcc-major-version-only --with-isl --with-linker-hash-style=gnu --with-tune=generic Processor Details- EPYC m5a.16xlarge: CPU Microcode: 0x8001250- EPYC m5a.8xlarge: CPU Microcode: 0x8001250- Xeon m5.16xlarge: CPU Microcode: 0x2000069- Xeon m5.8xlarge: CPU Microcode: 0x2000069- EPYC c5a.16xlarge: CPU Microcode: 0x8301025- EPYC c5a.8xlarge: CPU Microcode: 0x8301025Java Details- OpenJDK Runtime Environment (build 1.8.0_252-b09)Python Details- Python 2.7.16Security Details- EPYC m5a.16xlarge: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline STIBP: disabled RSB filling + tsx_async_abort: Not affected- EPYC m5a.8xlarge: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline STIBP: disabled RSB filling + tsx_async_abort: Not affected- Xeon m5.16xlarge: itlb_multihit: KVM: Vulnerable + l1tf: Mitigation of PTE Inversion + mds: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + meltdown: Mitigation of PTI + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline STIBP: disabled RSB filling + tsx_async_abort: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown - Xeon m5.8xlarge: itlb_multihit: KVM: Vulnerable + l1tf: Mitigation of PTE Inversion + mds: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown + meltdown: Mitigation of PTI + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full generic retpoline STIBP: disabled RSB filling + tsx_async_abort: Vulnerable: Clear buffers attempted no microcode; SMT Host state unknown - Arm m6g.16xlarge: 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 + tsx_async_abort: Not affected- Arm m6g.8xlarge: 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 + tsx_async_abort: Not affected- EPYC c5a.16xlarge: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + tsx_async_abort: Not affected- EPYC c5a.8xlarge: itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling + tsx_async_abort: Not affected

Graviton2 EC2 AMD EPYC 7002 Rome Comparison Benchmarksbuild-linux-kernel: Time To Compilecompress-7zip: Compress Speed Testpgbench: Buffer Test - Normal Load - Read Onlycassandra: Writesredis: SETredis: GETphpbench: PHP Benchmark Suitebuild2: Time To Compilecoremark: CoreMark Size 666 - Iterations Per Secondstress-ng: CPU Stressstress-ng: Cryptostress-ng: Memory Copyingstress-ng: Vector Mathstress-ng: Context Switchinggraphics-magick: Resizinggraphics-magick: Rotatejohn-the-ripper: MD5john-the-ripper: Blowfishnpb: EP.Cnpb: MG.Crodinia: OpenMP CFD Solverrodinia: OpenMP LavaMDparboil: OpenMP CUTCPEPYC m5a.16xlargeEPYC m5a.8xlargeXeon m5.16xlargeXeon m5.8xlargeArm m6g.16xlargeArm m6g.8xlargeEPYC c5a.16xlargeEPYC c5a.8xlarge42.28898846345160.3537761288111104632.501420554.3312397879.956829028.0061547197.135455.442397.51144985.8911406293.381539494151400045853601.0214236.8611.45217.9192.22289471.30451165200236.437774688781169289.371412453.75125036106.099405902.8139363493.952664.122250.6470231.524769836.9199648171873321655293.867637.1619.22035.7433.16487434.749141018561776.7056031721021485890.71829222.5817667567.313948869.00878611187.315650.5910529.53164932.9510663301.131833539405866746843878.9716504.2010.33317.5751.96213356.40774000325343.9101251183141546195.501837582.1217618985.650472840.2729115520.392792.925757.7081561.795400871.061147529205700023467440.319835.0114.80233.2112.82520374.356204359148403.1280282647461369966.881760567.0412817365.5081252073.1779717324.3711016.348628.77335135.4525407351.5134006511266667432371716.7225817.308.22611.0521.590478106.640107317345900.2596351282141360596.121759532.0012806073.003627000.2458553655.445510.308795.04167581.9113142582.76217565969725321964862.4012423.869.69920.6742.54317529.551165444539486.6444551967561332685.501676375.2213388260.4541194350.4775599615.867300.996371.34192241.1516211956.402236520334866763084763.8814839.5510.10912.7581.35826846.99889095293611.0580341096701365072.841717757.2113504873.282598183.2902864860.473644.814714.9095984.327301087.071455521169166731695381.228455.0414.34924.5881.944253OpenBenchmarking.org

