Kernel LLVM AMD Phenom II X4 965 testing with a MSI 890FXA-GD70 (MS-7640) v1.0 (V1.8 BIOS) and MSI NVIDIA GeForce GTX 650 Ti 1GB on ManjaroLinux 21.0.4 via the Phoronix Test Suite.
HTML result view exported from: https://openbenchmarking.org/result/2105128-HA-2105107HA77&grs .
Kernel LLVM Processor Motherboard Chipset Memory Disk Graphics Audio Network OS Kernel Desktop Display Server Display Driver OpenGL Vulkan Compiler File-System Screen Resolution Kernel LLVM Kernel GCC Kernel LLVM 11.1 AMD Phenom II X4 965 @ 3.30GHz (4 Cores) MSI 890FXA-GD70 (MS-7640) v1.0 (V1.8 BIOS) AMD RD890 PCI-e GFX Hydra part 16GB 116GB Samsung SSD 840 + 200GB Seagate ST3200822AS MSI NVIDIA GeForce GTX 650 Ti 1GB Realtek ALC892 Realtek RTL8111/8168/8411 ManjaroLinux 21.0.4 5.12.2-153-tkg-pds-llvm (x86_64) KDE Plasma 5.21.4 X Server 1.20.11 NVIDIA 465.27 4.6.0 1.2.168 GCC 10.2.0 + Clang 13.0.0 btrfs 1920x1080 5.12.2-153-tkg-pds (x86_64) 5.12.2-153-tkg-pds-llvm (x86_64) GCC 10.2.0 + Clang 11.1.0 + LLVM 11.1.0 OpenBenchmarking.org Kernel Details - Transparent Huge Pages: always Environment Details - __GL_ExperimentalPerfStrategy=1 __GL_IGNORE_GLSL_EXT_REQS=1 __GL_THREADED_OPTIMIZATIONS=1 __GL_SHADER_DISK_CACHE=1 __GL_SHADER_CACHE=1 __GL_ALLOW_UNOFFICIAL_PROTOCOL=1 Disk Details - BFQ / commit=120,compress=zstd:3,discard=async,lazytime,noatime,rw,space_cache,ssd_spread,subvol=/@home,subvolid=258 / Block Size: 4096 Processor Details - Scaling Governor: acpi-cpufreq performance - CPU Microcode: 0x10000c8 Security Details - itlb_multihit: Not affected + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Not affected + spectre_v1: Vulnerable: __user pointer sanitization and usercopy barriers only; no swapgs barriers + spectre_v2: Vulnerable STIBP: disabled + srbds: Not affected + tsx_async_abort: Not affected
Kernel LLVM perf-bench: Futex Lock-Pi hackbench: 16 - Thread ipc-benchmark: FIFO Named Pipe - 128 hackbench: 8 - Thread hackbench: 8 - Process ipc-benchmark: Unnamed Pipe - 128 hackbench: 16 - Process hackbench: 4 - Process hackbench: 4 - Thread hackbench: 1 - Thread stress-ng: System V Message Passing hackbench: 2 - Thread ipc-benchmark: Unnamed Unix Domain Socket - 2048 apache: Static Web Page Serving stress-ng: Context Switching wireguard: hackbench: 2 - Process perf-bench: Epoll Wait osbench: Create Files osbench: Launch Programs ipc-benchmark: FIFO Named Pipe - 256 ipc-benchmark: Unnamed Unix Domain Socket - 4096 leveldb: Rand Read sqlite-speedtest: Timed Time - Size 1,000 mbw: Memory Copy, Fixed Block Size - 512 MiB ipc-benchmark: Unnamed Unix Domain Socket - 512 iperf: 5201 - 10 Seconds - UDP - 1000Mbit Objective - 32 t-test1: 1 leveldb: Overwrite leveldb: Overwrite stress-ng: Socket Activity iperf: 5201 - 10 Seconds - TCP - 1 mbw: Memory Copy, Fixed Block Size - 1024 MiB leveldb: Rand Fill postmark: Disk Transaction Performance ipc-benchmark: Unnamed Unix Domain Socket - 1024 stress-ng: SENDFILE leveldb: Rand Fill osbench: Memory Allocations stress-ng: Semaphores ipc-benchmark: FIFO Named Pipe - 4096 osbench: Create Processes perf-bench: Sched Pipe leveldb: Fill Sync pmbench: 4 - 100% Reads ipc-benchmark: TCP Socket - 1024 leveldb: Hot Read mbw: Memory Copy, Fixed Block Size - 4096 MiB pmbench: 4 - 50% stress-ng: Forking leveldb: Fill Sync mbw: Memory Copy - 4096 MiB leveldb: Seq Fill leveldb: Seq Fill pmbench: 1 - 80% Reads 20% Writes ipc-benchmark: Unnamed Pipe - 256 ipc-benchmark: Unnamed Pipe - 1024 iperf: 5201 - 10 Seconds - UDP - 1000Mbit Objective - 64 pmbench: 1 - 100% Reads ipc-benchmark: Unnamed Unix Domain Socket - 128 perf-bench: Futex Hash mbw: Memory Copy, Fixed Block Size - 128 MiB pmbench: 2 - 100% Writes ipc-benchmark: Unnamed Unix Domain Socket - 256 mbw: Memory Copy - 512 MiB ipc-benchmark: Unnamed Pipe - 4096 leveldb: Rand Delete ipc-benchmark: TCP Socket - 4096 rocksdb: Rand Fill ipc-benchmark: Unnamed Pipe - 512 mbw: Memory Copy - 1024 MiB rocksdb: Rand Fill Sync perf-bench: Syscall Basic mbw: Memory Copy - 128 MiB t-test1: 2 rocksdb: Seq Fill pmbench: 2 - 50% rocksdb: Rand Read ipc-benchmark: FIFO Named Pipe - 512 pmbench: 1 - 100% Writes leveldb: Seek Rand stress-ng: Glibc C String Functions pmbench: 2 - 80% Reads 20% Writes stress-ng: MEMFD ipc-benchmark: TCP Socket - 256 openssl: RSA 4096-bit Performance ipc-benchmark: TCP Socket - 128 pmbench: 4 - 100% Writes ipc-benchmark: FIFO Named Pipe - 1024 pmbench: 2 - 100% Reads mutex: Mutex Lock Unlock std::mutex stress-ng: Glibc Qsort Data Sorting stress-ng: Malloc mutex: Shared Mutex Lock Shared perf-bench: Memcpy 1MB mutex: Mutex Lock Unlock spinlock ipc-benchmark: TCP Socket - 512 pmbench: 1 - 50% mutex: Mutex Lock Unlock std::mutex ipc-benchmark: Unnamed Pipe - 2048 ctx-clock: Context Switch Time ipc-benchmark: TCP Socket - 2048 stress-ng: Atomic mutex: Mutex Lock Unlock spinlock_amd mutex: Semaphore Release And Acquire ipc-benchmark: FIFO Named Pipe - 2048 stress-ng: Memory Copying pmbench: 4 - 80% Reads 20% Writes tinymembench: Standard Memset stress-ng: CPU Stress mutex: Mutex Lock Unlock ticket_spinlock stress-ng: Crypto tinymembench: Standard Memcpy perf-bench: Memset 1MB stress-ng: Matrix Math stress-ng: Vector Math stress-ng: NUMA mutex: Mutex Lock Unlock pthread_mutex pgbench: 100 - 1 - Read Only - Average Latency pgbench: 100 - 1 - Read Only pgbench: 1 - 250 - Read Only - Average Latency pgbench: 1 - 250 - Read Only pgbench: 1 - 100 - Read Only - Average Latency pgbench: 1 - 100 - Read Only pgbench: 1 - 50 - Read Only - Average Latency pgbench: 1 - 50 - Read Only pgbench: 1 - 1 - Read Write - Average Latency pgbench: 1 - 1 - Read Write pgbench: 1 - 1 - Read Only - Average Latency pgbench: 1 - 1 - Read Only iperf: 5201 - 10 Seconds - UDP - 64 iperf: 5201 - 10 Seconds - UDP - 32 iperf: 5201 - 10 Seconds - UDP - 1 iperf: 5201 - 10 Seconds - UDP - 100Mbit Objective - 64 iperf: 5201 - 10 Seconds - UDP - 100Mbit Objective - 32 iperf: 5201 - 10 Seconds - UDP - 1000Mbit Objective - 1 iperf: 5201 - 10 Seconds - UDP - 100Mbit Objective - 1 schbench: 8 - 4 rocksdb: Read While Writing stress-ng: CPU Cache stress-ng: MMAP hackbench: 32 - Process hackbench: 1 - Process osbench: Create Threads iperf: 5201 - 10 Seconds - TCP - 64 iperf: 5201 - 10 Seconds - TCP - 32 Kernel LLVM Kernel GCC Kernel LLVM 11.1 9738 168.261 2801157 88.294 79.201 2891727 159.749 40.361 44.099 11.979 4225731.83 23.060 520300 16754.67 548376.44 550.068 20.910 365409 38.822397 95.507304 2278454 457959 2.293 132.190 