AMD EPYC 72F3 Retpoline benchmarks by Michael Larabel.
New Default - Generic Retpolines Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vDisk Notes: NONE / errors=remount-ro,relatime,rw / Block Size: 4096Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001119Java Notes: OpenJDK Runtime Environment (build 11.0.13+8-Ubuntu-0ubuntu1.21.04)Python Notes: Python 3.9.5Security Notes: 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 Processor: AMD EPYC 72F3 8-Core @ 3.70GHz (8 Cores / 16 Threads), Motherboard: Supermicro H12SSL-i v1.01 (2.0 BIOS), Chipset: AMD Starship/Matisse, Memory: 128GB, Disk: 3841GB Micron_9300_MTFDHAL3T8TDP + 1000GB Corsair Force MP600, Graphics: ASPEED, Monitor: VE228, Network: 2 x Broadcom NetXtreme BCM5720 2-port PCIe
OS: Ubuntu 21.04, Kernel: 5.17.0-051700rc7daily20220310-generic (x86_64), Desktop: GNOME Shell 3.38.4, Display Server: X Server, Compiler: GCC 10.3.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vDisk Notes: NONE / errors=remount-ro,relatime,rw / Block Size: 4096Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001119Java Notes: OpenJDK Runtime Environment (build 11.0.13+8-Ubuntu-0ubuntu1.21.04)Python Notes: Python 3.9.5Security Notes: 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 5.17 OpenBenchmarking.org Phoronix Test Suite AMD EPYC 72F3 8-Core @ 3.70GHz (8 Cores / 16 Threads) Supermicro H12SSL-i v1.01 (2.0 BIOS) AMD Starship/Matisse 128GB 3841GB Micron_9300_MTFDHAL3T8TDP + 1000GB Corsair Force MP600 ASPEED VE228 2 x Broadcom NetXtreme BCM5720 2-port PCIe Ubuntu 21.04 5.17.0-051700rc7daily20220310-generic (x86_64) GNOME Shell 3.38.4 X Server GCC 10.3.0 ext4 1920x1080 Processor Motherboard Chipset Memory Disk Graphics Monitor Network OS Kernel Desktop Display Server Compiler File-System Screen Resolution EPYC 72F3 Retpolines Change In Linux 5.17 Benchmarks System Logs - Transparent Huge Pages: madvise - --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 - NONE / errors=remount-ro,relatime,rw / Block Size: 4096 - Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001119 - OpenJDK Runtime Environment (build 11.0.13+8-Ubuntu-0ubuntu1.21.04) - Python 3.9.5 - 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
New Default - Generic Retpolines vs. AMD Retpolines Former Default Comparison Phoronix Test Suite Baseline +2.2% +2.2% +4.4% +4.4% +6.6% +6.6% +8.8% +8.8% 8.9% 7.2% 4.9% 4.2% 3.9% 3.4% 3.3% 3% 2.5% 2.5% 2.3% 2% Throughput Create Files D.T.P 8 - Process 3.8% SENDFILE 3.7% 200 3.7% 100 3.6% Timed Time - Size 1,000 GET 2.8% 3 - D.S Rand Write - IO_uring - No - Yes - 4KB Seq Fill 2.4% 8 - D.S 2.3% Tradebeans 20 2.2% Seq Fill Sockperf OSBench PostMark KeyDB Hackbench ebizzy Stress-NG nginx nginx WireGuard + Linux Networking Stack Stress Test Node.js Express HTTP Load Test SQLite Speedtest Redis LZ4 Compression Flexible IO Tester LevelDB Zstd Compression DaCapo Benchmark nginx Facebook RocksDB New Default - Generic Retpolines AMD Retpolines Former Default
Stress-NG Stress-NG is a Linux stress tool developed by Colin King of Canonical. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.13.02 Test: SENDFILE AMD Retpolines Former Default New Default - Generic Retpolines 40K 80K 120K 160K 200K SE +/- 443.51, N = 3 SE +/- 510.20, N = 3 193652.98 200771.29 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.13.02 Test: Semaphores AMD Retpolines Former Default New Default - Generic Retpolines 300K 600K 900K 1200K 1500K SE +/- 10886.78, N = 3 SE +/- 4542.73, N = 3 1342170.84 1326767.92 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.13.02 Test: Socket Activity AMD Retpolines Former Default New Default - Generic Retpolines 2K 4K 6K 8K 10K SE +/- 13.84, N = 3 SE +/- 12.31, N = 3 10567.09 10751.13 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.13.02 Test: Context Switching AMD Retpolines Former Default New Default - Generic Retpolines 900K 1800K 2700K 3600K 4500K SE +/- 26201.07, N = 3 SE +/- 18653.91, N = 3 4326244.50 4296558.49 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic
OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.13.02 Test: System V Message Passing AMD Retpolines Former Default New Default - Generic Retpolines 1000K 2000K 3000K 4000K 5000K SE +/- 10539.69, N = 3 SE +/- 16005.42, N = 3 4530010.92 4521841.84 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lsctp -lz -ldl -pthread -lc -latomic
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org byte/s, More Is Better OpenSSL 3.0 Algorithm: SHA256 AMD Retpolines Former Default New Default - Generic Retpolines 3000M 6000M 9000M 12000M 15000M SE +/- 9115647.24, N = 3 SE +/- 6459514.72, N = 3 13234833737 13238722860 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
Flexible IO Tester FIO, the Flexible I/O Tester, is an advanced Linux disk benchmark supporting multiple I/O engines and a wealth of options. FIO was written by Jens Axboe for testing of the Linux I/O subsystem and schedulers. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester 3.29 Type: Random Read - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory AMD Retpolines Former Default New Default - Generic Retpolines 50K 100K 150K 200K 250K SE +/- 577.35, N = 3 SE +/- 333.33, N = 3 243000 239667 1. (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
OpenBenchmarking.org IOPS, More Is Better Flexible IO Tester 3.29 Type: Random Write - Engine: IO_uring - Buffered: No - Direct: Yes - Block Size: 4KB - Disk Target: Default Test Directory AMD Retpolines Former Default New Default - Generic Retpolines 50K 100K 150K 200K 250K SE +/- 2958.04, N = 4 SE +/- 3027.60, N = 15 229500 223933 1. (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 LevelDB is a key-value storage library developed by Google that supports making use of Snappy for data compression and has other modern features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Overwrite AMD Retpolines Former Default New Default - Generic Retpolines 4 8 12 16 20 SE +/- 0.11, N = 14 SE +/- 0.15, N = 15 16.9 17.1 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Random Fill AMD Retpolines Former Default New Default - Generic Retpolines 4 8 12 16 20 SE +/- 0.22, N = 3 SE +/- 0.15, N = 15 17.1 17.2 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Sequential Fill AMD Retpolines Former Default New Default - Generic Retpolines 4 8 12 16 20 SE +/- 0.09, N = 3 SE +/- 0.07, N = 3 16.5 16.9 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 3 - Decompression Speed AMD Retpolines Former Default New Default - Generic Retpolines 3K 6K 9K 12K 15K SE +/- 22.12, N = 3 SE +/- 24.84, N = 15 14817.3 14455.1 1. (CC) gcc options: -O3
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 9 - Compression Speed AMD Retpolines Former Default New Default - Generic Retpolines 13 26 39 52 65 SE +/- 0.04, N = 3 SE +/- 0.63, N = 3 59.69 59.57 1. (CC) gcc options: -O3
OpenBenchmarking.org MB/s, More Is Better LZ4 Compression 1.9.3 Compression Level: 9 - Decompression Speed AMD Retpolines Former Default New Default - Generic Retpolines 3K 6K 9K 12K 15K SE +/- 61.43, N = 3 SE +/- 17.39, N = 3 14807.4 14586.2 1. (CC) gcc options: -O3
Zstd Compression This test measures the time needed to compress/decompress a sample file (a FreeBSD disk image - FreeBSD-12.2-RELEASE-amd64-memstick.img) using Zstd compression with options for different compression levels / settings. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.0 Compression Level: 8 - Compression Speed AMD Retpolines Former Default New Default - Generic Retpolines 300 600 900 1200 1500 SE +/- 10.10, N = 3 SE +/- 3.86, N = 3 1338.6 1343.6 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.0 Compression Level: 8 - Decompression Speed AMD Retpolines Former Default New Default - Generic Retpolines 900 1800 2700 3600 4500 SE +/- 4.29, N = 3 SE +/- 54.59, N = 3 4160.1 4255.1 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.0 Compression Level: 19 - Compression Speed AMD Retpolines Former Default New Default - Generic Retpolines 12 24 36 48 60 SE +/- 0.45, N = 15 SE +/- 0.47, N = 15 51.9 51.6 1. (CC) gcc options: -O3 -pthread -lz -llzma
OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.5.0 Compression Level: 19 - Decompression Speed AMD Retpolines Former Default New Default - Generic Retpolines 800 1600 2400 3200 4000 SE +/- 14.48, N = 15 SE +/- 18.20, N = 15 3668.5 3694.5 1. (CC) gcc options: -O3 -pthread -lz -llzma
Sockperf This is a network socket API performance benchmark developed by Mellanox. This test profile runs both the client and server on the local host for evaluating individual system performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Messages Per Second, More Is Better Sockperf 3.7 Test: Throughput AMD Retpolines Former Default New Default - Generic Retpolines 130K 260K 390K 520K 650K SE +/- 5255.38, N = 25 SE +/- 4906.55, N = 25 587523 539724 1. (CXX) g++ options: --param -O3 -rdynamic -ldl -lpthread
OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 21.06 Test: Decompression Rating AMD Retpolines Former Default New Default - Generic Retpolines 15K 30K 45K 60K 75K SE +/- 59.62, N = 5 SE +/- 8.76, N = 3 69308 69393 1. (CXX) g++ options: -lpthread -ldl -O2 -fPIC
Facebook RocksDB This is a benchmark of Facebook's RocksDB as an embeddable persistent key-value store for fast storage based on Google's LevelDB. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.22.1 Test: Random Fill AMD Retpolines Former Default New Default - Generic Retpolines 130K 260K 390K 520K 650K SE +/- 3544.93, N = 3 SE +/- 5871.64, N = 3 617250 605572 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.22.1 Test: Random Read AMD Retpolines Former Default New Default - Generic Retpolines 13M 26M 39M 52M 65M SE +/- 78029.55, N = 3 SE +/- 195189.61, N = 3 61586232 61331721 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.22.1 Test: Update Random AMD Retpolines Former Default New Default - Generic Retpolines 90K 180K 270K 360K 450K SE +/- 1232.53, N = 3 SE +/- 2036.18, N = 3 398526 393101 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.22.1 Test: Sequential Fill AMD Retpolines Former Default New Default - Generic Retpolines 140K 280K 420K 560K 700K SE +/- 3478.74, N = 3 SE +/- 1947.52, N = 3 670760 657523 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.22.1 Test: Read While Writing AMD Retpolines Former Default New Default - Generic Retpolines 400K 800K 1200K 1600K 2000K SE +/- 20451.18, N = 4 SE +/- 24226.58, N = 3 1746524 1734878 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.22.1 Test: Read Random Write Random AMD Retpolines Former Default New Default - Generic Retpolines 300K 600K 900K 1200K 1500K SE +/- 6491.63, N = 3 SE +/- 2292.69, N = 3 1386586 1391000 1. (CXX) g++ options: -O3 -march=native -pthread -fno-builtin-memcmp -fno-rtti -lpthread
KeyDB A benchmark of KeyDB as a multi-threaded fork of the Redis server. The KeyDB benchmark is conducted using memtier-benchmark. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better KeyDB 6.2.0 AMD Retpolines Former Default New Default - Generic Retpolines 30K 60K 90K 120K 150K SE +/- 1055.00, N = 15 SE +/- 1054.95, N = 10 142573.34 136793.59 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
Memtier_benchmark Memtier_benchmark is a NoSQL Redis/Memcache traffic generation plus benchmarking tool. This current test profile currently just stresses the Redis protocol and basic options exposed wotj a 1:1 Set/Get ratio, 30 pipeline, 100 clients per thread, and thread count equal to the number of CPU cores/threads present. Patches to extend the test are welcome as always. Currently this test profile uses Memtier_benchmark 1.3 and Redis 6. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Ops/sec, More Is Better Memtier_benchmark 1.3 Protocol: Redis AMD Retpolines Former Default New Default - Generic Retpolines 300K 600K 900K 1200K 1500K SE +/- 1363.28, N = 3 SE +/- 3316.11, N = 3 1313054.11 1324327.32 1. (CXX) g++ options: -O2 -levent_openssl -levent -lcrypto -lssl -lpthread -lz -lpcre
