EPYC 72F3 Retpolines Change In Linux 5.17

AMD EPYC 72F3 Retpoline benchmarks by Michael Larabel.

HTML result view exported from: https://openbenchmarking.org/result/2203111-NE-EPYC72F3R09&sro&grs.

EPYC 72F3 Retpolines Change In Linux 5.17ProcessorMotherboardChipsetMemoryDiskGraphicsMonitorNetworkOSKernelDesktopDisplay ServerCompilerFile-SystemScreen ResolutionNew Default - Generic RetpolinesAMD Retpolines Former DefaultAMD EPYC 72F3 8-Core @ 3.70GHz (8 Cores / 16 Threads)Supermicro H12SSL-i v1.01 (2.0 BIOS)AMD Starship/Matisse128GB3841GB Micron_9300_MTFDHAL3T8TDP + 1000GB Corsair Force MP600ASPEEDVE2282 x Broadcom NetXtreme BCM5720 2-port PCIeUbuntu 21.045.17.0-051700rc7daily20220310-generic (x86_64)GNOME Shell 3.38.4X ServerGCC 10.3.0ext41920x1080OpenBenchmarking.orgKernel Details- Transparent Huge Pages: madviseCompiler 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++,m2 --enable-libphobos-checking=release --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-link-mutex --enable-multiarch --enable-multilib --enable-nls --enable-objc-gc=auto --enable-offload-targets=nvptx-none=/build/gcc-10-gDeRY6/gcc-10-10.3.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-10-gDeRY6/gcc-10-10.3.0/debian/tmp-gcn/usr,hsa --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-build-config=bootstrap-lto-lean --with-default-libstdcxx-abi=new --with-gcc-major-version-only --with-multilib-list=m32,m64,mx32 --with-target-system-zlib=auto --with-tune=generic --without-cuda-driver -v Disk Details- NONE / errors=remount-ro,relatime,rw / Block Size: 4096Processor Details- Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001119Java Details- OpenJDK Runtime Environment (build 11.0.13+8-Ubuntu-0ubuntu1.21.04)Python Details- Python 3.9.5Security Details- New Default - Generic Retpolines: 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 usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Retpolines IBPB: conditional IBRS_FW STIBP: always-on RSB filling + srbds: Not affected + tsx_async_abort: Not affected - AMD Retpolines Former Default: 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 usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Vulnerable: LFENCE + srbds: Not affected + tsx_async_abort: Not affected