3548.362 1151175 21412 31.999 30.3 14.570 1877.82 22637 3575.719 14.481 3713 936841 53687.47 30.5 141.110659 1229760.03 465636 40.930907 214271 17.679 0.0655 1453827 2.310 3600.373 0.1191 19252.07 24.9 5105.868 31.6 13.957 0.1175 2351761 1285487 20179 0.0623 1363561 7193341 3623.863 0.1065 1204312 5146.671 468875 13.453 550089 452057 1772787 5152.911 490063 16627566 5172.974 11.041 534770 0.1161 13123433 1746678 0.1038 3.359 112501.83 0.1200 9.01 2539979 441.7 2985001 0.1132 1280989 0.0640 25.5 36.82 11148830.46 31.9 5.969103 42.9 2062968 0.1128 35.0 814387 264 950880 167420.30 46.9 19.0 807480 1019.43 0.1240 5702.3 825.59 24.9 644.00 5825.3 12.033610 11754.15 9619.98 38 21.3 0.050 19914 4.305 58083 1.445 69196 0.685 73013 7.225 138 0.041 24331 67.1 33.6 1.05 6400 3200 1000 100 33519 244439 6.84 12.45 369.548 10.387 16.550064 18133 20013 13155 153.611 3043686 77.877 72.842 3040261 144.751 37.050 39.901 10.963 4626477.31 21.534 559665 18155.45 530566.53 513.194 19.354 394078 36.750389 91.060797 2299703 473559 2.411 125.962 3719.911 1171213 21434 31.274 31.2 14.125 1957.06 22822 3724.204 14.137 3866 964065 55850.50 31.2 135.699351 1277296.32 483511 39.486885 209932 17.435 0.0649 1421362 2.389 3722.106 0.1206 19887.81 25.1 5267.712 32.6 13.547 0.1159 2337377 1316913 20411 0.0632 1374745 7113178 3715.488 0.1059 1233688 5267.520 479683 13.161 555085 458917 1809330 5257.927 499640 16907939 5259.714 10.939 544572 0.1176 13347834 1776333 0.1032 3.393 113135.64 0.1209 9.0 2517133 448.0 2971343 0.1129 1297216 0.0637 25.2 36.74 11170312.20 31.6 5.913378 42.5 2054086 0.1135 34.7 819679 266 944138 167641.36 46.6 19.1 811570 1014.42 0.1235 5691.5 828.24 24.8 642.70 5826.1 11.998912 11754.94 9629.61 38 21.3 67.1 33.6 1.05 6400 3200 1000 100 28597 239799 6.90 13.13 317.405 10.084 15.745640 18304 20700 13043 176.700 2649855 88.562 81.684 2722030 161.291 41.255 44.300 12.146 4244091.68 23.498 515245 16830.43 574464.00 555.201 20.888 375558 39.395617 97.432931 2150341 448144 2.353 131.720 3693.634 1120322 20504 32.656 29.9 14.731 1912.64 21899 3705.678 14.722 3713 926277 54398.51 30.0 137.383382 1274218.42 469668 40.313403 217433 18.057 0.0672 1471645 2.317 3689.686 0.1231 19280.77 24.3 5240.097 32.2 13.725 0.1194 2283738 1279652 19835 0.0641 1336817 7006665 3648.544 0.1085 1209161 5233.318 471693 13.462 562576 449078 1798509 5252.537 490666 16586283 5160.392 11.144 535796 0.1182 13283954 1747103 0.1049 3.412 111451.53 0.1218 8.88 2503889 447.8 2943227 0.1145 1281112 0.0645 25.3 36.44 11061908.02 31.9 5.954150 42.5 2044677 0.1138 35.0 812768 264 948187 168572.72 46.7 19.0 808658 1014.43 0.1234 5677.3 824.83 24.8 644.6 5809.3 12.017047 11766.64 9625.42 38 21.3 67.1 33.6 1.05 6400 3200 1000 100 28747 247556 6.69 12.71 349.937 11.458 16.019980 16653 17735 OpenBenchmarking.org
perf-bench Benchmark: Futex Lock-Pi OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Futex Lock-Pi Kernel LLVM Kernel GCC Kernel LLVM 11.1 3K 6K 9K 12K 15K SE +/- 7.69, N = 3 SE +/- 20.00, N = 3 SE +/- 12.14, N = 3 9738 13155 13043 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
Hackbench Count: 16 - Type: Thread OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 16 - Type: Thread Kernel LLVM Kernel GCC Kernel LLVM 11.1 40 80 120 160 200 SE +/- 1.75, N = 3 SE +/- 0.69, N = 3 SE +/- 2.17, N = 4 168.26 153.61 176.70 1. (CC) gcc options: -lpthread
IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 128 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 128 Kernel LLVM Kernel GCC Kernel LLVM 11.1 700K 1400K 2100K 2800K 3500K SE +/- 18345.59, N = 3 SE +/- 29770.09, N = 3 SE +/- 35347.02, N = 3 2801157 3043686 2649855
Hackbench Count: 8 - Type: Thread OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 8 - Type: Thread Kernel LLVM Kernel GCC Kernel LLVM 11.1 20 40 60 80 100 SE +/- 0.97, N = 3 SE +/- 0.23, N = 3 SE +/- 0.81, N = 15 88.29 77.88 88.56 1. (CC) gcc options: -lpthread
Hackbench Count: 8 - Type: Process OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 8 - Type: Process Kernel LLVM Kernel GCC Kernel LLVM 11.1 20 40 60 80 100 SE +/- 0.46, N = 3 SE +/- 0.85, N = 3 SE +/- 0.85, N = 4 79.20 72.84 81.68 1. (CC) gcc options: -lpthread
IPC_benchmark Type: Unnamed Pipe - Message Bytes: 128 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 128 Kernel LLVM Kernel GCC Kernel LLVM 11.1 700K 1400K 2100K 2800K 3500K SE +/- 22099.44, N = 3 SE +/- 34085.75, N = 4 SE +/- 11315.58, N = 3 2891727 3040261 2722030
Hackbench Count: 16 - Type: Process OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 16 - Type: Process Kernel LLVM Kernel GCC Kernel LLVM 11.1 40 80 120 160 200 SE +/- 0.62, N = 3 SE +/- 0.22, N = 3 SE +/- 1.83, N = 3 159.75 144.75 161.29 1. (CC) gcc options: -lpthread
Hackbench Count: 4 - Type: Process OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 4 - Type: Process Kernel LLVM Kernel GCC Kernel LLVM 11.1 9 18 27 36 45 SE +/- 0.08, N = 3 SE +/- 0.12, N = 3 SE +/- 0.16, N = 3 40.36 37.05 41.26 1. (CC) gcc options: -lpthread
Hackbench Count: 4 - Type: Thread OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 4 - Type: Thread Kernel LLVM Kernel GCC Kernel LLVM 11.1 10 20 30 40 50 SE +/- 0.16, N = 3 SE +/- 0.35, N = 3 SE +/- 0.11, N = 3 44.10 39.90 44.30 1. (CC) gcc options: -lpthread
Hackbench Count: 1 - Type: Thread OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 1 - Type: Thread Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.06, N = 3 11.98 10.96 12.15 1. (CC) gcc options: -lpthread
Stress-NG Test: System V Message Passing OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: System V Message Passing Kernel LLVM Kernel GCC Kernel LLVM 11.1 1000K 2000K 3000K 4000K 5000K SE +/- 2913.62, N = 3 SE +/- 32996.49, N = 12 SE +/- 18462.77, N = 3 4225731.83 4626477.31 4244091.68 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Hackbench Count: 2 - Type: Thread OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 2 - Type: Thread Kernel LLVM Kernel GCC Kernel LLVM 11.1 6 12 18 24 30 SE +/- 0.01, N = 3 SE +/- 0.16, N = 3 SE +/- 0.03, N = 3 23.06 21.53 23.50 1. (CC) gcc options: -lpthread
IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 2048 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 2048 Kernel LLVM Kernel GCC Kernel LLVM 11.1 120K 240K 360K 480K 600K SE +/- 359.60, N = 3 SE +/- 1280.33, N = 3 SE +/- 1696.10, N = 3 520300 559665 515245
