Intel Core i9-10900K compiler benchmarks by Michael Larabel for a future article.
GCC 8.4: -O2 Environment Notes: CXXFLAGS=-O2 CFLAGS=-O2Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 8.4: -O3 -march=native Environment Notes: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 8.4: -O3 -march=native -flto Processor: Intel Core i9-10900K @ 5.30GHz (10 Cores / 20 Threads), Motherboard: Gigabyte Z490 AORUS MASTER (F3 BIOS), Chipset: Intel Comet Lake PCH, Memory: 16GB, Disk: Samsung SSD 970 EVO 250GB, Graphics: Gigabyte AMD Radeon RX 5500/5500M / Pro 5500M 8GB (1890/875MHz), Audio: Realtek ALC1220, Monitor: DELL P2415Q, Network: Intel Device 15f3 + Intel Wi-Fi 6 AX201
OS: Ubuntu 20.04, Kernel: 5.8.0-050800daily20200707-generic (x86_64), Desktop: GNOME Shell 3.36.3, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-ef67218 2020-07-07 focal-oibaf-ppa) (LLVM 10.0.1), OpenCL: OpenCL 1.1 Mesa 20.2.0-devel (git-c977567db6), Vulkan: 1.2.131, Compiler: GCC 8.4.0, File-System: ext4, Screen Resolution: 3840x2160
Environment Notes: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 9.3: -O2 Environment Notes: CXXFLAGS=-O2 CFLAGS=-O2Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 9.3: -O3 -march=native Environment Notes: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 9.3: -O3 -march=native -flto OS: Ubuntu 20.04, Kernel: 5.8.0-050800daily20200707-generic (x86_64), Desktop: GNOME Shell 3.36.3, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-ef67218 2020-07-07 focal-oibaf-ppa) (LLVM 10.0.1), OpenCL: OpenCL 1.1 Mesa 20.2.0-devel (git-c977567db6), Vulkan: 1.2.131, Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 3840x2160
Environment Notes: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 10.2: -O2 Environment Notes: CXXFLAGS=-O2 CFLAGS=-O2Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 10.2: -O3 -march=native Environment Notes: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
GCC 10.2: -O3 -march=native -flto OS: Ubuntu 20.04, Kernel: 5.8.0-050800daily20200707-generic (x86_64), Desktop: GNOME Shell 3.36.3, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-ef67218 2020-07-07 focal-oibaf-ppa) (LLVM 10.0.1), OpenCL: OpenCL 1.1 Mesa 20.2.0-devel (git-c977567db6), Vulkan: 1.2.131, Compiler: GCC 10.1.1 20200715, File-System: ext4, Screen Resolution: 3840x2160
Environment Notes: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
MariaDB This is a MariaDB MySQL database server benchmark making use of mysqlslap. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 10.5.2 Clients: 16 -O2 -O3 -march=native -O3 -march=native -flto 60 120 180 240 300 SE +/- 2.46, N = 3 SE +/- 1.64, N = 3 SE +/- 2.97, N = 3 SE +/- 1.50, N = 3 SE +/- 1.65, N = 3 SE +/- 2.98, N = 3 SE +/- 2.64, N = 3 SE +/- 2.82, N = 6 SE +/- 2.54, N = 3 242 240 242 237 249 244 259 237 237 1. (CXX) g++ options: -O2 -pie -fPIC -fstack-protector -lpthread -llzma -lbz2 -lsnappy -laio -lnuma -lpcre2-8 -lcrypt -lz -lm -lssl -lcrypto -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 10.5.2 Clients: 16 -O2 -O3 -march=native -O3 -march=native -flto 50 100 150 200 250 Min: 236.8 / Avg: 241.56 / Max: 245.04 Min: 237.02 / Avg: 240.29 / Max: 242.01 Min: 237.53 / Avg: 242.03 / Max: 247.65 Min: 235.85 / Avg: 237.44 / Max: 240.44 Min: 246.06 / Avg: 249.29 / Max: 251.51 Min: 238.21 / Avg: 244.08 / Max: 247.89 Min: 254.45 / Avg: 258.59 / Max: 263.5 Min: 224.47 / Avg: 237.48 / Max: 242.78 Min: 233.43 / Avg: 237.19 / Max: 242.01 1. (CXX) g++ options: -O2 -pie -fPIC -fstack-protector -lpthread -llzma -lbz2 -lsnappy -laio -lnuma -lpcre2-8 -lcrypt -lz -lm -lssl -lcrypto -ldl
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 10.5.2 Clients: 1 -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 SE +/- 18.38, N = 9 SE +/- 24.51, N = 9 SE +/- 18.27, N = 9 SE +/- 16.66, N = 9 SE +/- 31.77, N = 9 SE +/- 20.37, N = 9 SE +/- 15.10, N = 9 SE +/- 16.10, N = 9 SE +/- 16.40, N = 9 706 699 730 699 714 704 860 709 709 1. (CXX) g++ options: -O2 -pie -fPIC -fstack-protector -lpthread -llzma -lbz2 -lsnappy -laio -lnuma -lpcre2-8 -lcrypt -lz -lm -lssl -lcrypto -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Queries Per Second, More Is Better MariaDB 10.5.2 Clients: 1 -O2 -O3 -march=native -O3 -march=native -flto 150 300 450 600 750 Min: 590.32 / Avg: 705.56 / Max: 754.72 Min: 535.62 / Avg: 698.55 / Max: 764.53 Min: 598.8 / Avg: 730.47 / Max: 803.21 Min: 573.39 / Avg: 699.26 / Max: 733.68 Min: 532.77 / Avg: 714.01 / Max: 773.4 Min: 583.43 / Avg: 703.89 / Max: 757 Min: 742.94 / Avg: 859.69 / Max: 898.47 Min: 593.47 / Avg: 709.11 / Max: 762.2 Min: 586.51 / Avg: 708.53 / Max: 749.63 1. (CXX) g++ options: -O2 -pie -fPIC -fstack-protector -lpthread -llzma -lbz2 -lsnappy -laio -lnuma -lpcre2-8 -lcrypt -lz -lm -lssl -lcrypto -ldl
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Random Numbers -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 SE +/- 0.32, N = 3 SE +/- 0.32, N = 3 SE +/- 1.08, N = 3 SE +/- 0.27, N = 3 SE +/- 0.06, N = 3 SE +/- 6.69, N = 3 SE +/- 8.79, N = 3 SE +/- 0.05, N = 3 SE +/- 0.15, N = 3 917.06 907.93 876.85 891.75 886.82 871.57 1039.89 852.98 887.49 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Random Numbers -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 Min: 916.73 / Avg: 917.06 / Max: 917.69 Min: 907.43 / Avg: 907.93 / Max: 908.52 Min: 875.33 / Avg: 876.85 / Max: 878.94 Min: 891.42 / Avg: 891.75 / Max: 892.28 Min: 886.75 / Avg: 886.82 / Max: 886.95 Min: 864.81 / Avg: 871.57 / Max: 884.95 Min: 1022.34 / Avg: 1039.89 / Max: 1049.49 Min: 852.88 / Avg: 852.98 / Max: 853.07 Min: 887.19 / Avg: 887.49 / Max: 887.7 1. (CXX) g++ options: -std=c++11 -O3
Crypto++ Crypto++ is a C++ class library of cryptographic algorithms. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: All Algorithms -O2 -O3 -march=native -O3 -march=native -flto 500 1000 1500 2000 2500 SE +/- 0.72, N = 3 SE +/- 0.44, N = 3 SE +/- 1.23, N = 3 SE +/- 0.42, N = 3 SE +/- 0.78, N = 3 SE +/- 0.32, N = 3 SE +/- 0.70, N = 3 SE +/- 0.91, N = 3 SE +/- 0.49, N = 3 1980.05 2105.42 1991.14 1985.80 2042.72 2012.97 1932.89 2042.32 2009.83 1. (CXX) g++ options: -fPIC -pthread -pipe Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: All Algorithms -O2 -O3 -march=native -O3 -march=native -flto 400 800 1200 1600 2000 Min: 1979.15 / Avg: 1980.05 / Max: 1981.48 Min: 2104.55 / Avg: 2105.42 / Max: 2105.95 Min: 1988.73 / Avg: 1991.14 / Max: 1992.76 Min: 1984.96 / Avg: 1985.8 / Max: 1986.27 Min: 2041.55 / Avg: 2042.72 / Max: 2044.21 Min: 2012.33 / Avg: 2012.97 / Max: 2013.35 Min: 1931.49 / Avg: 1932.89 / Max: 1933.71 Min: 2041.07 / Avg: 2042.32 / Max: 2044.1 Min: 2009.07 / Avg: 2009.83 / Max: 2010.76 1. (CXX) g++ options: -fPIC -pthread -pipe
Basis Universal Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 2 + RDO Post-Processing -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 SE +/- 0.77, N = 3 SE +/- 0.00, N = 3 SE +/- 0.11, N = 3 SE +/- 0.11, N = 3 SE +/- 0.37, N = 3 SE +/- 0.22, N = 3 SE +/- 15.61, N = 3 SE +/- 0.51, N = 3 SE +/- 0.37, N = 3 617.61 624.21 616.65 575.76 575.82 565.98 1022.70 575.75 574.23 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 2 + RDO Post-Processing -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 Min: 616.08 / Avg: 617.61 / Max: 618.46 Min: 624.21 / Avg: 624.21 / Max: 624.21 Min: 616.49 / Avg: 616.65 / Max: 616.85 Min: 575.54 / Avg: 575.76 / Max: 575.9 Min: 575.37 / Avg: 575.82 / Max: 576.55 Min: 565.7 / Avg: 565.98 / Max: 566.42 Min: 996.87 / Avg: 1022.7 / Max: 1050.82 Min: 574.76 / Avg: 575.75 / Max: 576.46 Min: 573.61 / Avg: 574.23 / Max: 574.88 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 2D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 6K 12K 18K 24K 30K SE +/- 197.08, N = 3 SE +/- 326.38, N = 9 SE +/- 426.01, N = 3 SE +/- 322.22, N = 9 SE +/- 208.05, N = 3 SE +/- 370.06, N = 3 SE +/- 385.52, N = 4 SE +/- 269.53, N = 3 SE +/- 258.50, N = 9 27739 27241 26506 27536 25475 25068 27311 27832 27004 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 2D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 5K 10K 15K 20K 25K Min: 27372 / Avg: 27739 / Max: 28047 Min: 25200 / Avg: 27240.78 / Max: 28260 Min: 25683 / Avg: 26505.67 / Max: 27109 Min: 26282 / Avg: 27536.11 / Max: 29028 Min: 25064 / Avg: 25474.67 / Max: 25738 Min: 24383 / Avg: 25068.33 / Max: 25653 Min: 26209 / Avg: 27310.75 / Max: 27921 Min: 27336 / Avg: 27831.67 / Max: 28263 Min: 25849 / Avg: 27003.78 / Max: 28297 1. (CC) gcc options: -pthread -lm
Crypto++ Crypto++ is a C++ class library of cryptographic algorithms. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: Keyed Algorithms -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 SE +/- 0.10, N = 3 SE +/- 0.92, N = 3 SE +/- 0.87, N = 3 SE +/- 0.15, N = 3 SE +/- 0.23, N = 3 SE +/- 0.52, N = 3 SE +/- 0.91, N = 3 SE +/- 7.47, N = 3 SE +/- 0.55, N = 3 786.25 799.35 724.68 765.95 781.29 728.08 709.54 787.60 727.18 1. (CXX) g++ options: -fPIC -pthread -pipe Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: Keyed Algorithms -O2 -O3 -march=native -O3 -march=native -flto 140 280 420 560 700 Min: 786.07 / Avg: 786.25 / Max: 786.41 Min: 797.65 / Avg: 799.35 / Max: 800.79 Min: 723 / Avg: 724.68 / Max: 725.87 Min: 765.71 / Avg: 765.95 / Max: 766.21 Min: 780.89 / Avg: 781.29 / Max: 781.67 Min: 727.08 / Avg: 728.08 / Max: 728.83 Min: 707.93 / Avg: 709.54 / Max: 711.09 Min: 776.92 / Avg: 787.6 / Max: 801.99 Min: 726.14 / Avg: 727.18 / Max: 728.01 1. (CXX) g++ options: -fPIC -pthread -pipe