Redis Redis is an open-source in-memory data structure store, used as a database, cache, and message broker. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: GET AMD Retpolines Former Default New Default - Generic Retpolines 500K 1000K 1500K 2000K 2500K SE +/- 22769.87, N = 6 SE +/- 15897.48, N = 3 2438721.71 2506720.25 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
OpenBenchmarking.org Requests Per Second, More Is Better Redis 6.0.9 Test: SET AMD Retpolines Former Default New Default - Generic Retpolines 400K 800K 1200K 1600K 2000K SE +/- 18125.41, N = 7 SE +/- 2446.32, N = 3 1995692.89 1998825.92 1. (CXX) g++ options: -MM -MT -g3 -fvisibility=hidden -O3
nginx This is a benchmark of the lightweight Nginx HTTP(S) web-server. This Nginx web server benchmark test profile makes use of the Golang "Bombardier" program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better nginx 1.21.1 Concurrent Requests: 20 AMD Retpolines Former Default New Default - Generic Retpolines 50K 100K 150K 200K 250K SE +/- 247.73, N = 3 SE +/- 627.93, N = 3 210605.86 215287.79 1. (CC) gcc options: -ldl -lpthread -lcrypt -lz -O3 -march=native
OpenBenchmarking.org Requests Per Second, More Is Better nginx 1.21.1 Concurrent Requests: 100 AMD Retpolines Former Default New Default - Generic Retpolines 50K 100K 150K 200K 250K SE +/- 369.20, N = 3 SE +/- 1627.81, N = 3 207615.65 215190.56 1. (CC) gcc options: -ldl -lpthread -lcrypt -lz -O3 -march=native
OpenBenchmarking.org Requests Per Second, More Is Better nginx 1.21.1 Concurrent Requests: 200 AMD Retpolines Former Default New Default - Generic Retpolines 50K 100K 150K 200K 250K SE +/- 466.89, N = 3 SE +/- 275.98, N = 3 206147.39 213684.40 1. (CC) gcc options: -ldl -lpthread -lcrypt -lz -O3 -march=native
Apache HTTP Server This is a test of the Apache HTTPD web server. This Apache HTTPD web server benchmark test profile makes use of the Golang "Bombardier" program for facilitating the HTTP requests over a fixed period time with a configurable number of concurrent clients. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Requests Per Second, More Is Better Apache HTTP Server 2.4.48 Concurrent Requests: 100 AMD Retpolines Former Default New Default - Generic Retpolines 20K 40K 60K 80K 100K SE +/- 311.80, N = 3 SE +/- 56.43, N = 3 83397.57 82500.75 1. (CC) gcc options: -shared -fPIC -O2 -pthread
OpenBenchmarking.org Requests Per Second, More Is Better Apache HTTP Server 2.4.48 Concurrent Requests: 200 AMD Retpolines Former Default New Default - Generic Retpolines 20K 40K 60K 80K 100K SE +/- 123.84, N = 3 SE +/- 557.73, N = 3 86408.48 86362.18 1. (CC) gcc options: -shared -fPIC -O2 -pthread
OpenBenchmarking.org Requests Per Second, More Is Better Apache HTTP Server 2.4.48 Concurrent Requests: 500 AMD Retpolines Former Default New Default - Generic Retpolines 20K 40K 60K 80K 100K SE +/- 149.43, N = 3 SE +/- 264.49, N = 3 80970.15 80298.15 1. (CC) gcc options: -shared -fPIC -O2 -pthread
Node.js V8 Web Tooling Benchmark Running the V8 project's Web-Tooling-Benchmark under Node.js. The Web-Tooling-Benchmark stresses JavaScript-related workloads common to web developers like Babel and TypeScript and Babylon. This test profile can test the system's JavaScript performance with Node.js. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org runs/s, More Is Better Node.js V8 Web Tooling Benchmark AMD Retpolines Former Default New Default - Generic Retpolines 3 6 9 12 15 SE +/- 0.03, N = 3 SE +/- 0.05, N = 3 11.31 11.26
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org sign/s, More Is Better OpenSSL 3.0 Algorithm: RSA4096 AMD Retpolines Former Default New Default - Generic Retpolines 500 1000 1500 2000 2500 SE +/- 0.15, N = 3 SE +/- 2.17, N = 3 2374.1 2372.3 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