Apache Benchmark Static Web Page Serving OpenBenchmarking.org Requests Per Second, More Is Better Apache Benchmark 2.4.29 Static Web Page Serving Kernel LLVM Kernel GCC Kernel LLVM 11.1 4K 8K 12K 16K 20K SE +/- 28.17, N = 3 SE +/- 31.87, N = 3 SE +/- 18.45, N = 3 16754.67 18155.45 16830.43 1. (CC) gcc options: -shared -fPIC -O2 -pthread
Stress-NG Test: Context Switching OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Context Switching Kernel LLVM Kernel GCC Kernel LLVM 11.1 120K 240K 360K 480K 600K SE +/- 2752.66, N = 3 SE +/- 4193.19, N = 3 SE +/- 2034.95, N = 3 548376.44 530566.53 574464.00 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
WireGuard + Linux Networking Stack Stress Test OpenBenchmarking.org Seconds, Fewer Is Better WireGuard + Linux Networking Stack Stress Test Kernel LLVM Kernel GCC Kernel LLVM 11.1 120 240 360 480 600 SE +/- 0.12, N = 3 SE +/- 1.35, N = 3 SE +/- 0.65, N = 3 550.07 513.19 555.20
Hackbench Count: 2 - Type: Process OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 2 - Type: Process Kernel LLVM Kernel GCC Kernel LLVM 11.1 5 10 15 20 25 SE +/- 0.22, N = 14 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 20.91 19.35 20.89 1. (CC) gcc options: -lpthread
perf-bench Benchmark: Epoll Wait OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Epoll Wait Kernel LLVM Kernel GCC Kernel LLVM 11.1 80K 160K 240K 320K 400K SE +/- 2596.83, N = 3 SE +/- 4664.03, N = 3 SE +/- 1651.40, N = 3 365409 394078 375558 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
OSBench Test: Create Files OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Files Kernel LLVM Kernel GCC Kernel LLVM 11.1 9 18 27 36 45 SE +/- 0.24, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 38.82 36.75 39.40 -lm -lm 1. (CC) gcc options:
OSBench Test: Launch Programs OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Launch Programs Kernel LLVM Kernel GCC Kernel LLVM 11.1 20 40 60 80 100 SE +/- 0.20, N = 3 SE +/- 0.19, N = 3 SE +/- 0.64, N = 3 95.51 91.06 97.43 -lm -lm 1. (CC) gcc options:
IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 256 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 256 Kernel LLVM Kernel GCC Kernel LLVM 11.1 500K 1000K 1500K 2000K 2500K SE +/- 5587.94, N = 3 SE +/- 4874.11, N = 3 SE +/- 4551.71, N = 3 2278454 2299703 2150341
IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 4096 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 4096 Kernel LLVM Kernel GCC Kernel LLVM 11.1 100K 200K 300K 400K 500K SE +/- 1305.06, N = 3 SE +/- 1755.44, N = 3 SE +/- 680.61, N = 3 457959 473559 448144
LevelDB Benchmark: Random Read OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Read Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.5425 1.085 1.6275 2.17 2.7125 SE +/- 0.027, N = 4 SE +/- 0.019, N = 9 SE +/- 0.021, N = 7 2.293 2.411 2.353 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
SQLite Speedtest Timed Time - Size 1,000 OpenBenchmarking.org Seconds, Fewer Is Better SQLite Speedtest 3.30 Timed Time - Size 1,000 Kernel LLVM Kernel GCC Kernel LLVM 11.1 30 60 90 120 150 SE +/- 0.52, N = 3 SE +/- 0.26, N = 3 SE +/- 0.40, N = 3 132.19 125.96 131.72 1. (CC) gcc options: -O2 -ldl -lz -lpthread
MBW Test: Memory Copy, Fixed Block Size - Array Size: 512 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy, Fixed Block Size - Array Size: 512 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 800 1600 2400 3200 4000 SE +/- 1.95, N = 3 SE +/- 4.41, N = 3 SE +/- 6.66, N = 3 3548.36 3719.91 3693.63 1. (CC) gcc options: -O3 -march=native
IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 512 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 512 Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 1165.49, N = 3 SE +/- 4109.25, N = 3 SE +/- 7819.75, N = 12 1151175 1171213 1120322
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 32 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 32 Kernel LLVM Kernel GCC Kernel LLVM 11.1 5K 10K 15K 20K 25K SE +/- 114.44, N = 3 SE +/- 131.93, N = 3 SE +/- 267.82, N = 3 21412 21434 20504 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
t-test1 Threads: 1 OpenBenchmarking.org Seconds, Fewer Is Better t-test1 2017-01-13 Threads: 1 Kernel LLVM Kernel GCC Kernel LLVM 11.1 8 16 24 32 40 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 32.00 31.27 32.66 1. (CC) gcc options: -pthread
LevelDB Benchmark: Overwrite OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Overwrite Kernel LLVM Kernel GCC Kernel LLVM 11.1 7 14 21 28 35 SE +/- 0.19, N = 3 SE +/- 0.23, N = 3 SE +/- 0.17, N = 3 30.3 31.2 29.9 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
LevelDB Benchmark: Overwrite OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Overwrite Kernel LLVM Kernel GCC Kernel LLVM 11.1 4 8 12 16 20 SE +/- 0.09, N = 3 SE +/- 0.12, N = 3 SE +/- 0.09, N = 3 14.57 14.13 14.73 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Stress-NG Test: Socket Activity OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Socket Activity Kernel LLVM Kernel GCC Kernel LLVM 11.1 400 800 1200 1600 2000 SE +/- 1.75, N = 3 SE +/- 2.69, N = 3 SE +/- 1.60, N = 3 1877.82 1957.06 1912.64 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 1 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 1 Kernel LLVM Kernel GCC Kernel LLVM 11.1 5K 10K 15K 20K 25K SE +/- 273.77, N = 15 SE +/- 85.32, N = 3 SE +/- 212.79, N = 3 22637 22822 21899 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
MBW Test: Memory Copy, Fixed Block Size - Array Size: 1024 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy, Fixed Block Size - Array Size: 1024 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 800 1600 2400 3200 4000 SE +/- 5.31, N = 3 SE +/- 1.03, N = 3 SE +/- 3.36, N = 3 3575.72 3724.20 3705.68 1. (CC) gcc options: -O3 -march=native
LevelDB Benchmark: Random Fill OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Fill Kernel LLVM Kernel GCC Kernel LLVM 11.1 4 8 12 16 20 SE +/- 0.06, N = 3 SE +/- 0.11, N = 3 SE +/- 0.07, N = 3 14.48 14.14 14.72 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
PostMark Disk Transaction Performance OpenBenchmarking.org TPS, More Is Better PostMark 1.51 Disk Transaction Performance Kernel LLVM Kernel GCC Kernel LLVM 11.1 800 1600 2400 3200 4000 SE +/- 18.33, N = 3 SE +/- 20.00, N = 3 SE +/- 18.33, N = 3 3713 3866 3713 1. (CC) gcc options: -O3
IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 1024 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 1024 Kernel LLVM Kernel GCC Kernel LLVM 11.1 200K 400K 600K 800K 1000K SE +/- 1701.21, N = 3 SE +/- 4520.85, N = 3 SE +/- 3059.81, N = 3 936841 964065 926277