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 30K 60K 90K 120K 150K SE +/- 0.62, N = 3 SE +/- 2.38, N = 3 SE +/- 0.24, N = 3 SE +/- 0.18, N = 3 SE +/- 1.21, N = 3 SE +/- 0.45, N = 3 SE +/- 1832.49, N = 9 SE +/- 0.67, N = 3 SE +/- 0.90, N = 3 235.48 234.17 234.91 231.14 233.88 234.05 124814.00 230.23 233.99 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Recurrent Neural Network Training - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 20K 40K 60K 80K 100K Min: 234.63 / Avg: 235.48 / Max: 236.68 Min: 229.44 / Avg: 234.17 / Max: 237 Min: 234.49 / Avg: 234.91 / Max: 235.32 Min: 230.79 / Avg: 231.14 / Max: 231.4 Min: 231.76 / Avg: 233.88 / Max: 235.95 Min: 233.16 / Avg: 234.05 / Max: 234.59 Min: 116092 / Avg: 124814.44 / Max: 130651 Min: 229.32 / Avg: 230.23 / Max: 231.55 Min: 232.42 / Avg: 233.99 / Max: 235.53 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill -O2 -O3 -march=native -O3 -march=native -flto 200K 400K 600K 800K 1000K SE +/- 19698.59, N = 15 SE +/- 19115.99, N = 12 SE +/- 24759.46, N = 15 SE +/- 26818.73, N = 15 SE +/- 10300.40, N = 15 SE +/- 31100.37, N = 15 SE +/- 3996.19, N = 3 SE +/- 27169.86, N = 15 SE +/- 20950.15, N = 15 705993 691247 701478 689652 899209 738257 275752 675757 663613 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill -O2 -O3 -march=native -O3 -march=native -flto 160K 320K 480K 640K 800K Min: 553143 / Avg: 705992.87 / Max: 825973 Min: 553160 / Avg: 691247.42 / Max: 821596 Min: 527925 / Avg: 701478.27 / Max: 830283 Min: 508251 / Avg: 689651.53 / Max: 836500 Min: 836189 / Avg: 899208.67 / Max: 974885 Min: 527006 / Avg: 738256.6 / Max: 946112 Min: 268362 / Avg: 275752.33 / Max: 282083 Min: 518780 / Avg: 675756.67 / Max: 840679 Min: 509561 / Avg: 663612.53 / Max: 767901 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Math Library -O2 -O3 -march=native -O3 -march=native -flto 90 180 270 360 450 SE +/- 0.71, N = 3 SE +/- 0.43, N = 3 SE +/- 1.11, N = 3 SE +/- 0.63, N = 3 SE +/- 0.40, N = 3 SE +/- 0.20, N = 3 SE +/- 0.09, N = 3 SE +/- 0.39, N = 3 SE +/- 0.14, N = 3 255.38 253.70 248.01 251.89 248.25 240.61 415.04 232.83 239.73 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Math Library -O2 -O3 -march=native -O3 -march=native -flto 70 140 210 280 350 Min: 253.97 / Avg: 255.38 / Max: 256.14 Min: 253.01 / Avg: 253.7 / Max: 254.48 Min: 246.41 / Avg: 248.01 / Max: 250.15 Min: 251.17 / Avg: 251.89 / Max: 253.14 Min: 247.74 / Avg: 248.25 / Max: 249.05 Min: 240.32 / Avg: 240.61 / Max: 241 Min: 414.93 / Avg: 415.04 / Max: 415.21 Min: 232.26 / Avg: 232.83 / Max: 233.58 Min: 239.59 / Avg: 239.73 / Max: 240 1. (CXX) g++ options: -std=c++11 -O3