EPYC 72F3 Retpolines Change In Linux 5.17sockperf: Throughputosbench: Create Filespostmark: Disk Transaction Performancekeydb: hackbench: 8 - Processebizzy: stress-ng: SENDFILEnginx: 200nginx: 100wireguard: node-express-loadtest: sqlite-speedtest: Timed Time - Size 1,000redis: GETcompress-lz4: 3 - Decompression Speedfio: Rand Write - IO_uring - No - Yes - 4KB - Default Test Directoryleveldb: Seq Fillcompress-zstd: 8 - Decompression Speeddacapobench: Tradebeansnginx: 20rocksdb: Seq Fillrocksdb: Rand Fillrenaissance: In-Memory Database Shootoutosbench: Create Threadsstress-ng: Socket Activityleveldb: Seq Filldacapobench: Jythonhackbench: 32 - Processcompress-lz4: 9 - Decompression Speedosbench: Launch Programsfio: Rand Read - IO_uring - No - Yes - 4KB - Default Test Directoryrocksdb: Update Randleveldb: Overwriteleveldb: Overwritestress-ng: Semaphorescompress-pbzip2: FreeBSD-13.0-RELEASE-amd64-memstick.img Compressionapache: 100compress-7zip: Compression Ratingbuild-eigen: Time To Compiledacapobench: H2leveldb: Rand Deletedacapobench: Tradesoapmemtier-benchmark: Redisapache: 500leveldb: Fill Syncleveldb: Seek Randhackbench: 8 - Threadhackbench: 16 - Processdraco: Church Facadecompress-zstd: 19 - Decompression Speedstress-ng: Context Switchingsockperf: Latency Under Loadrocksdb: Read While Writingpgbench: 100 - 50 - Read Only - Average Latencyleveldb: Rand Fillcompress-zstd: 19 - Compression Speedpgbench: 100 - 100 - Read Writecassandra: Writespgbench: 100 - 100 - Read Write - Average Latencyosbench: Create Processesrenaissance: Finagle HTTP Requestsleveldb: Rand Fillnode-web-tooling: rocksdb: Rand Readcompress-zstd: 8 - Compression Speeddraco: Lionleveldb: Hot Readrocksdb: Read Rand Write Randleveldb: Rand Readpgbench: 100 - 100 - Read Only - Average Latencybuild-mplayer: Time To Compilehackbench: 16 - Threadinfluxdb: 4 - 10000 - 2,5000,1 - 10000compress-lz4: 9 - Compression Speedcompress-lz4: 3 - Compression Speedpgbench: 100 - 50 - Read Onlystress-ng: System V Message Passingredis: SETbuild-gdb: Time To Compilecompress-7zip: Decompression Ratinginfluxdb: 64 - 10000 - 2,5000,1 - 10000openssl: RSA4096build-godot: Time To Compileapache: 200pgbench: 100 - 100 - Read Onlybuild-linux-kernel: defconfigopenssl: SHA256openssl: RSA4096New Default - Generic RetpolinesAMD Retpolines Former Default53972414.7065386944136793.5929.950619271200771.29213684.40215190.56257.104669250.4832506720.2514455.122393316.94255.15762215287.796575236055724788.217.37674110751.13105.1163761119.83114586.249.362977239667393101103.36517.11326767.9210.78482500.758635761.685426899.87636601324327.3280298.151820.48817.13931.27558.78266973694.54296558.493.66617348780.17017.251.610383864209.64143.1402532378.7103.21711.26613317211343.6508810.537139100010.5770.33130.70962.2031326859.159.5761.242951794521841.841998825.9254.955693931638195.72372.3111.50786362.1830252378.61813238722860155003.158752313.7230147282142573.3428.828596554193652.98206147.39207615.65248.756691049.0072438721.7114817.322950016.54160.15634210605.866707606172504878.717.70337410567.09106.9463697117.98414807.448.627853243000398526104.60816.91342170.8410.66483397.578721061.098422898.95336281313054.1180970.151806.75917.01131.04658.35666493668.54326244.503.64117465240.16917.151.910440859569.59043.3627772366.9103.68011.31615862321338.6507010.502138658610.5450.33030.80262.0291323460.259.6961.362957304530010.921995692.8955.032693081639910.02374.1111.58086408.4830239978.59113234833737155031.8OpenBenchmarking.org

Sockperf

Test: Throughput

OpenBenchmarking.orgMessages Per Second, More Is BetterSockperf 3.7Test: ThroughputAMD Retpolines Former DefaultNew Default - Generic Retpolines130K260K390K520K650KSE +/- 5255.38, N = 25SE +/- 4906.55, N = 255875235397241. (CXX) g++ options: --param -O3 -rdynamic -ldl -lpthread

OSBench

Test: Create Files

OpenBenchmarking.orgus Per Event, Fewer Is BetterOSBenchTest: Create FilesAMD Retpolines Former DefaultNew Default - Generic Retpolines48121620SE +/- 0.01, N = 3SE +/- 0.04, N = 313.7214.711. (CC) gcc options: -lm

PostMark

Disk Transaction Performance

OpenBenchmarking.orgTPS, More Is BetterPostMark 1.51Disk Transaction PerformanceAMD Retpolines Former DefaultNew Default - Generic Retpolines16003200480064008000SE +/- 70.00, N = 3728269441. (CC) gcc options: -O3

KeyDB

OpenBenchmarking.orgOps/sec, More Is BetterKeyDB 6.2.0AMD Retpolines Former DefaultNew Default - Generic Retpolines30K60K90K120K150KSE +/- 1055.00, N = 15SE +/- 1054.95, N = 10142573.34136793.591. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre

Hackbench

Count: 8 - Type: Process

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 8 - Type: ProcessAMD Retpolines Former DefaultNew Default - Generic Retpolines714212835SE +/- 0.17, N = 3SE +/- 0.17, N = 328.8329.951. (CC) gcc options: -lpthread

ebizzy

OpenBenchmarking.orgRecords/s, More Is Betterebizzy 0.3AMD Retpolines Former DefaultNew Default - Generic Retpolines130K260K390K520K650KSE +/- 6095.03, N = 15SE +/- 5150.50, N = 155965546192711. (CC) gcc options: -pthread -lpthread -O3 -march=native