Stress-NG Test: SENDFILE OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: SENDFILE Kernel LLVM Kernel GCC Kernel LLVM 11.1 12K 24K 36K 48K 60K SE +/- 107.63, N = 3 SE +/- 144.65, N = 3 SE +/- 46.31, N = 3 53687.47 55850.50 54398.51 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
LevelDB Benchmark: Random Fill OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Random Fill Kernel LLVM Kernel GCC Kernel LLVM 11.1 7 14 21 28 35 SE +/- 0.13, N = 3 SE +/- 0.24, N = 3 SE +/- 0.12, N = 3 30.5 31.2 30.0 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
OSBench Test: Memory Allocations OpenBenchmarking.org Ns Per Event, Fewer Is Better OSBench Test: Memory Allocations Kernel LLVM Kernel GCC Kernel LLVM 11.1 30 60 90 120 150 SE +/- 0.44, N = 3 SE +/- 0.11, N = 3 SE +/- 0.34, N = 3 141.11 135.70 137.38 -lm -lm 1. (CC) gcc options:
Stress-NG Test: Semaphores OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Semaphores Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 10868.11, N = 15 SE +/- 9507.55, N = 15 SE +/- 14472.05, N = 3 1229760.03 1277296.32 1274218.42 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 4096 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 4096 Kernel LLVM Kernel GCC Kernel LLVM 11.1 100K 200K 300K 400K 500K SE +/- 2657.74, N = 3 SE +/- 1945.98, N = 3 SE +/- 3051.56, N = 3 465636 483511 469668
OSBench Test: Create Processes OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Processes Kernel LLVM Kernel GCC Kernel LLVM 11.1 9 18 27 36 45 SE +/- 0.18, N = 3 SE +/- 0.28, N = 3 SE +/- 0.16, N = 3 40.93 39.49 40.31 -lm -lm 1. (CC) gcc options:
perf-bench Benchmark: Sched Pipe OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Sched Pipe Kernel LLVM Kernel GCC Kernel LLVM 11.1 50K 100K 150K 200K 250K SE +/- 896.38, N = 3 SE +/- 569.18, N = 3 SE +/- 861.60, N = 3 214271 209932 217433 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
LevelDB Benchmark: Fill Sync OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Fill Sync Kernel LLVM Kernel GCC Kernel LLVM 11.1 4 8 12 16 20 SE +/- 0.15, N = 3 SE +/- 0.26, N = 4 SE +/- 0.19, N = 3 17.68 17.44 18.06 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 100% Reads OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 100% Reads Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0151 0.0302 0.0453 0.0604 0.0755 SE +/- 0.0007, N = 3 SE +/- 0.0008, N = 3 SE +/- 0.0006, N = 3 0.0655 0.0649 0.0672 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
IPC_benchmark Type: TCP Socket - Message Bytes: 1024 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: TCP Socket - Message Bytes: 1024 Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 2398.52, N = 3 SE +/- 2423.58, N = 3 SE +/- 3846.75, N = 3 1453827 1421362 1471645
LevelDB Benchmark: Hot Read OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Hot Read Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.5375 1.075 1.6125 2.15 2.6875 SE +/- 0.006, N = 3 SE +/- 0.023, N = 3 SE +/- 0.027, N = 3 2.310 2.389 2.317 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
MBW Test: Memory Copy, Fixed Block Size - Array Size: 4096 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy, Fixed Block Size - Array Size: 4096 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 800 1600 2400 3200 4000 SE +/- 37.05, N = 3 SE +/- 0.67, N = 3 SE +/- 3.81, N = 3 3600.37 3722.11 3689.69 1. (CC) gcc options: -O3 -march=native
pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 50% OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 50% Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0277 0.0554 0.0831 0.1108 0.1385 SE +/- 0.0004, N = 3 SE +/- 0.0006, N = 3 SE +/- 0.0016, N = 3 0.1191 0.1206 0.1231 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
Stress-NG Test: Forking OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Forking Kernel LLVM Kernel GCC Kernel LLVM 11.1 4K 8K 12K 16K 20K SE +/- 39.67, N = 3 SE +/- 41.17, N = 3 SE +/- 97.53, N = 3 19252.07 19887.81 19280.77 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
LevelDB Benchmark: Fill Sync OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Fill Sync Kernel LLVM Kernel GCC Kernel LLVM 11.1 6 12 18 24 30 SE +/- 0.21, N = 3 SE +/- 0.31, N = 4 SE +/- 0.26, N = 3 24.9 25.1 24.3 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
MBW Test: Memory Copy - Array Size: 4096 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy - Array Size: 4096 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 1100 2200 3300 4400 5500 SE +/- 62.83, N = 3 SE +/- 5.16, N = 3 SE +/- 9.07, N = 3 5105.87 5267.71 5240.10 1. (CC) gcc options: -O3 -march=native
LevelDB Benchmark: Sequential Fill OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Sequential Fill Kernel LLVM Kernel GCC Kernel LLVM 11.1 8 16 24 32 40 SE +/- 0.12, N = 3 SE +/- 0.10, N = 3 SE +/- 0.12, N = 3 31.6 32.6 32.2 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
LevelDB Benchmark: Sequential Fill OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Sequential Fill Kernel LLVM Kernel GCC Kernel LLVM 11.1 4 8 12 16 20 SE +/- 0.05, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 13.96 13.55 13.73 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 80% Reads 20% Writes OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 80% Reads 20% Writes Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0269 0.0538 0.0807 0.1076 0.1345 SE +/- 0.0012, N = 4 SE +/- 0.0011, N = 3 SE +/- 0.0010, N = 3 0.1175 0.1159 0.1194 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
IPC_benchmark Type: Unnamed Pipe - Message Bytes: 256 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 256 Kernel LLVM Kernel GCC Kernel LLVM 11.1 500K 1000K 1500K 2000K 2500K SE +/- 31763.45, N = 3 SE +/- 17228.51, N = 3 SE +/- 18436.29, N = 3 2351761 2337377 2283738
IPC_benchmark Type: Unnamed Pipe - Message Bytes: 1024 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 1024 Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 3654.53, N = 3 SE +/- 9574.67, N = 3 SE +/- 16798.04, N = 3 1285487 1316913 1279652