Crypto++ Crypto++ is a C++ class library of cryptographic algorithms. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: Integer + Elliptic Curve Public Key Algorithms -O2 -O3 -march=native -O3 -march=native -flto 1400 2800 4200 5600 7000 SE +/- 55.58, N = 3 SE +/- 2.38, N = 3 SE +/- 2.47, N = 3 SE +/- 67.06, N = 3 SE +/- 2.07, N = 3 SE +/- 8.70, N = 3 SE +/- 1.61, N = 3 SE +/- 4.38, N = 3 SE +/- 5.42, N = 3 6394.23 6606.14 6427.55 6322.53 6392.77 6563.73 6287.20 6591.08 6512.26 1. (CXX) g++ options: -fPIC -pthread -pipe Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: Integer + Elliptic Curve Public Key Algorithms -O2 -O3 -march=native -O3 -march=native -flto 1100 2200 3300 4400 5500 Min: 6283.29 / Avg: 6394.23 / Max: 6455.87 Min: 6602.76 / Avg: 6606.14 / Max: 6610.73 Min: 6422.62 / Avg: 6427.55 / Max: 6430.34 Min: 6245.49 / Avg: 6322.53 / Max: 6456.12 Min: 6388.75 / Avg: 6392.77 / Max: 6395.64 Min: 6552.69 / Avg: 6563.73 / Max: 6580.9 Min: 6284.06 / Avg: 6287.2 / Max: 6289.41 Min: 6582.72 / Avg: 6591.08 / Max: 6597.54 Min: 6502.25 / Avg: 6512.26 / Max: 6520.86 1. (CXX) g++ options: -fPIC -pthread -pipe
PostgreSQL pgbench This is a simple benchmark of PostgreSQL using pgbench. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 12.0 Scaling: Buffer Test - Test: Normal Load - Mode: Read Write -O2 -O3 -march=native -O3 -march=native -flto 1300 2600 3900 5200 6500 SE +/- 71.87, N = 5 SE +/- 54.83, N = 10 SE +/- 61.79, N = 8 SE +/- 77.49, N = 5 SE +/- 53.86, N = 11 SE +/- 53.90, N = 10 SE +/- 47.10, N = 14 SE +/- 52.02, N = 11 SE +/- 59.32, N = 15 5789.05 5858.13 5842.45 5796.13 6176.79 5929.80 6016.35 5939.67 5821.51 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 12.0 Scaling: Buffer Test - Test: Normal Load - Mode: Read Write -O2 -O3 -march=native -O3 -march=native -flto 1100 2200 3300 4400 5500 Min: 5508.59 / Avg: 5789.05 / Max: 5895.87 Min: 5467.87 / Avg: 5858.13 / Max: 6025.35 Min: 5532.94 / Avg: 5842.45 / Max: 6014.39 Min: 5501.21 / Avg: 5796.13 / Max: 5947.9 Min: 5672.78 / Avg: 6176.79 / Max: 6360.07 Min: 5478.27 / Avg: 5929.8 / Max: 6076.75 Min: 5609.3 / Avg: 6016.35 / Max: 6249.62 Min: 5486.05 / Avg: 5939.67 / Max: 6110.49 Min: 5153.24 / Avg: 5821.51 / Max: 6039.14 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm
GROMACS The GROMACS (GROningen MAchine for Chemical Simulations) molecular dynamics package testing on the CPU with the water_GMX50 data. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2020.1 Water Benchmark -O2 -O3 -march=native -O3 -march=native -flto 0.1967 0.3934 0.5901 0.7868 0.9835 SE +/- 0.000, N = 3 SE +/- 0.001, N = 3 SE +/- 0.000, N = 3 SE +/- 0.001, N = 3 SE +/- 0.001, N = 3 SE +/- 0.001, N = 3 SE +/- 0.004, N = 3 SE +/- 0.001, N = 3 SE +/- 0.001, N = 3 0.874 0.870 0.874 0.866 0.866 0.869 0.576 0.870 0.874 1. (CXX) g++ options: -O3 -pthread -lrt -lpthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Ns Per Day, More Is Better GROMACS 2020.1 Water Benchmark -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 Min: 0.87 / Avg: 0.87 / Max: 0.88 Min: 0.87 / Avg: 0.87 / Max: 0.87 Min: 0.87 / Avg: 0.87 / Max: 0.88 Min: 0.87 / Avg: 0.87 / Max: 0.87 Min: 0.86 / Avg: 0.87 / Max: 0.87 Min: 0.87 / Avg: 0.87 / Max: 0.87 Min: 0.57 / Avg: 0.58 / Max: 0.58 Min: 0.87 / Avg: 0.87 / Max: 0.87 Min: 0.87 / Avg: 0.87 / Max: 0.88 1. (CXX) g++ options: -O3 -pthread -lrt -lpthread -lm