Timed Linux Kernel Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 5.4Time To CompileEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlargeArm m6g.16xlargeArm m6g.8xlarge20406080100SE +/- 0.35, N = 6SE +/- 0.77, N = 3SE +/- 0.20, N = 3SE +/- 0.51, N = 3SE +/- 0.51, N = 3SE +/- 0.87, N = 3SE +/- 1.45, N = 3SE +/- 0.92, N = 329.5534.7542.2947.0056.4171.3074.36106.64

7-Zip Compression

Compress Speed Test

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 16.02Compress Speed TestArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.16xlargeArm m6g.8xlargeEPYC m5a.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge40K80K120K160K200KSE +/- 4153.99, N = 3SE +/- 32.19, N = 3SE +/- 232.74, N = 3SE +/- 293.58, N = 3SE +/- 142.35, N = 3SE +/- 84.54, N = 3SE +/- 180.00, N = 3SE +/- 310.73, N = 3204359165444141018107317988468909574000511651. (CXX) g++ options: -pipe -lpthread

PostgreSQL pgbench

Scaling: Buffer Test - Test: Normal Load - Mode: Read Only

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL pgbench 12.0Scaling: Buffer Test - Test: Normal Load - Mode: Read OnlyXeon m5.16xlargeEPYC c5a.16xlargeArm m6g.8xlargeEPYC m5a.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC m5a.8xlargeArm m6g.16xlarge120K240K360K480K600KSE +/- 780.81, N = 3SE +/- 323.59, N = 3SE +/- 1098.93, N = 3SE +/- 3550.73, N = 3SE +/- 388.76, N = 3SE +/- 456.60, N = 3SE +/- 315.67, N = 3SE +/- 5347.77, N = 3561776.71539486.64345900.26345160.35325343.91293611.06200236.44148403.131. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm

Apache Cassandra

Test: Writes

OpenBenchmarking.orgOp/s, More Is BetterApache Cassandra 3.11.4Test: WritesArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeArm m6g.8xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC m5a.8xlarge60K120K180K240K300KSE +/- 1849.41, N = 3SE +/- 3069.85, N = 3SE +/- 20850.03, N = 3SE +/- 1363.81, N = 3SE +/- 895.50, N = 3SE +/- 172.88, N = 3SE +/- 1438.92, N = 3SE +/- 649.61, N = 326474619675617210212881112821411831410967068878

Redis

Test: SET

OpenBenchmarking.orgRequests Per Second, More Is BetterRedis 5.0.5Test: SETXeon m5.8xlargeXeon m5.16xlargeArm m6g.16xlargeEPYC c5a.8xlargeArm m6g.8xlargeEPYC c5a.16xlargeEPYC m5a.8xlargeEPYC m5a.16xlarge300K600K900K1200K1500KSE +/- 21406.86, N = 3SE +/- 2211.17, N = 3SE +/- 8411.40, N = 3SE +/- 11513.00, N = 3SE +/- 5939.03, N = 3SE +/- 19011.69, N = 3SE +/- 9493.93, N = 3SE +/- 6089.90, N = 31546195.501485890.701369966.881365072.841360596.121332685.501169289.371104632.501. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3

Redis

Test: GET

OpenBenchmarking.orgRequests Per Second, More Is BetterRedis 5.0.5Test: GETXeon m5.8xlargeXeon m5.16xlargeArm m6g.16xlargeArm m6g.8xlargeEPYC c5a.8xlargeEPYC c5a.16xlargeEPYC m5a.16xlargeEPYC m5a.8xlarge400K800K1200K1600K2000KSE +/- 38131.52, N = 3SE +/- 31468.04, N = 3SE +/- 1789.57, N = 3SE +/- 1031.38, N = 3SE +/- 21419.20, N = 3SE +/- 15001.13, N = 11SE +/- 8434.32, N = 3SE +/- 4144.84, N = 31837582.121829222.581760567.041759532.001717757.211676375.221420554.331412453.751. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3