Stress-NG

Test: SENDFILE

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.13.02Test: SENDFILEAMD Retpolines Former DefaultNew Default - Generic Retpolines40K80K120K160K200KSE +/- 443.51, N = 3SE +/- 510.20, N = 3193652.98200771.291. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic

nginx

Concurrent Requests: 200

OpenBenchmarking.orgRequests Per Second, More Is Betternginx 1.21.1Concurrent Requests: 200AMD Retpolines Former DefaultNew Default - Generic Retpolines50K100K150K200K250KSE +/- 466.89, N = 3SE +/- 275.98, N = 3206147.39213684.401. (CC) gcc options: -ldl -lpthread -lcrypt -lz -O3 -march=native

nginx

Concurrent Requests: 100

OpenBenchmarking.orgRequests Per Second, More Is Betternginx 1.21.1Concurrent Requests: 100AMD Retpolines Former DefaultNew Default - Generic Retpolines50K100K150K200K250KSE +/- 369.20, N = 3SE +/- 1627.81, N = 3207615.65215190.561. (CC) gcc options: -ldl -lpthread -lcrypt -lz -O3 -march=native

WireGuard + Linux Networking Stack Stress Test

OpenBenchmarking.orgSeconds, Fewer Is BetterWireGuard + Linux Networking Stack Stress TestAMD Retpolines Former DefaultNew Default - Generic Retpolines60120180240300SE +/- 0.99, N = 3SE +/- 1.39, N = 3248.76257.10

Node.js Express HTTP Load Test

OpenBenchmarking.orgRequests Per Second, More Is BetterNode.js Express HTTP Load TestAMD Retpolines Former DefaultNew Default - Generic Retpolines15003000450060007500SE +/- 43.97, N = 3SE +/- 34.82, N = 369106692

SQLite Speedtest

Timed Time - Size 1,000

OpenBenchmarking.orgSeconds, Fewer Is BetterSQLite Speedtest 3.30Timed Time - Size 1,000AMD Retpolines Former DefaultNew Default - Generic Retpolines1122334455SE +/- 0.05, N = 3SE +/- 0.10, N = 349.0150.481. (CC) gcc options: -O2 -ldl -lz -lpthread

Redis

Test: GET

OpenBenchmarking.orgRequests Per Second, More Is BetterRedis 6.0.9Test: GETAMD Retpolines Former DefaultNew Default - Generic Retpolines500K1000K1500K2000K2500KSE +/- 22769.87, N = 6SE +/- 15897.48, N = 32438721.712506720.251. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3

LZ4 Compression

Compression Level: 3 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Decompression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines3K6K9K12K15KSE +/- 22.12, N = 3SE +/- 24.84, N = 1514817.314455.11. (CC) gcc options: -O3

Flexible IO Tester

Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory

OpenBenchmarking.orgIOPS, More Is BetterFlexible IO Tester 3.29Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test DirectoryAMD Retpolines Former DefaultNew Default - Generic Retpolines50K100K150K200K250KSE +/- 2958.04, N = 4SE +/- 3027.60, N = 152295002239331. (CC) gcc options: -rdynamic -ltcmalloc -ll -lnuma -lrt -lz -lpthread -lm -ldl -laio -lcurl -lssl -lcrypto -std=gnu99 -ffast-math -include -O3 -fcommon -U_FORTIFY_SOURCE -march=native

LevelDB

Benchmark: Sequential Fill

OpenBenchmarking.orgMB/s, More Is BetterLevelDB 1.22Benchmark: Sequential FillAMD Retpolines Former DefaultNew Default - Generic Retpolines48121620SE +/- 0.09, N = 3SE +/- 0.07, N = 316.516.91. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

Zstd Compression

Compression Level: 8 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 8 - Decompression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines9001800270036004500SE +/- 4.29, N = 3SE +/- 54.59, N = 34160.14255.11. (CC) gcc options: -O3 -pthread -lz -llzma

DaCapo Benchmark

Java Test: Tradebeans

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: TradebeansAMD Retpolines Former DefaultNew Default - Generic Retpolines12002400360048006000SE +/- 38.99, N = 4SE +/- 34.00, N = 456345762

nginx

Concurrent Requests: 20

OpenBenchmarking.orgRequests Per Second, More Is Betternginx 1.21.1Concurrent Requests: 20AMD Retpolines Former DefaultNew Default - Generic Retpolines50K100K150K200K250KSE +/- 247.73, N = 3SE +/- 627.93, N = 3210605.86215287.791. (CC) gcc options: -ldl -lpthread -lcrypt -lz -O3 -march=native