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 64 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 64 Kernel LLVM Kernel GCC Kernel LLVM 11.1 4K 8K 12K 16K 20K SE +/- 67.90, N = 3 SE +/- 36.96, N = 3 SE +/- 13.02, N = 3 20179 20411 19835 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 100% Reads OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 100% Reads Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0144 0.0288 0.0432 0.0576 0.072 SE +/- 0.0005, N = 15 SE +/- 0.0006, N = 7 SE +/- 0.0003, N = 3 0.0623 0.0632 0.0641 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 128 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 128 Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 4775.86, N = 3 SE +/- 4747.41, N = 3 SE +/- 5815.34, N = 3 1363561 1374745 1336817
perf-bench Benchmark: Futex Hash OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Futex Hash Kernel LLVM Kernel GCC Kernel LLVM 11.1 1.5M 3M 4.5M 6M 7.5M SE +/- 9923.07, N = 3 SE +/- 5828.57, N = 3 SE +/- 9731.70, N = 3 7193341 7113178 7006665 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
MBW Test: Memory Copy, Fixed Block Size - Array Size: 128 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy, Fixed Block Size - Array Size: 128 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 800 1600 2400 3200 4000 SE +/- 3.01, N = 3 SE +/- 2.94, N = 3 SE +/- 1.33, N = 3 3623.86 3715.49 3648.54 1. (CC) gcc options: -O3 -march=native
pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 100% Writes OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 100% Writes Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0244 0.0488 0.0732 0.0976 0.122 SE +/- 0.0014, N = 3 SE +/- 0.0011, N = 3 SE +/- 0.0009, N = 3 0.1065 0.1059 0.1085 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 256 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Unix Domain Socket - Message Bytes: 256 Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 4752.07, N = 3 SE +/- 1670.87, N = 3 SE +/- 7325.27, N = 3 1204312 1233688 1209161
MBW Test: Memory Copy - Array Size: 512 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy - Array Size: 512 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 1100 2200 3300 4400 5500 SE +/- 12.30, N = 3 SE +/- 8.07, N = 3 SE +/- 9.52, N = 3 5146.67 5267.52 5233.32 1. (CC) gcc options: -O3 -march=native
IPC_benchmark Type: Unnamed Pipe - Message Bytes: 4096 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 4096 Kernel LLVM Kernel GCC Kernel LLVM 11.1 100K 200K 300K 400K 500K SE +/- 3477.35, N = 3 SE +/- 4438.60, N = 3 SE +/- 1532.35, N = 3 468875 479683 471693
LevelDB Benchmark: Random Delete OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Delete Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.15, N = 3 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 13.45 13.16 13.46 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
IPC_benchmark Type: TCP Socket - Message Bytes: 4096 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: TCP Socket - Message Bytes: 4096 Kernel LLVM Kernel GCC Kernel LLVM 11.1 120K 240K 360K 480K 600K SE +/- 1970.04, N = 3 SE +/- 4984.28, N = 3 SE +/- 1685.44, N = 3 550089 555085 562576
Facebook RocksDB Test: Random Fill OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill Kernel LLVM Kernel GCC Kernel LLVM 11.1 100K 200K 300K 400K 500K SE +/- 3333.72, N = 3 SE +/- 3750.21, N = 3 SE +/- 3376.28, N = 3 452057 458917 449078 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
IPC_benchmark Type: Unnamed Pipe - Message Bytes: 512 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 512 Kernel LLVM Kernel GCC Kernel LLVM 11.1 400K 800K 1200K 1600K 2000K SE +/- 1897.32, N = 3 SE +/- 16098.73, N = 3 SE +/- 15480.54, N = 15 1772787 1809330 1798509
MBW Test: Memory Copy - Array Size: 1024 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy - Array Size: 1024 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 1100 2200 3300 4400 5500 SE +/- 18.98, N = 3 SE +/- 1.90, N = 3 SE +/- 7.65, N = 3 5152.91 5257.93 5252.54 1. (CC) gcc options: -O3 -march=native
Facebook RocksDB Test: Random Fill Sync OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill Sync Kernel LLVM Kernel GCC Kernel LLVM 11.1 110K 220K 330K 440K 550K SE +/- 2169.84, N = 3 SE +/- 1981.98, N = 3 SE +/- 4275.49, N = 3 490063 499640 490666 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
perf-bench Benchmark: Syscall Basic OpenBenchmarking.org ops/sec, More Is Better perf-bench Benchmark: Syscall Basic Kernel LLVM Kernel GCC Kernel LLVM 11.1 4M 8M 12M 16M 20M SE +/- 44903.68, N = 3 SE +/- 6821.55, N = 3 SE +/- 12906.55, N = 3 16627566 16907939 16586283 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
MBW Test: Memory Copy - Array Size: 128 MiB OpenBenchmarking.org MiB/s, More Is Better MBW 2018-09-08 Test: Memory Copy - Array Size: 128 MiB Kernel LLVM Kernel GCC Kernel LLVM 11.1 1100 2200 3300 4400 5500 SE +/- 12.43, N = 3 SE +/- 7.70, N = 3 SE +/- 19.56, N = 3 5172.97 5259.71 5160.39 1. (CC) gcc options: -O3 -march=native
t-test1 Threads: 2 OpenBenchmarking.org Seconds, Fewer Is Better t-test1 2017-01-13 Threads: 2 Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.03, N = 3 11.04 10.94 11.14 1. (CC) gcc options: -pthread
Facebook RocksDB Test: Sequential Fill OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Sequential Fill Kernel LLVM Kernel GCC Kernel LLVM 11.1 120K 240K 360K 480K 600K SE +/- 3023.00, N = 3 SE +/- 2578.31, N = 3 SE +/- 6929.94, N = 3 534770 544572 535796 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 50% OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 50% Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0266 0.0532 0.0798 0.1064 0.133 SE +/- 0.0014, N = 3 SE +/- 0.0014, N = 4 SE +/- 0.0000, N = 3 0.1161 0.1176 0.1182 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
Facebook RocksDB Test: Random Read OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Read Kernel LLVM Kernel GCC Kernel LLVM 11.1 3M 6M 9M 12M 15M SE +/- 59134.09, N = 3 SE +/- 85590.38, N = 3 SE +/- 47642.98, N = 3 13123433 13347834 13283954 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 512 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 512 Kernel LLVM Kernel GCC Kernel LLVM 11.1 400K 800K 1200K 1600K 2000K SE +/- 24548.29, N = 3 SE +/- 19241.60, N = 3 SE +/- 8871.35, N = 3 1746678 1776333 1747103
pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 100% Writes OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 100% Writes Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0236 0.0472 0.0708 0.0944 0.118 SE +/- 0.0012, N = 3 SE +/- 0.0010, N = 3 SE +/- 0.0001, N = 3 0.1038 0.1032 0.1049 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