PHPBench

PHP Benchmark Suite

OpenBenchmarking.orgScore, More Is BetterPHPBench 0.8.1PHP Benchmark SuiteXeon m5.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC c5a.16xlargeArm m6g.16xlargeArm m6g.8xlargeEPYC m5a.8xlargeEPYC m5a.16xlarge40K80K120K160K200KSE +/- 1566.49, N = 3SE +/- 162.64, N = 3SE +/- 262.79, N = 3SE +/- 1621.31, N = 12SE +/- 35.18, N = 3SE +/- 58.00, N = 3SE +/- 377.35, N = 3SE +/- 1782.09, N = 3176675176189135048133882128173128060125036123978

Build2

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterBuild2 0.12Time To CompileEPYC c5a.16xlargeArm m6g.16xlargeXeon m5.16xlargeArm m6g.8xlargeEPYC c5a.8xlargeEPYC m5a.16xlargeXeon m5.8xlargeEPYC m5a.8xlarge20406080100SE +/- 0.02, N = 3SE +/- 0.15, N = 3SE +/- 0.36, N = 3SE +/- 0.18, N = 3SE +/- 0.14, N = 3SE +/- 0.83, N = 3SE +/- 0.12, N = 3SE +/- 0.42, N = 360.4565.5167.3173.0073.2879.9685.65106.10

Coremark

CoreMark Size 666 - Iterations Per Second

OpenBenchmarking.orgIterations/Sec, More Is BetterCoremark 1.0CoreMark Size 666 - Iterations Per SecondArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeArm m6g.8xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge300K600K900K1200K1500KSE +/- 2255.41, N = 3SE +/- 1719.33, N = 3SE +/- 4847.04, N = 3SE +/- 468.92, N = 3SE +/- 125.09, N = 3SE +/- 908.61, N = 3SE +/- 941.88, N = 3SE +/- 252.95, N = 31252073.181194350.48948869.01829028.01627000.25598183.29472840.27405902.811. (CC) gcc options: -O2 -lrt" -lrt

Stress-NG

Test: CPU Stress

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: CPU StressXeon m5.16xlargeEPYC c5a.16xlargeArm m6g.16xlargeEPYC m5a.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeArm m6g.8xlargeEPYC m5a.8xlarge2K4K6K8K10KSE +/- 3.06, N = 3SE +/- 39.32, N = 3SE +/- 4.38, N = 3SE +/- 9.72, N = 3SE +/- 6.44, N = 3SE +/- 19.59, N = 3SE +/- 2.09, N = 3SE +/- 2.75, N = 311187.319615.867324.377197.135520.394860.473655.443493.951. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

Stress-NG

Test: Crypto

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: CryptoArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.16xlargeArm m6g.8xlargeEPYC m5a.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge2K4K6K8K10KSE +/- 1.26, N = 3SE +/- 0.25, N = 3SE +/- 0.53, N = 3SE +/- 1.85, N = 3SE +/- 0.31, N = 3SE +/- 2.06, N = 3SE +/- 0.80, N = 3SE +/- 0.40, N = 311016.347300.995650.595510.305455.443644.812792.922664.121. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