Facebook RocksDB

Test: Sequential Fill

OpenBenchmarking.orgOp/s, More Is BetterFacebook RocksDB 6.22.1Test: Sequential FillAMD Retpolines Former DefaultNew Default - Generic Retpolines140K280K420K560K700KSE +/- 3478.74, N = 3SE +/- 1947.52, N = 36707606575231. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Facebook RocksDB

Test: Random Fill

OpenBenchmarking.orgOp/s, More Is BetterFacebook RocksDB 6.22.1Test: Random FillAMD Retpolines Former DefaultNew Default - Generic Retpolines130K260K390K520K650KSE +/- 3544.93, N = 3SE +/- 5871.64, N = 36172506055721. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Renaissance

Test: In-Memory Database Shootout

OpenBenchmarking.orgms, Fewer Is BetterRenaissance 0.12Test: In-Memory Database ShootoutAMD Retpolines Former DefaultNew Default - Generic Retpolines10002000300040005000SE +/- 64.75, N = 3SE +/- 22.81, N = 34878.74788.2MIN: 4512.55 / MAX: 5542.48MIN: 4452.21 / MAX: 5160.31

OSBench

Test: Create Threads

OpenBenchmarking.orgus Per Event, Fewer Is BetterOSBenchTest: Create ThreadsAMD Retpolines Former DefaultNew Default - Generic Retpolines48121620SE +/- 0.03, N = 3SE +/- 0.09, N = 317.7017.381. (CC) gcc options: -lm

Stress-NG

Test: Socket Activity

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.13.02Test: Socket ActivityAMD Retpolines Former DefaultNew Default - Generic Retpolines2K4K6K8K10KSE +/- 13.84, N = 3SE +/- 12.31, N = 310567.0910751.131. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic

LevelDB

Benchmark: Sequential Fill

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Sequential FillAMD Retpolines Former DefaultNew Default - Generic Retpolines20406080100SE +/- 0.61, N = 3SE +/- 0.43, N = 3106.95105.121. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

DaCapo Benchmark

Java Test: Jython

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: JythonAMD Retpolines Former DefaultNew Default - Generic Retpolines8001600240032004000SE +/- 24.35, N = 20SE +/- 28.58, N = 436973761

Hackbench

Count: 32 - Type: Process

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 32 - Type: ProcessAMD Retpolines Former DefaultNew Default - Generic Retpolines306090120150SE +/- 0.59, N = 3SE +/- 0.30, N = 3117.98119.831. (CC) gcc options: -lpthread

LZ4 Compression

Compression Level: 9 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Decompression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines3K6K9K12K15KSE +/- 61.43, N = 3SE +/- 17.39, N = 314807.414586.21. (CC) gcc options: -O3

OSBench

Test: Launch Programs

OpenBenchmarking.orgus Per Event, Fewer Is BetterOSBenchTest: Launch ProgramsAMD Retpolines Former DefaultNew Default - Generic Retpolines1122334455SE +/- 0.53, N = 5SE +/- 0.10, N = 348.6349.361. (CC) gcc options: -lm

Flexible IO Tester

Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory

OpenBenchmarking.orgIOPS, More Is BetterFlexible IO Tester 3.29Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test DirectoryAMD Retpolines Former DefaultNew Default - Generic Retpolines50K100K150K200K250KSE +/- 577.35, N = 3SE +/- 333.33, N = 32430002396671. (CC) gcc options: -rdynamic -ltcmalloc -ll -lnuma -lrt -lz -lpthread -lm -ldl -laio -lcurl -lssl -lcrypto -std=gnu99 -ffast-math -include -O3 -fcommon -U_FORTIFY_SOURCE -march=native

Facebook RocksDB

Test: Update Random

OpenBenchmarking.orgOp/s, More Is BetterFacebook RocksDB 6.22.1Test: Update RandomAMD Retpolines Former DefaultNew Default - Generic Retpolines90K180K270K360K450KSE +/- 1232.53, N = 3SE +/- 2036.18, N = 33985263931011. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

LevelDB

Benchmark: Overwrite

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: OverwriteAMD Retpolines Former DefaultNew Default - Generic Retpolines20406080100SE +/- 0.70, N = 14SE +/- 0.87, N = 15104.61103.371. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

LevelDB

Benchmark: Overwrite

OpenBenchmarking.orgMB/s, More Is BetterLevelDB 1.22Benchmark: OverwriteAMD Retpolines Former DefaultNew Default - Generic Retpolines48121620SE +/- 0.11, N = 14SE +/- 0.15, N = 1516.917.11. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