PostMark This is a test of NetApp's PostMark benchmark designed to simulate small-file testing similar to the tasks endured by web and mail servers. This test profile will set PostMark to perform 25,000 transactions with 500 files simultaneously with the file sizes ranging between 5 and 512 kilobytes. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org TPS, More Is Better PostMark 1.51 Disk Transaction Performance AMD Retpolines Former Default New Default - Generic Retpolines 1600 3200 4800 6400 8000 SE +/- 70.00, N = 3 7282 6944 1. (CC) gcc options: -O3
PostgreSQL pgbench This is a benchmark of PostgreSQL using pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 14.0 Scaling Factor: 100 - Clients: 50 - Mode: Read Only AMD Retpolines Former Default New Default - Generic Retpolines 60K 120K 180K 240K 300K SE +/- 3223.57, N = 12 SE +/- 2460.01, N = 12 295730 295179 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 14.0 Scaling Factor: 100 - Clients: 100 - Mode: Read Only AMD Retpolines Former Default New Default - Generic Retpolines 60K 120K 180K 240K 300K SE +/- 303.18, N = 3 SE +/- 3587.18, N = 4 302399 302523 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 14.0 Scaling Factor: 100 - Clients: 100 - Mode: Read Write AMD Retpolines Former Default New Default - Generic Retpolines 2K 4K 6K 8K 10K SE +/- 108.62, N = 12 SE +/- 102.24, N = 12 10440 10383 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
InfluxDB This is a benchmark of the InfluxDB open-source time-series database optimized for fast, high-availability storage for IoT and other use-cases. The InfluxDB test profile makes use of InfluxDB Inch for facilitating the benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org val/sec, More Is Better InfluxDB 1.8.2 Concurrent Streams: 4 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000 AMD Retpolines Former Default New Default - Generic Retpolines 300K 600K 900K 1200K 1500K SE +/- 2326.89, N = 3 SE +/- 1561.10, N = 3 1323460.2 1326859.1
OpenBenchmarking.org val/sec, More Is Better InfluxDB 1.8.2 Concurrent Streams: 64 - Batch Size: 10000 - Tags: 2,5000,1 - Points Per Series: 10000 AMD Retpolines Former Default New Default - Generic Retpolines 400K 800K 1200K 1600K 2000K SE +/- 1145.84, N = 3 SE +/- 3111.87, N = 3 1639910.0 1638195.7
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test profile makes use of the built-in "openssl speed" benchmarking capabilities. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org verify/s, More Is Better OpenSSL 3.0 Algorithm: RSA4096 AMD Retpolines Former Default New Default - Generic Retpolines 30K 60K 90K 120K 150K SE +/- 8.58, N = 3 SE +/- 7.94, N = 3 155031.8 155003.1 1. (CC) gcc options: -pthread -m64 -O3 -lssl -lcrypto -ldl
LevelDB LevelDB is a key-value storage library developed by Google that supports making use of Snappy for data compression and has other modern features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Hot Read AMD Retpolines Former Default New Default - Generic Retpolines 3 6 9 12 15 SE +/- 0.05, N = 3 SE +/- 0.11, N = 6 10.50 10.54 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Fill Sync AMD Retpolines Former Default New Default - Generic Retpolines 400 800 1200 1600 2000 SE +/- 3.85, N = 3 SE +/- 8.19, N = 3 1806.76 1820.49 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Overwrite AMD Retpolines Former Default New Default - Generic Retpolines 20 40 60 80 100 SE +/- 0.70, N = 14 SE +/- 0.87, N = 15 104.61 103.37 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Fill AMD Retpolines Former Default New Default - Generic Retpolines 20 40 60 80 100 SE +/- 1.34, N = 3 SE +/- 0.90, N = 15 103.68 103.22 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Read AMD Retpolines Former Default New Default - Generic Retpolines 3 6 9 12 15 SE +/- 0.13, N = 3 SE +/- 0.07, N = 3 10.55 10.58 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Seek Random AMD Retpolines Former Default New Default - Generic Retpolines 4 8 12 16 20 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 17.01 17.14 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Delete AMD Retpolines Former Default New Default - Generic Retpolines 20 40 60 80 100 SE +/- 1.08, N = 4 SE +/- 0.42, N = 3 98.95 99.88 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Sequential Fill AMD Retpolines Former Default New Default - Generic Retpolines 20 40 60 80 100 SE +/- 0.61, N = 3 SE +/- 0.43, N = 3 106.95 105.12 1. (CXX) g++ options: -O3 -lsnappy -ltcmalloc -lpthread