LevelDB Benchmark: Seek Random OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Seek Random Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.7677 1.5354 2.3031 3.0708 3.8385 SE +/- 0.029, N = 3 SE +/- 0.029, N = 3 SE +/- 0.028, N = 3 3.359 3.393 3.412 -lsqlite3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Stress-NG Test: Glibc C String Functions OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Glibc C String Functions Kernel LLVM Kernel GCC Kernel LLVM 11.1 20K 40K 60K 80K 100K SE +/- 495.62, N = 3 SE +/- 236.51, N = 3 SE +/- 415.80, N = 3 112501.83 113135.64 111451.53 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 80% Reads 20% Writes OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 80% Reads 20% Writes Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0274 0.0548 0.0822 0.1096 0.137 SE +/- 0.0012, N = 3 SE +/- 0.0003, N = 3 SE +/- 0.0004, N = 3 0.1200 0.1209 0.1218 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
Stress-NG Test: MEMFD OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: MEMFD Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 9.01 9.00 8.88 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
IPC_benchmark Type: TCP Socket - Message Bytes: 256 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: TCP Socket - Message Bytes: 256 Kernel LLVM Kernel GCC Kernel LLVM 11.1 500K 1000K 1500K 2000K 2500K SE +/- 9721.39, N = 3 SE +/- 22225.96, N = 7 SE +/- 16570.75, N = 15 2539979 2517133 2503889
OpenSSL RSA 4096-bit Performance OpenBenchmarking.org Signs Per Second, More Is Better OpenSSL 1.1.1 RSA 4096-bit Performance Kernel LLVM Kernel GCC Kernel LLVM 11.1 100 200 300 400 500 SE +/- 4.10, N = 7 SE +/- 0.18, N = 3 SE +/- 0.03, N = 3 441.7 448.0 447.8
IPC_benchmark Type: TCP Socket - Message Bytes: 128 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: TCP Socket - Message Bytes: 128 Kernel LLVM Kernel GCC Kernel LLVM 11.1 600K 1200K 1800K 2400K 3000K SE +/- 10452.71, N = 3 SE +/- 24608.70, N = 9 SE +/- 41324.96, N = 3 2985001 2971343 2943227
pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 100% Writes OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 100% Writes Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0258 0.0516 0.0774 0.1032 0.129 SE +/- 0.0014, N = 3 SE +/- 0.0001, N = 3 SE +/- 0.0013, N = 3 0.1132 0.1129 0.1145 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 1024 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 1024 Kernel LLVM Kernel GCC Kernel LLVM 11.1 300K 600K 900K 1200K 1500K SE +/- 2998.97, N = 3 SE +/- 1557.87, N = 3 SE +/- 2437.79, N = 3 1280989 1297216 1281112
pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 100% Reads OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 2 - Read-Write Ratio: 100% Reads Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0145 0.029 0.0435 0.058 0.0725 SE +/- 0.0000, N = 3 SE +/- 0.0000, N = 3 SE +/- 0.0008, N = 3 0.0640 0.0637 0.0645 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
BenchmarkMutex Benchmark: Mutex Lock Unlock std::mutex OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Mutex Lock Unlock std::mutex Kernel LLVM Kernel GCC Kernel LLVM 11.1 6 12 18 24 30 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.12, N = 3 25.5 25.2 25.3 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
Stress-NG Test: Glibc Qsort Data Sorting OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Glibc Qsort Data Sorting Kernel LLVM Kernel GCC Kernel LLVM 11.1 8 16 24 32 40 SE +/- 0.24, N = 3 SE +/- 0.37, N = 3 SE +/- 0.44, N = 4 36.82 36.74 36.44 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Stress-NG Test: Malloc OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Malloc Kernel LLVM Kernel GCC Kernel LLVM 11.1 2M 4M 6M 8M 10M SE +/- 12093.70, N = 3 SE +/- 13169.62, N = 3 SE +/- 10798.66, N = 3 11148830.46 11170312.20 11061908.02 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
BenchmarkMutex Benchmark: Shared Mutex Lock Shared OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Shared Mutex Lock Shared Kernel LLVM Kernel GCC Kernel LLVM 11.1 7 14 21 28 35 SE +/- 0.09, N = 3 SE +/- 0.07, N = 3 SE +/- 0.15, N = 3 31.9 31.6 31.9 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
perf-bench Benchmark: Memcpy 1MB OpenBenchmarking.org GB/sec, More Is Better perf-bench Benchmark: Memcpy 1MB Kernel LLVM Kernel GCC Kernel LLVM 11.1 1.343 2.686 4.029 5.372 6.715 SE +/- 0.007854, N = 3 SE +/- 0.051475, N = 3 SE +/- 0.003067, N = 3 5.969103 5.913378 5.954150 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
BenchmarkMutex Benchmark: Mutex Lock Unlock spinlock OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Mutex Lock Unlock spinlock Kernel LLVM Kernel GCC Kernel LLVM 11.1 10 20 30 40 50 SE +/- 0.12, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 42.9 42.5 42.5 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
IPC_benchmark Type: TCP Socket - Message Bytes: 512 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: TCP Socket - Message Bytes: 512 Kernel LLVM Kernel GCC Kernel LLVM 11.1 400K 800K 1200K 1600K 2000K SE +/- 11997.79, N = 3 SE +/- 2605.04, N = 3 SE +/- 21799.46, N = 4 2062968 2054086 2044677
pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 50% OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 1 - Read-Write Ratio: 50% Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0256 0.0512 0.0768 0.1024 0.128 SE +/- 0.0015, N = 3 SE +/- 0.0015, N = 3 SE +/- 0.0007, N = 3 0.1128 0.1135 0.1138 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
BenchmarkMutex Benchmark: Mutex Lock Unlock std::mutex OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Mutex Lock Unlock std::mutex Kernel LLVM Kernel GCC Kernel LLVM 11.1 8 16 24 32 40 SE +/- 0.12, N = 3 SE +/- 0.12, N = 3 SE +/- 0.13, N = 3 35.0 34.7 35.0 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
IPC_benchmark Type: Unnamed Pipe - Message Bytes: 2048 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: Unnamed Pipe - Message Bytes: 2048 Kernel LLVM Kernel GCC Kernel LLVM 11.1 200K 400K 600K 800K 1000K SE +/- 1382.62, N = 3 SE +/- 856.11, N = 3 SE +/- 4500.55, N = 3 814387 819679 812768
ctx_clock Context Switch Time OpenBenchmarking.org Clocks, Fewer Is Better ctx_clock Context Switch Time Kernel LLVM Kernel GCC Kernel LLVM 11.1 60 120 180 240 300 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 264 266 264