Stress-NG

Test: Semaphores

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.13.02Test: SemaphoresAMD Retpolines Former DefaultNew Default - Generic Retpolines300K600K900K1200K1500KSE +/- 10886.78, N = 3SE +/- 4542.73, N = 31342170.841326767.921. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic

Parallel BZIP2 Compression

FreeBSD-13.0-RELEASE-amd64-memstick.img Compression

OpenBenchmarking.orgSeconds, Fewer Is BetterParallel BZIP2 Compression 1.1.13FreeBSD-13.0-RELEASE-amd64-memstick.img CompressionAMD Retpolines Former DefaultNew Default - Generic Retpolines3691215SE +/- 0.15, N = 3SE +/- 0.11, N = 610.6610.781. (CXX) g++ options: -O2 -pthread -lbz2 -lpthread

Apache HTTP Server

Concurrent Requests: 100

OpenBenchmarking.orgRequests Per Second, More Is BetterApache HTTP Server 2.4.48Concurrent Requests: 100AMD Retpolines Former DefaultNew Default - Generic Retpolines20K40K60K80K100KSE +/- 311.80, N = 3SE +/- 56.43, N = 383397.5782500.751. (CC) gcc options: -shared -fPIC -O2 -pthread

7-Zip Compression

Test: Compression Rating

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 21.06Test: Compression RatingAMD Retpolines Former DefaultNew Default - Generic Retpolines20K40K60K80K100KSE +/- 870.19, N = 5SE +/- 438.70, N = 387210863571. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

Timed Eigen Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Eigen Compilation 3.3.9Time To CompileAMD Retpolines Former DefaultNew Default - Generic Retpolines1428425670SE +/- 0.02, N = 3SE +/- 0.02, N = 361.1061.69

DaCapo Benchmark

Java Test: H2

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: H2AMD Retpolines Former DefaultNew Default - Generic Retpolines9001800270036004500SE +/- 25.85, N = 16SE +/- 26.89, N = 442284268

LevelDB

Benchmark: Random Delete

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Random DeleteAMD Retpolines Former DefaultNew Default - Generic Retpolines20406080100SE +/- 1.08, N = 4SE +/- 0.42, N = 398.9599.881. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

DaCapo Benchmark

Java Test: Tradesoap

OpenBenchmarking.orgmsec, Fewer Is BetterDaCapo Benchmark 9.12-MR1Java Test: TradesoapAMD Retpolines Former DefaultNew Default - Generic Retpolines8001600240032004000SE +/- 33.50, N = 4SE +/- 40.14, N = 436283660

Memtier_benchmark

Protocol: Redis

OpenBenchmarking.orgOps/sec, More Is BetterMemtier_benchmark 1.3Protocol: RedisAMD Retpolines Former DefaultNew Default - Generic Retpolines300K600K900K1200K1500KSE +/- 1363.28, N = 3SE +/- 3316.11, N = 31313054.111324327.321. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre

Apache HTTP Server

Concurrent Requests: 500

OpenBenchmarking.orgRequests Per Second, More Is BetterApache HTTP Server 2.4.48Concurrent Requests: 500AMD Retpolines Former DefaultNew Default - Generic Retpolines20K40K60K80K100KSE +/- 149.43, N = 3SE +/- 264.49, N = 380970.1580298.151. (CC) gcc options: -shared -fPIC -O2 -pthread

LevelDB

Benchmark: Fill Sync

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Fill SyncAMD Retpolines Former DefaultNew Default - Generic Retpolines400800120016002000SE +/- 3.85, N = 3SE +/- 8.19, N = 31806.761820.491. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

LevelDB

Benchmark: Seek Random

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Seek RandomAMD Retpolines Former DefaultNew Default - Generic Retpolines48121620SE +/- 0.09, N = 3SE +/- 0.05, N = 317.0117.141. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

Hackbench

Count: 8 - Type: Thread

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 8 - Type: ThreadAMD Retpolines Former DefaultNew Default - Generic Retpolines714212835SE +/- 0.14, N = 3SE +/- 0.16, N = 331.0531.281. (CC) gcc options: -lpthread

Hackbench

Count: 16 - Type: Process

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 16 - Type: ProcessAMD Retpolines Former DefaultNew Default - Generic Retpolines1326395265SE +/- 0.23, N = 3SE +/- 0.16, N = 358.3658.781. (CC) gcc options: -lpthread