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 2D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 2K 4K 6K 8K 10K SE +/- 50.26, N = 3 SE +/- 96.60, N = 12 SE +/- 22.37, N = 3 SE +/- 25.15, N = 3 SE +/- 10.63, N = 3 SE +/- 47.08, N = 3 SE +/- 32.19, N = 3 SE +/- 79.24, N = 3 SE +/- 14.68, N = 3 6792.1 8325.7 8049.2 7068.9 7542.0 8064.3 6837.6 7661.9 8318.2 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 2D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 1400 2800 4200 5600 7000 Min: 6722.5 / Avg: 6792.1 / Max: 6889.7 Min: 7817.1 / Avg: 8325.74 / Max: 8720.2 Min: 8009 / Avg: 8049.23 / Max: 8086.3 Min: 7026.1 / Avg: 7068.9 / Max: 7113.2 Min: 7528.5 / Avg: 7542.03 / Max: 7563 Min: 7987.4 / Avg: 8064.27 / Max: 8149.8 Min: 6775.8 / Avg: 6837.6 / Max: 6884.1 Min: 7510.1 / Avg: 7661.93 / Max: 7777.2 Min: 8288.9 / Avg: 8318.2 / Max: 8334.4 1. (CC) gcc options: -pthread -lm
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 11K 22K 33K 44K 55K SE +/- 0.12, N = 3 SE +/- 0.28, N = 3 SE +/- 0.25, N = 3 SE +/- 0.82, N = 3 SE +/- 0.10, N = 3 SE +/- 0.80, N = 3 SE +/- 2670.40, N = 9 SE +/- 0.53, N = 3 SE +/- 0.35, N = 3 104.12 103.18 104.69 105.20 104.00 104.51 51173.40 103.39 103.79 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Recurrent Neural Network Inference - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 9K 18K 27K 36K 45K Min: 103.89 / Avg: 104.12 / Max: 104.32 Min: 102.75 / Avg: 103.18 / Max: 103.7 Min: 104.37 / Avg: 104.69 / Max: 105.17 Min: 103.72 / Avg: 105.2 / Max: 106.56 Min: 103.83 / Avg: 104 / Max: 104.16 Min: 103.37 / Avg: 104.51 / Max: 106.06 Min: 33946.8 / Avg: 51173.39 / Max: 61394.5 Min: 102.44 / Avg: 103.39 / Max: 104.28 Min: 103.39 / Avg: 103.79 / Max: 104.48 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Read While Writing -O2 -O3 -march=native -O3 -march=native -flto 600K 1200K 1800K 2400K 3000K SE +/- 15080.50, N = 3 SE +/- 13502.20, N = 3 SE +/- 34614.47, N = 3 SE +/- 6775.07, N = 3 SE +/- 51165.38, N = 3 SE +/- 23831.88, N = 12 SE +/- 34693.17, N = 15 SE +/- 47176.39, N = 3 SE +/- 41402.09, N = 4 2749381 2703851 2722395 2718543 2968384 2780546 1602068 2867917 2829235 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Read While Writing -O2 -O3 -march=native -O3 -march=native -flto 500K 1000K 1500K 2000K 2500K Min: 2728091 / Avg: 2749381.33 / Max: 2778528 Min: 2679865 / Avg: 2703850.67 / Max: 2726588 Min: 2683497 / Avg: 2722394.67 / Max: 2791439 Min: 2705478 / Avg: 2718542.67 / Max: 2728188 Min: 2869270 / Avg: 2968383.67 / Max: 3039987 Min: 2649480 / Avg: 2780546.25 / Max: 2966525 Min: 1442397 / Avg: 1602068.33 / Max: 1800802 Min: 2774606 / Avg: 2867916.67 / Max: 2926683 Min: 2770806 / Avg: 2829235 / Max: 2951533 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