Renaissance Renaissance is a suite of benchmarks designed to test the Java JVM from Apache Spark to a Twitter-like service to Scala and other features. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Renaissance 0.12 Test: Finagle HTTP Requests AMD Retpolines Former Default New Default - Generic Retpolines 500 1000 1500 2000 2500 SE +/- 15.84, N = 3 SE +/- 24.88, N = 3 2366.9 2378.7 MIN: 2178.29 / MAX: 2434.21 MIN: 2177.37 / MAX: 2449.07
OpenBenchmarking.org ms, Fewer Is Better Renaissance 0.12 Test: In-Memory Database Shootout AMD Retpolines Former Default New Default - Generic Retpolines 1000 2000 3000 4000 5000 SE +/- 64.75, N = 3 SE +/- 22.81, N = 3 4878.7 4788.2 MIN: 4512.55 / MAX: 5542.48 MIN: 4452.21 / MAX: 5160.31
PostgreSQL pgbench This is a benchmark of PostgreSQL using pgbench for facilitating the database benchmarks. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 14.0 Scaling Factor: 100 - Clients: 50 - Mode: Read Only - Average Latency AMD Retpolines Former Default New Default - Generic Retpolines 0.0383 0.0766 0.1149 0.1532 0.1915 SE +/- 0.002, N = 12 SE +/- 0.001, N = 12 0.169 0.170 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 14.0 Scaling Factor: 100 - Clients: 100 - Mode: Read Only - Average Latency AMD Retpolines Former Default New Default - Generic Retpolines 0.0745 0.149 0.2235 0.298 0.3725 SE +/- 0.000, N = 3 SE +/- 0.004, N = 4 0.330 0.331 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 14.0 Scaling Factor: 100 - Clients: 100 - Mode: Read Write - Average Latency AMD Retpolines Former Default New Default - Generic Retpolines 3 6 9 12 15 SE +/- 0.099, N = 12 SE +/- 0.094, N = 12 9.590 9.641 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
Google Draco Draco is a library developed by Google for compressing/decompressing 3D geometric meshes and point clouds. This test profile uses some Artec3D PLY models as the sample 3D model input formats for Draco compression/decompression. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org ms, Fewer Is Better Google Draco 1.5.0 Model: Lion AMD Retpolines Former Default New Default - Generic Retpolines 1100 2200 3300 4400 5500 SE +/- 0.88, N = 3 SE +/- 7.37, N = 3 5070 5088 1. (CXX) g++ options: -O3
OpenBenchmarking.org ms, Fewer Is Better Google Draco 1.5.0 Model: Church Facade AMD Retpolines Former Default New Default - Generic Retpolines 1400 2800 4200 5600 7000 SE +/- 5.21, N = 3 SE +/- 17.74, N = 3 6649 6697 1. (CXX) g++ options: -O3
WireGuard + Linux Networking Stack Stress Test This is a benchmark of the WireGuard secure VPN tunnel and Linux networking stack stress test. The test runs on the local host but does require root permissions to run. The way it works is it creates three namespaces. ns0 has a loopback device. ns1 and ns2 each have wireguard devices. Those two wireguard devices send traffic through the loopback device of ns0. The end result of this is that tests wind up testing encryption and decryption at the same time -- a pretty CPU and scheduler-heavy workflow. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org Seconds, Fewer Is Better WireGuard + Linux Networking Stack Stress Test AMD Retpolines Former Default New Default - Generic Retpolines 60 120 180 240 300 SE +/- 0.99, N = 3 SE +/- 1.39, N = 3 248.76 257.10
OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 16 - Type: Thread AMD Retpolines Former Default New Default - Generic Retpolines 14 28 42 56 70 SE +/- 0.11, N = 3 SE +/- 0.19, N = 3 62.03 62.20 1. (CC) gcc options: -lpthread
OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 8 - Type: Process AMD Retpolines Former Default New Default - Generic Retpolines 7 14 21 28 35 SE +/- 0.17, N = 3 SE +/- 0.17, N = 3 28.83 29.95 1. (CC) gcc options: -lpthread
OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 16 - Type: Process AMD Retpolines Former Default New Default - Generic Retpolines 13 26 39 52 65 SE +/- 0.23, N = 3 SE +/- 0.16, N = 3 58.36 58.78 1. (CC) gcc options: -lpthread
OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 32 - Type: Process AMD Retpolines Former Default New Default - Generic Retpolines 30 60 90 120 150 SE +/- 0.59, N = 3 SE +/- 0.30, N = 3 117.98 119.83 1. (CC) gcc options: -lpthread
OSBench OSBench is a collection of micro-benchmarks for measuring operating system primitives like time to create threads/processes, launching programs, creating files, and memory allocation. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Files AMD Retpolines Former Default New Default - Generic Retpolines 4 8 12 16 20 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 13.72 14.71 1. (CC) gcc options: -lm
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Threads AMD Retpolines Former Default New Default - Generic Retpolines 4 8 12 16 20 SE +/- 0.03, N = 3 SE +/- 0.09, N = 3 17.70 17.38 1. (CC) gcc options: -lm
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Launch Programs AMD Retpolines Former Default New Default - Generic Retpolines 11 22 33 44 55 SE +/- 0.53, N = 5 SE +/- 0.10, N = 3 48.63 49.36 1. (CC) gcc options: -lm
OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Processes AMD Retpolines Former Default New Default - Generic Retpolines 10 20 30 40 50 SE +/- 0.54, N = 3 SE +/- 0.23, N = 3 43.36 43.14 1. (CC) gcc options: -lm
Sockperf This is a network socket API performance benchmark developed by Mellanox. This test profile runs both the client and server on the local host for evaluating individual system performance. Learn more via the OpenBenchmarking.org test page.
OpenBenchmarking.org usec, Fewer Is Better Sockperf 3.7 Test: Latency Under Load AMD Retpolines Former Default New Default - Generic Retpolines 0.8249 1.6498 2.4747 3.2996 4.1245 SE +/- 0.033, N = 25 SE +/- 0.035, N = 5 3.641 3.666 1. (CXX) g++ options: --param -O3 -rdynamic -ldl -lpthread
New Default - Generic Retpolines Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vDisk Notes: NONE / errors=remount-ro,relatime,rw / Block Size: 4096Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001119Java Notes: OpenJDK Runtime Environment (build 11.0.13+8-Ubuntu-0ubuntu1.21.04)Python Notes: Python 3.9.5Security Notes: 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
Testing initiated at 10 March 2022 15:32 by user phoronix.
AMD Retpolines Former Default Processor: AMD EPYC 72F3 8-Core @ 3.70GHz (8 Cores / 16 Threads), Motherboard: Supermicro H12SSL-i v1.01 (2.0 BIOS), Chipset: AMD Starship/Matisse, Memory: 128GB, Disk: 3841GB Micron_9300_MTFDHAL3T8TDP + 1000GB Corsair Force MP600, Graphics: ASPEED, Monitor: VE228, Network: 2 x Broadcom NetXtreme BCM5720 2-port PCIe
OS: Ubuntu 21.04, Kernel: 5.17.0-051700rc7daily20220310-generic (x86_64), Desktop: GNOME Shell 3.38.4, Display Server: X Server, Compiler: GCC 10.3.0, File-System: ext4, Screen Resolution: 1920x1080
Kernel Notes: Transparent Huge Pages: madviseCompiler Notes: --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 -vDisk Notes: NONE / errors=remount-ro,relatime,rw / Block Size: 4096Processor Notes: Scaling Governor: acpi-cpufreq schedutil (Boost: Enabled) - CPU Microcode: 0xa001119Java Notes: OpenJDK Runtime Environment (build 11.0.13+8-Ubuntu-0ubuntu1.21.04)Python Notes: Python 3.9.5Security Notes: 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
Testing initiated at 11 March 2022 04:16 by user phoronix.