Google Draco

Model: Church Facade

OpenBenchmarking.orgms, Fewer Is BetterGoogle Draco 1.5.0Model: Church FacadeAMD Retpolines Former DefaultNew Default - Generic Retpolines14002800420056007000SE +/- 5.21, N = 3SE +/- 17.74, N = 3664966971. (CXX) g++ options: -O3

Zstd Compression

Compression Level: 19 - Decompression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 19 - Decompression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines8001600240032004000SE +/- 14.48, N = 15SE +/- 18.20, N = 153668.53694.51. (CC) gcc options: -O3 -pthread -lz -llzma

Stress-NG

Test: Context Switching

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.13.02Test: Context SwitchingAMD Retpolines Former DefaultNew Default - Generic Retpolines900K1800K2700K3600K4500KSE +/- 26201.07, N = 3SE +/- 18653.91, N = 34326244.504296558.491. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic

Sockperf

Test: Latency Under Load

OpenBenchmarking.orgusec, Fewer Is BetterSockperf 3.7Test: Latency Under LoadAMD Retpolines Former DefaultNew Default - Generic Retpolines0.82491.64982.47473.29964.1245SE +/- 0.033, N = 25SE +/- 0.035, N = 53.6413.6661. (CXX) g++ options: --param -O3 -rdynamic -ldl -lpthread

Facebook RocksDB

Test: Read While Writing

OpenBenchmarking.orgOp/s, More Is BetterFacebook RocksDB 6.22.1Test: Read While WritingAMD Retpolines Former DefaultNew Default - Generic Retpolines400K800K1200K1600K2000KSE +/- 20451.18, N = 4SE +/- 24226.58, N = 3174652417348781. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 50 - Mode: Read Only - Average Latency

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL pgbench 14.0Scaling Factor: 100 - Clients: 50 - Mode: Read Only - Average LatencyAMD Retpolines Former DefaultNew Default - Generic Retpolines0.03830.07660.11490.15320.1915SE +/- 0.002, N = 12SE +/- 0.001, N = 120.1690.1701. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

LevelDB

Benchmark: Random Fill

OpenBenchmarking.orgMB/s, More Is BetterLevelDB 1.22Benchmark: Random FillAMD Retpolines Former DefaultNew Default - Generic Retpolines48121620SE +/- 0.22, N = 3SE +/- 0.15, N = 1517.117.21. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

Zstd Compression

Compression Level: 19 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 19 - Compression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines1224364860SE +/- 0.45, N = 15SE +/- 0.47, N = 1551.951.61. (CC) gcc options: -O3 -pthread -lz -llzma

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 100 - Mode: Read Write

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL pgbench 14.0Scaling Factor: 100 - Clients: 100 - Mode: Read WriteAMD Retpolines Former DefaultNew Default - Generic Retpolines2K4K6K8K10KSE +/- 108.62, N = 12SE +/- 102.24, N = 1210440103831. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

Apache Cassandra

Test: Writes

OpenBenchmarking.orgOp/s, More Is BetterApache Cassandra 4.0Test: WritesAMD Retpolines Former DefaultNew Default - Generic Retpolines20K40K60K80K100KSE +/- 518.75, N = 3SE +/- 31.48, N = 38595686420

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 100 - Mode: Read Write - Average Latency

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL pgbench 14.0Scaling Factor: 100 - Clients: 100 - Mode: Read Write - Average LatencyAMD Retpolines Former DefaultNew Default - Generic Retpolines3691215SE +/- 0.099, N = 12SE +/- 0.094, N = 129.5909.6411. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

OSBench

Test: Create Processes

OpenBenchmarking.orgus Per Event, Fewer Is BetterOSBenchTest: Create ProcessesAMD Retpolines Former DefaultNew Default - Generic Retpolines1020304050SE +/- 0.54, N = 3SE +/- 0.23, N = 343.3643.141. (CC) gcc options: -lm

Renaissance

Test: Finagle HTTP Requests

OpenBenchmarking.orgms, Fewer Is BetterRenaissance 0.12Test: Finagle HTTP RequestsAMD Retpolines Former DefaultNew Default - Generic Retpolines5001000150020002500SE +/- 15.84, N = 3SE +/- 24.88, N = 32366.92378.7MIN: 2178.29 / MAX: 2434.21MIN: 2177.37 / MAX: 2449.07