IPC_benchmark Type: TCP Socket - Message Bytes: 2048 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: TCP Socket - Message Bytes: 2048 Kernel LLVM Kernel GCC Kernel LLVM 11.1 200K 400K 600K 800K 1000K SE +/- 5079.92, N = 3 SE +/- 8681.93, N = 3 SE +/- 4947.37, N = 3 950880 944138 948187
Stress-NG Test: Atomic OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Atomic Kernel LLVM Kernel GCC Kernel LLVM 11.1 40K 80K 120K 160K 200K SE +/- 324.01, N = 3 SE +/- 368.77, N = 3 SE +/- 560.17, N = 3 167420.30 167641.36 168572.72 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
BenchmarkMutex Benchmark: Mutex Lock Unlock spinlock_amd OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Mutex Lock Unlock spinlock_amd Kernel LLVM Kernel GCC Kernel LLVM 11.1 11 22 33 44 55 SE +/- 0.07, N = 3 SE +/- 0.00, N = 3 SE +/- 0.07, N = 3 46.9 46.6 46.7 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
BenchmarkMutex Benchmark: Semaphore Release And Acquire OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Semaphore Release And Acquire Kernel LLVM Kernel GCC Kernel LLVM 11.1 5 10 15 20 25 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.06, N = 3 19.0 19.1 19.0 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 2048 OpenBenchmarking.org Messages Per Second, More Is Better IPC_benchmark Type: FIFO Named Pipe - Message Bytes: 2048 Kernel LLVM Kernel GCC Kernel LLVM 11.1 200K 400K 600K 800K 1000K SE +/- 1912.41, N = 3 SE +/- 1585.30, N = 3 SE +/- 5943.58, N = 3 807480 811570 808658
Stress-NG Test: Memory Copying OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Memory Copying Kernel LLVM Kernel GCC Kernel LLVM 11.1 200 400 600 800 1000 SE +/- 6.09, N = 3 SE +/- 1.52, N = 3 SE +/- 2.74, N = 3 1019.43 1014.42 1014.43 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 80% Reads 20% Writes OpenBenchmarking.org us - Average Page Latency, Fewer Is Better pmbench Concurrent Worker Threads: 4 - Read-Write Ratio: 80% Reads 20% Writes Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.0279 0.0558 0.0837 0.1116 0.1395 SE +/- 0.0016, N = 3 SE +/- 0.0012, N = 3 SE +/- 0.0014, N = 3 0.1240 0.1235 0.1234 1. (CC) gcc options: -lm -luuid -lxml2 -m64 -pthread
Tinymembench Standard Memset OpenBenchmarking.org MB/s, More Is Better Tinymembench 2018-05-28 Standard Memset Kernel LLVM Kernel GCC Kernel LLVM 11.1 1200 2400 3600 4800 6000 SE +/- 15.70, N = 3 SE +/- 2.89, N = 3 SE +/- 3.21, N = 3 5702.3 5691.5 5677.3 1. (CC) gcc options: -O2 -lm
Stress-NG Test: CPU Stress OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: CPU Stress Kernel LLVM Kernel GCC Kernel LLVM 11.1 200 400 600 800 1000 SE +/- 2.22, N = 3 SE +/- 0.36, N = 3 SE +/- 1.66, N = 3 825.59 828.24 824.83 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
BenchmarkMutex Benchmark: Mutex Lock Unlock ticket_spinlock OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Mutex Lock Unlock ticket_spinlock Kernel LLVM Kernel GCC Kernel LLVM 11.1 6 12 18 24 30 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 24.9 24.8 24.8 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
Stress-NG Test: Crypto OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Crypto Kernel LLVM Kernel GCC Kernel LLVM 11.1 140 280 420 560 700 SE +/- 1.45, N = 3 SE +/- 0.12, N = 3 SE +/- 1.44, N = 3 644.00 642.70 644.60 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Tinymembench Standard Memcpy OpenBenchmarking.org MB/s, More Is Better Tinymembench 2018-05-28 Standard Memcpy Kernel LLVM Kernel GCC Kernel LLVM 11.1 1200 2400 3600 4800 6000 SE +/- 10.38, N = 3 SE +/- 8.19, N = 3 SE +/- 6.50, N = 3 5825.3 5826.1 5809.3 1. (CC) gcc options: -O2 -lm
perf-bench Benchmark: Memset 1MB OpenBenchmarking.org GB/sec, More Is Better perf-bench Benchmark: Memset 1MB Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 12.03 12.00 12.02 1. (CC) gcc options: -O6 -ggdb3 -funwind-tables -std=gnu99 -lunwind-x86_64 -lunwind -llzma -Xlinker -export-dynamic -lpthread -lrt -lm -ldl -lelf -ldebuginfod -ldw -lcrypto -lslang -lpython2.7 -lutil -lbfd -lopcodes -liberty -lz -lzstd -lcap -lnuma
Stress-NG Test: Matrix Math OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Matrix Math Kernel LLVM Kernel GCC Kernel LLVM 11.1 3K 6K 9K 12K 15K SE +/- 16.83, N = 3 SE +/- 17.54, N = 3 SE +/- 0.54, N = 3 11754.15 11754.94 11766.64 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Stress-NG Test: Vector Math OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: Vector Math Kernel LLVM Kernel GCC Kernel LLVM 11.1 2K 4K 6K 8K 10K SE +/- 10.59, N = 3 SE +/- 1.82, N = 3 SE +/- 0.92, N = 3 9619.98 9629.61 9625.42 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Stress-NG Test: NUMA OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: NUMA Kernel LLVM Kernel GCC Kernel LLVM 11.1 9 18 27 36 45 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 38 38 38 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
BenchmarkMutex Benchmark: Mutex Lock Unlock pthread_mutex OpenBenchmarking.org ns, Fewer Is Better BenchmarkMutex Benchmark: Mutex Lock Unlock pthread_mutex Kernel LLVM Kernel GCC Kernel LLVM 11.1 5 10 15 20 25 SE +/- 0.03, N = 3 SE +/- 0.07, N = 3 SE +/- 0.00, N = 3 21.3 21.3 21.3 1. (CXX) g++ options: -std=c++17 -lbenchmark -pthread
PostgreSQL pgbench Scaling Factor: 100 - Clients: 1 - Mode: Read Only - Average Latency OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 13.0 Scaling Factor: 100 - Clients: 1 - Mode: Read Only - Average Latency Kernel LLVM 0.0113 0.0226 0.0339 0.0452 0.0565 SE +/- 0.000, N = 3 0.050 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 100 - Clients: 1 - Mode: Read Only OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 13.0 Scaling Factor: 100 - Clients: 1 - Mode: Read Only Kernel LLVM 4K 8K 12K 16K 20K SE +/- 112.18, N = 3 19914 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 250 - Mode: Read Only - Average Latency OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 250 - Mode: Read Only - Average Latency Kernel LLVM 0.9686 1.9372 2.9058 3.8744 4.843 SE +/- 0.028, N = 3 4.305 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 250 - Mode: Read Only OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 250 - Mode: Read Only Kernel LLVM 12K 24K 36K 48K 60K SE +/- 383.60, N = 3 58083 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 100 - Mode: Read Only - Average Latency OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 100 - Mode: Read Only - Average Latency Kernel LLVM 0.3251 0.6502 0.9753 1.3004 1.6255 SE +/- 0.001, N = 3 1.445 