Stress-NG

Test: Memory Copying

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.11.07Test: Memory CopyingXeon m5.16xlargeArm m6g.8xlargeArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC m5a.16xlargeEPYC m5a.8xlarge2K4K6K8K10KSE +/- 22.69, N = 3SE +/- 8.37, N = 3SE +/- 9.79, N = 3SE +/- 41.89, N = 3SE +/- 29.24, N = 3SE +/- 23.97, N = 3SE +/- 4.75, N = 3SE +/- 7.27, N = 310529.538795.048628.776371.345757.704714.902397.512250.641. (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 MathArm m6g.16xlargeEPYC c5a.16xlargeArm m6g.8xlargeXeon m5.16xlargeEPYC m5a.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge70K140K210K280K350KSE +/- 2.35, N = 3SE +/- 23.54, N = 3SE +/- 1.36, N = 3SE +/- 77.26, N = 3SE +/- 310.18, N = 3SE +/- 58.01, N = 3SE +/- 23.81, N = 3SE +/- 15.59, N = 3335135.45192241.15167581.91164932.95144985.8995984.3281561.7970231.521. (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 SwitchingArm m6g.16xlargeEPYC c5a.16xlargeArm m6g.8xlargeEPYC m5a.16xlargeXeon m5.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge5M10M15M20M25MSE +/- 124150.70, N = 3SE +/- 83665.67, N = 3SE +/- 15489.60, N = 3SE +/- 12844.44, N = 3SE +/- 207947.62, N = 3SE +/- 10376.30, N = 3SE +/- 103473.24, N = 3SE +/- 294612.10, N = 325407351.5116211956.4013142582.7611406293.3810663301.137301087.075400871.064769836.911. (CC) gcc options: -O2 -std=gnu99 -lm -lcrypt -lrt -lz -ldl -lpthread -lc

GraphicsMagick

Operation: Resizing

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.33Operation: ResizingArm m6g.16xlargeEPYC c5a.16xlargeArm m6g.8xlargeXeon m5.16xlargeEPYC m5a.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge7001400210028003500SE +/- 12.17, N = 3SE +/- 36.61, N = 3SE +/- 8.35, N = 3SE +/- 12.91, N = 3SE +/- 25.69, N = 3SE +/- 1.20, N = 3SE +/- 2.00, N = 334002236217518331539145511479961. (CC) gcc options: -fopenmp -O2 -pthread -ljpeg -lz -lm -lpthread

GraphicsMagick

Operation: Rotate

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.33Operation: RotateArm m6g.8xlargeArm m6g.16xlargeXeon m5.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC c5a.16xlargeEPYC m5a.16xlargeEPYC m5a.8xlarge140280420560700SE +/- 1.53, N = 3SE +/- 4.10, N = 3SE +/- 5.29, N = 3SE +/- 7.02, N = 3SE +/- 1.86, N = 3SE +/- 2.85, N = 3SE +/- 4.67, N = 3SE +/- 0.67, N = 36596515395295215204944811. (CC) gcc options: -fopenmp -O2 -pthread -ljpeg -lz -lm -lpthread

John The Ripper

Test: MD5

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 1.9.0-jumbo-1Test: MD5Xeon m5.16xlargeEPYC c5a.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC m5a.16xlargeArm m6g.16xlargeEPYC m5a.8xlargeArm m6g.8xlarge900K1800K2700K3600K4500KSE +/- 666.67, N = 3SE +/- 5364.49, N = 3SE +/- 1527.53, N = 3SE +/- 1333.33, N = 3SE +/- 5507.57, N = 3SE +/- 3844.19, N = 3SE +/- 830.17, N = 3SE +/- 1593.87, N = 3405866733486672057000169166715140001266667718733697253-m64-m64-m64-m64-m64-m641. (CC) gcc options: -lssl -lcrypto -fopenmp -pthread -lm -lz -ldl -lcrypt

John The Ripper

Test: Blowfish

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 1.9.0-jumbo-1Test: BlowfishEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeArm m6g.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeArm m6g.8xlargeEPYC m5a.8xlarge14K28K42K56K70KSE +/- 23.07, N = 3SE +/- 24.33, N = 3SE +/- 112.47, N = 3SE +/- 14.73, N = 3SE +/- 3.67, N = 3SE +/- 2.67, N = 3SE +/- 6.94, N = 3SE +/- 3.67, N = 36308446843458534323731695234672196421655-m64-m64-m64-m64-m64-m641. (CC) gcc options: -lssl -lcrypto -fopenmp -pthread -lm -lz -ldl -lcrypt