LevelDB

Benchmark: Random Fill

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Random FillAMD Retpolines Former DefaultNew Default - Generic Retpolines20406080100SE +/- 1.34, N = 3SE +/- 0.90, N = 15103.68103.221. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

Node.js V8 Web Tooling Benchmark

OpenBenchmarking.orgruns/s, More Is BetterNode.js V8 Web Tooling BenchmarkAMD Retpolines Former DefaultNew Default - Generic Retpolines3691215SE +/- 0.03, N = 3SE +/- 0.05, N = 311.3111.26

Facebook RocksDB

Test: Random Read

OpenBenchmarking.orgOp/s, More Is BetterFacebook RocksDB 6.22.1Test: Random ReadAMD Retpolines Former DefaultNew Default - Generic Retpolines13M26M39M52M65MSE +/- 78029.55, N = 3SE +/- 195189.61, N = 361586232613317211. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

Zstd Compression

Compression Level: 8 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterZstd Compression 1.5.0Compression Level: 8 - Compression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines30060090012001500SE +/- 10.10, N = 3SE +/- 3.86, N = 31338.61343.61. (CC) gcc options: -O3 -pthread -lz -llzma

Google Draco

Model: Lion

OpenBenchmarking.orgms, Fewer Is BetterGoogle Draco 1.5.0Model: LionAMD Retpolines Former DefaultNew Default - Generic Retpolines11002200330044005500SE +/- 0.88, N = 3SE +/- 7.37, N = 3507050881. (CXX) g++ options: -O3

LevelDB

Benchmark: Hot Read

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Hot ReadAMD Retpolines Former DefaultNew Default - Generic Retpolines3691215SE +/- 0.05, N = 3SE +/- 0.11, N = 610.5010.541. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

Facebook RocksDB

Test: Read Random Write Random

OpenBenchmarking.orgOp/s, More Is BetterFacebook RocksDB 6.22.1Test: Read Random Write RandomAMD Retpolines Former DefaultNew Default - Generic Retpolines300K600K900K1200K1500KSE +/- 6491.63, N = 3SE +/- 2292.69, N = 3138658613910001. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread

LevelDB

Benchmark: Random Read

OpenBenchmarking.orgMicroseconds Per Op, Fewer Is BetterLevelDB 1.22Benchmark: Random ReadAMD Retpolines Former DefaultNew Default - Generic Retpolines3691215SE +/- 0.13, N = 3SE +/- 0.07, N = 310.5510.581. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 100 - Mode: Read Only - Average Latency

OpenBenchmarking.orgms, Fewer Is BetterPostgreSQL pgbench 14.0Scaling Factor: 100 - Clients: 100 - Mode: Read Only - Average LatencyAMD Retpolines Former DefaultNew Default - Generic Retpolines0.07450.1490.22350.2980.3725SE +/- 0.000, N = 3SE +/- 0.004, N = 40.3300.3311. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

Timed MPlayer Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed MPlayer Compilation 1.5Time To CompileAMD Retpolines Former DefaultNew Default - Generic Retpolines714212835SE +/- 0.04, N = 3SE +/- 0.02, N = 330.8030.71

Hackbench

Count: 16 - Type: Thread

OpenBenchmarking.orgSeconds, Fewer Is BetterHackbenchCount: 16 - Type: ThreadAMD Retpolines Former DefaultNew Default - Generic Retpolines1428425670SE +/- 0.11, N = 3SE +/- 0.19, N = 362.0362.201. (CC) gcc options: -lpthread

InfluxDB

Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000

OpenBenchmarking.orgval/sec, More Is BetterInfluxDB 1.8.2Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000AMD Retpolines Former DefaultNew Default - Generic Retpolines300K600K900K1200K1500KSE +/- 2326.89, N = 3SE +/- 1561.10, N = 31323460.21326859.1

LZ4 Compression

Compression Level: 9 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 9 - Compression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines1326395265SE +/- 0.04, N = 3SE +/- 0.63, N = 359.6959.571. (CC) gcc options: -O3

LZ4 Compression

Compression Level: 3 - Compression Speed

OpenBenchmarking.orgMB/s, More Is BetterLZ4 Compression 1.9.3Compression Level: 3 - Compression SpeedAMD Retpolines Former DefaultNew Default - Generic Retpolines1428425670SE +/- 0.30, N = 3SE +/- 0.47, N = 1561.3661.241. (CC) gcc options: -O3

