Cavium Thunderx vs Amlogic S905

David vs Goliath? (fixed stream to fit in RAM)

HTML result view exported from: https://openbenchmarking.org/result/1704098-RI-1702269RI43&grs&sro.

Cavium Thunderx vs Amlogic S905ProcessorMotherboardMemoryDiskGraphicsNetworkOSKernelDisplay ServerDisplay DriverOpenGLCompilerFile-SystemScreen ResolutionThunderxS905 GCC7Unknown (96 Cores)cavium ThunderX Unknown262144MB240GB INTEL SSDSC2BB24 + 4 x 1000GB HGST HTS721010A9ASPEED ASPEED FamilyCavium THUNDERX Interface + Cavium THUNDERX BGXUbuntu 16.044.2.0-pmu1-g749be54 (aarch64)X Server 1.18.3NVIDIA 352.55)1.4GCC 5.3.0 + Clang 3.9.0 + LLVM 3.9.0svnext42160x1920Unknown @ 1.54GHz (4 Cores)Amlogic2048MB16GB NCard + 32GB 00000OSD3.14.79-vegas95 (aarch64)GCC 6.3.1 20170316 + Clang 3.8.0-2ubuntu4 + LLVM 3.8.01280x1440OpenBenchmarking.orgCompiler Details- Thunderx: --disable-libsanitizer --disable-libsanitizer --disable-werror --disable-werror --enable-languages=c,c++,fortran,lto,objc --enable-languages=c,c++,fortran,objc --enable-lto --enable-lto --enable-modules --enable-modules --no-create --no-recursion --with-cpu=thunderx --with-cpu=thunderx - S905 GCC7: --build=aarch64-linux-gnu --disable-bootstrap --disable-browser-plugin --disable-libquadmath --disable-werror --enable-checking=release --enable-clocale=gnu --enable-gnu-unique-object --enable-languages=c,c++,fortran --enable-libstdcxx-time=yes --enable-multiarch --enable-nls --enable-plugin --enable-shared --enable-threads=posix --host=aarch64-linux-gnu --target=aarch64-linux-gnu --with-arch-directory=aarch64 --with-default-libstdcxx-abi=new Processor Details- S905 GCC7: Scaling Governor: meson_cpufreq interactive

Cavium Thunderx vs Amlogic S905stream: Scaleramspeed: Integer Scalecachebench: Read Cacheramspeed: Floating-Point Addramspeed: Integer Copyramspeed: Integer Addcachebench: Write Cachestream: Copystream: AddThunderxS905 GCC744096.672821.403311.912733.733807.173118.763384.2140248.5244608.564188.824606.182159.193820.274667.043782.993375.054129.444089.98OpenBenchmarking.org

Stream

Scale

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17ScaleS905 GCC7Thunderx9K18K27K36K45KSE +/- 26.77, N = 5SE +/- 454.77, N = 34188.8244096.67-O3 -march=native1. (CC) gcc options: -fopenmp

RAMspeed SMP

Integer Scale

OpenBenchmarking.orgMB/s, More Is BetterRAMspeed SMP 3.5.0Integer ScaleS905 GCC7Thunderx100020003000400050004606.182821.40

CacheBench

Read Cache

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchRead CacheS905 GCC7Thunderx7001400210028003500SE +/- 2.25, N = 3SE +/- 0.04, N = 32159.193311.911. (CC) gcc options: -lrt

RAMspeed SMP

Floating-Point Add

OpenBenchmarking.orgMB/s, More Is BetterRAMspeed SMP 3.5.0Floating-Point AddS905 GCC7Thunderx80016002400320040003820.272733.73

RAMspeed SMP

Integer Copy

OpenBenchmarking.orgMB/s, More Is BetterRAMspeed SMP 3.5.0Integer CopyS905 GCC7Thunderx100020003000400050004667.043807.17

RAMspeed SMP

Integer Add

OpenBenchmarking.orgMB/s, More Is BetterRAMspeed SMP 3.5.0Integer AddS905 GCC7Thunderx80016002400320040003782.993118.76

CacheBench

Write Cache

OpenBenchmarking.orgMB/s, More Is BetterCacheBenchWrite CacheS905 GCC7Thunderx7001400210028003500SE +/- 2.70, N = 3SE +/- 0.06, N = 33375.053384.211. (CC) gcc options: -lrt

Stream

Copy

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17CopyS905 GCC7Thunderx9K18K27K36K45KSE +/- 11.16, N = 5SE +/- 1983.34, N = 54129.4440248.52-O3 -march=native1. (CC) gcc options: -fopenmp

Stream

Add

OpenBenchmarking.orgMB/s, More Is BetterStream 2013-01-17AddS905 GCC7Thunderx10K20K30K40K50KSE +/- 15.19, N = 5SE +/- 2003.35, N = 104089.9844608.56-O3 -march=native1. (CC) gcc options: -fopenmp

Phoronix Test Suite v10.8.4