NAS Parallel Benchmarks

Test / Class: EP.C

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: EP.CArm m6g.16xlargeXeon m5.16xlargeArm m6g.8xlargeEPYC c5a.16xlargeEPYC m5a.16xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC m5a.8xlarge400800120016002000SE +/- 1.33, N = 3SE +/- 0.97, N = 3SE +/- 0.52, N = 3SE +/- 0.06, N = 3SE +/- 0.45, N = 3SE +/- 0.28, N = 3SE +/- 0.40, N = 3SE +/- 1.36, N = 31716.72878.97862.40763.88601.02440.31381.22293.861. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

NAS Parallel Benchmarks

Test / Class: MG.C

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.4Test / Class: MG.CArm m6g.16xlargeXeon m5.16xlargeEPYC c5a.16xlargeEPYC m5a.16xlargeArm m6g.8xlargeXeon m5.8xlargeEPYC c5a.8xlargeEPYC m5a.8xlarge6K12K18K24K30KSE +/- 1.02, N = 3SE +/- 13.26, N = 3SE +/- 64.13, N = 3SE +/- 200.97, N = 3SE +/- 4.54, N = 3SE +/- 2.10, N = 3SE +/- 3.38, N = 3SE +/- 268.64, N = 325817.3016504.2014839.5514236.8612423.869835.018455.047637.161. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

Rodinia

Test: OpenMP CFD Solver

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 2.4Test: OpenMP CFD SolverArm m6g.16xlargeArm m6g.8xlargeEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge510152025SE +/- 0.048, N = 3SE +/- 0.038, N = 3SE +/- 0.070, N = 3SE +/- 0.043, N = 3SE +/- 0.041, N = 3SE +/- 0.025, N = 3SE +/- 0.014, N = 3SE +/- 0.083, N = 38.2269.69910.10910.33311.45214.34914.80219.2201. (CXX) g++ options: -O3 -fopenmp

Rodinia

Test: OpenMP LavaMD

OpenBenchmarking.orgSeconds, Fewer Is BetterRodinia 2.4Test: OpenMP LavaMDArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeArm m6g.8xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge816243240SE +/- 0.01, N = 3SE +/- 0.01, N = 3SE +/- 0.11, N = 3SE +/- 0.01, N = 3SE +/- 0.00, N = 3SE +/- 0.01, N = 3SE +/- 0.31, N = 3SE +/- 0.02, N = 311.0512.7617.5817.9220.6724.5933.2135.741. (CXX) g++ options: -O3 -fopenmp

Parboil

Test: OpenMP CUTCP

OpenBenchmarking.orgSeconds, Fewer Is BetterParboil 2.5Test: OpenMP CUTCPEPYC c5a.16xlargeArm m6g.16xlargeEPYC c5a.8xlargeXeon m5.16xlargeEPYC m5a.16xlargeArm m6g.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge0.71211.42422.13632.84843.5605SE +/- 0.001329, N = 3SE +/- 0.001653, N = 3SE +/- 0.020206, N = 3SE +/- 0.008492, N = 3SE +/- 0.015619, N = 3SE +/- 0.009903, N = 3SE +/- 0.001005, N = 3SE +/- 0.030016, N = 31.3582681.5904781.9442531.9621332.2228942.5431752.8252033.1648741. (CXX) g++ options: -lm -lpthread -lgomp -O3 -ffast-math -fopenmp

Geometric Mean Of All Test Results

Result Composite - Graviton2 EC2 AMD EPYC 7002 Rome Comparison Benchmarks

OpenBenchmarking.orgGeometric Mean, More Is BetterGeometric Mean Of All Test ResultsResult Composite - Graviton2 EC2 AMD EPYC 7002 Rome Comparison BenchmarksArm m6g.16xlargeEPYC c5a.16xlargeXeon m5.16xlargeEPYC m5a.16xlargeArm m6g.8xlargeEPYC c5a.8xlargeXeon m5.8xlargeEPYC m5a.8xlarge2K4K6K8K10K11217.7310933.0310383.757793.097650.796867.586500.374779.42

Phoronix Test Suite v10.8.4