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 50 - Mode: Read Only

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL pgbench 14.0Scaling Factor: 100 - Clients: 50 - Mode: Read OnlyAMD Retpolines Former DefaultNew Default - Generic Retpolines60K120K180K240K300KSE +/- 3223.57, N = 12SE +/- 2460.01, N = 122957302951791. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

Stress-NG

Test: System V Message Passing

OpenBenchmarking.orgBogo Ops/s, More Is BetterStress-NG 0.13.02Test: System V Message PassingAMD Retpolines Former DefaultNew Default - Generic Retpolines1000K2000K3000K4000K5000KSE +/- 10539.69, N = 3SE +/- 16005.42, N = 34530010.924521841.841. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic

Redis

Test: SET

OpenBenchmarking.orgRequests Per Second, More Is BetterRedis 6.0.9Test: SETAMD Retpolines Former DefaultNew Default - Generic Retpolines400K800K1200K1600K2000KSE +/- 18125.41, N = 7SE +/- 2446.32, N = 31995692.891998825.921. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3

Timed GDB GNU Debugger Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed GDB GNU Debugger Compilation 10.2Time To CompileAMD Retpolines Former DefaultNew Default - Generic Retpolines1224364860SE +/- 0.16, N = 3SE +/- 0.13, N = 355.0354.96

7-Zip Compression

Test: Decompression Rating

OpenBenchmarking.orgMIPS, More Is Better7-Zip Compression 21.06Test: Decompression RatingAMD Retpolines Former DefaultNew Default - Generic Retpolines15K30K45K60K75KSE +/- 59.62, N = 5SE +/- 8.76, N = 369308693931. (CXX) g++ options: -lpthread -ldl -O2 -fPIC

InfluxDB

Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000

OpenBenchmarking.orgval/sec, More Is BetterInfluxDB 1.8.2Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000AMD Retpolines Former DefaultNew Default - Generic Retpolines400K800K1200K1600K2000KSE +/- 1145.84, N = 3SE +/- 3111.87, N = 31639910.01638195.7

OpenSSL

Algorithm: RSA4096

OpenBenchmarking.orgsign/s, More Is BetterOpenSSL 3.0Algorithm: RSA4096AMD Retpolines Former DefaultNew Default - Generic Retpolines5001000150020002500SE +/- 0.15, N = 3SE +/- 2.17, N = 32374.12372.31. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl

Timed Godot Game Engine Compilation

Time To Compile

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Godot Game Engine Compilation 3.2.3Time To CompileAMD Retpolines Former DefaultNew Default - Generic Retpolines20406080100SE +/- 0.09, N = 3SE +/- 0.22, N = 3111.58111.51

Apache HTTP Server

Concurrent Requests: 200

OpenBenchmarking.orgRequests Per Second, More Is BetterApache HTTP Server 2.4.48Concurrent Requests: 200AMD Retpolines Former DefaultNew Default - Generic Retpolines20K40K60K80K100KSE +/- 123.84, N = 3SE +/- 557.73, N = 386408.4886362.181. (CC) gcc options: -shared -fPIC -O2 -pthread

PostgreSQL pgbench

Scaling Factor: 100 - Clients: 100 - Mode: Read Only

OpenBenchmarking.orgTPS, More Is BetterPostgreSQL pgbench 14.0Scaling Factor: 100 - Clients: 100 - Mode: Read OnlyAMD Retpolines Former DefaultNew Default - Generic Retpolines60K120K180K240K300KSE +/- 303.18, N = 3SE +/- 3587.18, N = 43023993025231. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm

Timed Linux Kernel Compilation

Build: defconfig

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed Linux Kernel Compilation 5.16Build: defconfigAMD Retpolines Former DefaultNew Default - Generic Retpolines20406080100SE +/- 0.03, N = 3SE +/- 0.06, N = 378.5978.62

OpenSSL

Algorithm: SHA256

OpenBenchmarking.orgbyte/s, More Is BetterOpenSSL 3.0Algorithm: SHA256AMD Retpolines Former DefaultNew Default - Generic Retpolines3000M6000M9000M12000M15000MSE +/- 9115647.24, N = 3SE +/- 6459514.72, N = 313234833737132387228601. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl

OpenSSL

Algorithm: RSA4096

OpenBenchmarking.orgverify/s, More Is BetterOpenSSL 3.0Algorithm: RSA4096AMD Retpolines Former DefaultNew Default - Generic Retpolines30K60K90K120K150KSE +/- 8.58, N = 3SE +/- 7.94, N = 3155031.8155003.11. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl


Phoronix Test Suite v10.8.4