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 100 - Mode: Read Only OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 100 - Mode: Read Only Kernel LLVM 15K 30K 45K 60K 75K SE +/- 50.59, N = 3 69196 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 50 - Mode: Read Only - Average Latency OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 50 - Mode: Read Only - Average Latency Kernel LLVM 0.1541 0.3082 0.4623 0.6164 0.7705 SE +/- 0.003, N = 3 0.685 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 50 - Mode: Read Only OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 50 - Mode: Read Only Kernel LLVM 16K 32K 48K 64K 80K SE +/- 358.11, N = 3 73013 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 1 - Mode: Read Write - Average Latency OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 1 - Mode: Read Write - Average Latency Kernel LLVM 2 4 6 8 10 SE +/- 0.086, N = 3 7.225 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 1 - Mode: Read Write OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 1 - Mode: Read Write Kernel LLVM 30 60 90 120 150 SE +/- 1.65, N = 3 138 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 1 - Mode: Read Only - Average Latency OpenBenchmarking.org ms, Fewer Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 1 - Mode: Read Only - Average Latency Kernel LLVM 0.0092 0.0184 0.0276 0.0368 0.046 SE +/- 0.000, N = 3 0.041 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
PostgreSQL pgbench Scaling Factor: 1 - Clients: 1 - Mode: Read Only OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 13.0 Scaling Factor: 1 - Clients: 1 - Mode: Read Only Kernel LLVM 5K 10K 15K 20K 25K SE +/- 38.99, N = 3 24331 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O2 -lpgcommon -lpgport -lpq -lpthread -lrt -ldl -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 64 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 64 Kernel LLVM Kernel GCC Kernel LLVM 11.1 15 30 45 60 75 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 67.1 67.1 67.1 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 32 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 32 Kernel LLVM Kernel GCC Kernel LLVM 11.1 8 16 24 32 40 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 33.6 33.6 33.6 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 1 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - Parallel: 1 Kernel LLVM Kernel GCC Kernel LLVM 11.1 0.2363 0.4726 0.7089 0.9452 1.1815 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 1.05 1.05 1.05 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 64 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 64 Kernel LLVM Kernel GCC Kernel LLVM 11.1 1400 2800 4200 5600 7000 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 6400 6400 6400 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 32 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 32 Kernel LLVM Kernel GCC Kernel LLVM 11.1 700 1400 2100 2800 3500 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 3200 3200 3200 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 1 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 1000Mbit Objective - Parallel: 1 Kernel LLVM Kernel GCC Kernel LLVM 11.1 200 400 600 800 1000 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 1000 1000 1000 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 1 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: UDP - 100Mbit Objective - Parallel: 1 Kernel LLVM Kernel GCC Kernel LLVM 11.1 20 40 60 80 100 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 100 100 100 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
Schbench Message Threads: 8 - Workers Per Message Thread: 4 OpenBenchmarking.org usec, 99.9th Latency Percentile, Fewer Is Better Schbench Message Threads: 8 - Workers Per Message Thread: 4 Kernel LLVM Kernel GCC Kernel LLVM 11.1 7K 14K 21K 28K 35K SE +/- 2786.67, N = 15 SE +/- 267.31, N = 3 SE +/- 216.51, N = 3 33519 28597 28747 1. (CC) gcc options: -O2 -lpthread
Facebook RocksDB Test: Read While Writing OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Read While Writing Kernel LLVM Kernel GCC Kernel LLVM 11.1 50K 100K 150K 200K 250K SE +/- 5812.42, N = 14 SE +/- 772.88, N = 3 SE +/- 5773.77, N = 14 244439 239799 247556 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
Stress-NG Test: CPU Cache OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: CPU Cache Kernel LLVM Kernel GCC Kernel LLVM 11.1 2 4 6 8 10 SE +/- 0.06, N = 15 SE +/- 0.12, N = 15 SE +/- 0.11, N = 15 6.84 6.90 6.69 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Stress-NG Test: MMAP OpenBenchmarking.org Bogo Ops/s, More Is Better Stress-NG 0.11.07 Test: MMAP Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.25, N = 12 SE +/- 0.22, N = 15 SE +/- 0.19, N = 15 12.45 13.13 12.71 1. (CC) gcc options: -O2 -std=gnu99 -lm -laio -lbsd -lcrypt -lrt -lz -ldl -lpthread -lc
Hackbench Count: 32 - Type: Process OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 32 - Type: Process Kernel LLVM Kernel GCC Kernel LLVM 11.1 80 160 240 320 400 SE +/- 6.99, N = 9 SE +/- 6.91, N = 8 SE +/- 9.64, N = 9 369.55 317.41 349.94 1. (CC) gcc options: -lpthread
Hackbench Count: 1 - Type: Process OpenBenchmarking.org Seconds, Fewer Is Better Hackbench Count: 1 - Type: Process Kernel LLVM Kernel GCC Kernel LLVM 11.1 3 6 9 12 15 SE +/- 0.03, N = 3 SE +/- 0.23, N = 12 SE +/- 0.03, N = 3 10.39 10.08 11.46 1. (CC) gcc options: -lpthread
OSBench Test: Create Threads OpenBenchmarking.org us Per Event, Fewer Is Better OSBench Test: Create Threads Kernel LLVM Kernel GCC Kernel LLVM 11.1 4 8 12 16 20 SE +/- 0.68, N = 15 SE +/- 0.70, N = 15 SE +/- 0.64, N = 15 16.55 15.75 16.02 -lm -lm 1. (CC) gcc options:
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 64 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 64 Kernel LLVM Kernel GCC Kernel LLVM 11.1 4K 8K 12K 16K 20K SE +/- 203.43, N = 4 SE +/- 78.33, N = 3 SE +/- 425.25, N = 15 18133 18304 16653 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
iPerf Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 32 OpenBenchmarking.org Mbits/sec, More Is Better iPerf 3.7 Server Address: localhost - Server Port: 5201 - Duration: 10 Seconds - Test: TCP - Parallel: 32 Kernel LLVM Kernel GCC Kernel LLVM 11.1 4K 8K 12K 16K 20K SE +/- 187.37, N = 7 SE +/- 135.39, N = 3 SE +/- 647.29, N = 15 20013 20700 17735 1. (CC) gcc options: -O3 -march=native -lssl -lcrypto -lm
Phoronix Test Suite v10.8.4