AMD EPYC Rome vs. Xeon Spectre Mitigation Benchmarks

benchmarks for a future article with Spectre, Meltdown, and friends mitigation toggling.

HTML result view exported from: https://openbenchmarking.org/result/1908120-AS-MITIGATIO88&grr.

AMD EPYC Rome vs. Xeon Spectre Mitigation BenchmarksProcessorMotherboardChipsetMemoryDiskGraphicsMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLCompilerFile-SystemScreen Resolution2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations2 x AMD EPYC 7742 64-Core @ 2.25GHz (128 Cores / 256 Threads)AMD DAYTONA_X (RDY1001C BIOS)AMD Device 148016 x 32 GB DDR4-3200MT/s 36ASF4G72PZ-3G2E2280GB INTEL SSDPED1D280GAllvmpipe 504GBVE2282 x Mellanox MT27710Ubuntu 19.045.2.0-050200rc7-generic (x86_64) 20190630GNOME Shell 3.32.1X Server 1.20.4modesetting 1.20.43.3 Mesa 19.0.2 (LLVM 8.0 128 bits)GCC 8.3.0ext41920x10802 x Intel Xeon Platinum 8280 @ 4.00GHz (56 Cores / 112 Threads)GIGABYTE MD61-SC2-00 v01000100 (T15 BIOS)Intel Sky Lake-E DMI3 Registers12 x 32 GB DDR4-2933MT/s HMA84GR7CJR4N-WMllvmpipe 377GB2 x Intel X722 for 1GbE + 2 x QLogic FastLinQ QL41000 10/25/40/50GbE3.3 Mesa 19.0.2 (LLVM 8.0 256 bits)GNOME Shell 3.32.2OpenBenchmarking.orgCompiler Details- --build=x86_64-linux-gnu --disable-vtable-verify --disable-werror --enable-bootstrap --enable-checking=release --enable-clocale=gnu --enable-default-pie --enable-gnu-unique-object --enable-languages=c,ada,c++,go,brig,d,fortran,objc,obj-c++ --enable-libmpx --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none --enable-plugin --enable-shared --enable-threads=posix --host=x86_64-linux-gnu --program-prefix=x86_64-linux-gnu- --target=x86_64-linux-gnu --with-abi=m64 --with-arch-32=i686 --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib --with-tune=generic --without-cuda-driver -v Processor Details- 2 x EPYC 7742: Scaling Governor: acpi-cpufreq ondemand- 2 x EPYC 7742 - No Mitigations: Scaling Governor: acpi-cpufreq ondemand- 2 x Xeon Platinum 8280: Scaling Governor: intel_pstate powersave- 2 x Xeon Platinum 8280 - No Mitigations: Scaling Governor: intel_pstate powersaveJava Details- 2 x EPYC 7742: OpenJDK Runtime Environment (build 11.0.3+7-Ubuntu-1ubuntu219.04.1)- 2 x EPYC 7742 - No Mitigations: OpenJDK Runtime Environment (build 11.0.4+11-post-Ubuntu-1ubuntu219.04)- 2 x Xeon Platinum 8280: OpenJDK Runtime Environment (build 11.0.3+7-Ubuntu-1ubuntu219.04.1)- 2 x Xeon Platinum 8280 - No Mitigations: OpenJDK Runtime Environment (build 11.0.4+11-post-Ubuntu-1ubuntu219.04)Python Details- Python 2.7.16 + Python 3.7.3Security Details- 2 x EPYC 7742: l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Full AMD retpoline IBPB: conditional IBRS_FW STIBP: conditional RSB filling - 2 x EPYC 7742 - No Mitigations: l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled - 2 x Xeon Platinum 8280: l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling - 2 x Xeon Platinum 8280 - No Mitigations: l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Vulnerable + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Vulnerable IBPB: disabled STIBP: disabled Disk Details- 2 x Xeon Platinum 8280: NONE / errors=remount-ro,relatime,rw

AMD EPYC Rome vs. Xeon Spectre Mitigation Benchmarksbuild-gcc: Time To Compilerenaissance: Savina Reactors.IOnamd: ATPase Simulation - 327,506 Atomsjohn-the-ripper: Blowfishrenaissance: Scala Dottybuild-llvm: Time To Compilehackbench: 32 - Processredis: GETgimp: unsharp-maskgimp: auto-levelsgimp: rotategimp: resizenpb: EP.Cctx-clock: Context Switch Time2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations723.5130851.620.265951791667515.5179.3714.722230671.1728.2723.8717.6211.984144.36135720.9728583.130.265771849757125.4079.6014.332138833.9427.2022.7116.6011.624125.22135891.1127155.250.36046842386790.53106.7936.682541994.8528.7523.7118.3912.662494.40135878.3325932.140.35940838566691.56103.6937.582652682.3326.4721.9217.6811.852509.97134OpenBenchmarking.org

Timed GCC Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed GCC Compilation 8.2Time To Compile2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations2004006008001000SE +/- 2.80, N = 3SE +/- 0.63, N = 3SE +/- 0.69, N = 3SE +/- 1.09, N = 3723.51720.97891.11878.33

Renaissance

Test: Savina Reactors.IO

OpenBenchmarking.orgms, Fewer Is BetterRenaissance 0.9.0Test: Savina Reactors.IO2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations7K14K21K28K35KSE +/- 685.37, N = 40SE +/- 526.43, N = 40SE +/- 424.03, N = 40SE +/- 348.41, N = 4030851.6228583.1327155.2525932.14

NAMD

ATPase Simulation - 327,506 Atoms

OpenBenchmarking.orgdays/ns, Fewer Is BetterNAMD 2.13b1ATPase Simulation - 327,506 Atoms2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations0.08110.16220.24330.32440.4055SE +/- 0.00268, N = 3SE +/- 0.00317, N = 3SE +/- 0.00079, N = 7SE +/- 0.00013, N = 150.265950.265770.360460.35940

John The Ripper

Test: Blowfish

OpenBenchmarking.orgReal C/S, More Is BetterJohn The Ripper 1.9.0-jumbo-1Test: Blowfish2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations40K80K120K160K200KSE +/- 2249.31, N = 13SE +/- 680.60, N = 3SE +/- 465.36, N = 3SE +/- 461.73, N = 317916618497584238838561. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -pthread -lm -lz -ldl -lcrypt -lbz2

Renaissance

Test: Scala Dotty

OpenBenchmarking.orgms, Fewer Is BetterRenaissance 0.9.0Test: Scala Dotty2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations16003200480064008000SE +/- 68.73, N = 40SE +/- 37.73, N = 8SE +/- 49.97, N = 8SE +/- 50.57, N = 87515.517125.406790.536691.56

Timed LLVM Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed LLVM Compilation 6.0.1Time To Compile2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations2040608010079.3779.60106.79103.69

Hackbench

Count: 32 - Type: Process

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 32 - Type: Process2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations918273645SE +/- 0.17, N = 6SE +/- 0.08, N = 3SE +/- 0.62, N = 3SE +/- 0.52, N = 314.7214.3336.6837.581. (CC) gcc options: -lpthread

Redis

Test: GET

OpenBenchmarking.orgRequests Per Second, More Is BetterRedis 4.0.8Test: GET2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations600K1200K1800K2400K3000KSE +/- 14491.01, N = 3SE +/- 52801.81, N = 12SE +/- 24940.97, N = 15SE +/- 14701.59, N = 32230671.172138833.942541994.852652682.331. (CC) gcc options: -ggdb -rdynamic -lm -ldl -pthread

GIMP

Test: unsharp-mask

OpenBenchmarking.orgSeconds, Fewer Is BetterGIMP 2.10.8Test: unsharp-mask2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations714212835SE +/- 0.35, N = 3SE +/- 0.19, N = 3SE +/- 0.11, N = 3SE +/- 0.13, N = 328.2727.2028.7526.47

GIMP

Test: auto-levels

OpenBenchmarking.orgSeconds, Fewer Is BetterGIMP 2.10.8Test: auto-levels2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations612182430SE +/- 0.35, N = 4SE +/- 0.05, N = 3SE +/- 0.14, N = 3SE +/- 0.11, N = 323.8722.7123.7121.92

GIMP

Test: rotate

OpenBenchmarking.orgSeconds, Fewer Is BetterGIMP 2.10.8Test: rotate2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations510152025SE +/- 0.02, N = 3SE +/- 0.01, N = 3SE +/- 0.11, N = 3SE +/- 0.01, N = 317.6216.6018.3917.68

GIMP

Test: resize

OpenBenchmarking.orgSeconds, Fewer Is BetterGIMP 2.10.8Test: resize2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations3691215SE +/- 0.08, N = 3SE +/- 0.10, N = 3SE +/- 0.14, N = 3SE +/- 0.01, N = 311.9811.6212.6611.85

NAS Parallel Benchmarks

Test / Class: EP.C

OpenBenchmarking.orgTotal Mop/s, More Is BetterNAS Parallel Benchmarks 3.3.1Test / Class: EP.C2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations9001800270036004500SE +/- 8.63, N = 3SE +/- 19.98, N = 3SE +/- 5.65, N = 3SE +/- 6.26, N = 34144.364125.222494.402509.971. (F9X) gfortran options: -O3 -march=native -pthread -lmpi_usempif08 -lmpi_mpifh -lmpi

ctx_clock

Context Switch Time

OpenBenchmarking.orgClocks, Fewer Is Betterctx_clockContext Switch Time2 x EPYC 77422 x EPYC 7742 - No Mitigations2 x Xeon Platinum 82802 x Xeon Platinum 8280 - No Mitigations306090120150SE +/- 0.67, N = 3135135135134

Phoronix Test Suite v10.8.5