Minion Minion is an open-source constraint solver that is designed to be very scalable. This test profile uses Minion's integrated benchmarking problems to solve. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Quasigroup -O2 -O3 -march=native -O3 -march=native -flto 30 60 90 120 150 SE +/- 0.22, N = 3 SE +/- 0.56, N = 3 SE +/- 0.54, N = 3 SE +/- 0.14, N = 3 SE +/- 0.67, N = 3 SE +/- 0.11, N = 3 SE +/- 0.25, N = 3 SE +/- 0.78, N = 3 SE +/- 0.24, N = 3 95.20 93.95 95.35 94.79 92.86 95.42 130.00 89.90 94.26 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Quasigroup -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 Min: 94.76 / Avg: 95.2 / Max: 95.44 Min: 93.37 / Avg: 93.95 / Max: 95.07 Min: 94.29 / Avg: 95.35 / Max: 96.06 Min: 94.59 / Avg: 94.79 / Max: 95.06 Min: 91.98 / Avg: 92.86 / Max: 94.18 Min: 95.21 / Avg: 95.42 / Max: 95.59 Min: 129.64 / Avg: 130 / Max: 130.49 Min: 89.03 / Avg: 89.9 / Max: 91.45 Min: 93.89 / Avg: 94.26 / Max: 94.71 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic
Darmstadt Automotive Parallel Heterogeneous Suite DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Test Cases Per Minute, More Is Better Darmstadt Automotive Parallel Heterogeneous Suite Backend: OpenMP - Kernel: Points2Image -O2 -O3 -march=native -O3 -march=native -flto 7K 14K 21K 28K 35K SE +/- 49.76, N = 3 SE +/- 173.90, N = 3 SE +/- 361.69, N = 6 SE +/- 160.88, N = 3 SE +/- 276.23, N = 3 SE +/- 379.31, N = 3 SE +/- 8.04, N = 3 SE +/- 353.00, N = 3 SE +/- 207.45, N = 3 29651.57 31234.32 29679.48 29472.08 30890.21 31156.96 2129.03 29989.20 30726.55 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Test Cases Per Minute, More Is Better Darmstadt Automotive Parallel Heterogeneous Suite Backend: OpenMP - Kernel: Points2Image -O2 -O3 -march=native -O3 -march=native -flto 5K 10K 15K 20K 25K Min: 29570.34 / Avg: 29651.57 / Max: 29741.99 Min: 30889.3 / Avg: 31234.32 / Max: 31444.89 Min: 28624.04 / Avg: 29679.48 / Max: 31068.44 Min: 29163.45 / Avg: 29472.08 / Max: 29705.18 Min: 30430.44 / Avg: 30890.21 / Max: 31385.35 Min: 30417.17 / Avg: 31156.96 / Max: 31672.3 Min: 2113.99 / Avg: 2129.03 / Max: 2141.47 Min: 29299.31 / Avg: 29989.2 / Max: 30463.97 Min: 30353.16 / Avg: 30726.55 / Max: 31069.89 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp
Zstd Compression This test measures the time needed to compress a sample file (an Ubuntu ISO) using Zstd compression. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.5 Compression Level: 19 -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 SE +/- 0.03, N = 3 SE +/- 0.00, N = 3 SE +/- 0.07, N = 3 SE +/- 0.03, N = 3 SE +/- 0.00, N = 3 28.4 28.5 28.5 28.4 28.5 28.5 28.4 28.4 28.5 1. (CC) gcc options: -pthread -lz -llzma Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.5 Compression Level: 19 -O2 -O3 -march=native -O3 -march=native -flto 6 12 18 24 30 Min: 28.4 / Avg: 28.4 / Max: 28.4 Min: 28.5 / Avg: 28.5 / Max: 28.5 Min: 28.5 / Avg: 28.53 / Max: 28.6 Min: 28.3 / Avg: 28.4 / Max: 28.5 Min: 28.4 / Avg: 28.47 / Max: 28.5 Min: 28.5 / Avg: 28.5 / Max: 28.5 Min: 28.3 / Avg: 28.43 / Max: 28.5 Min: 28.4 / Avg: 28.43 / Max: 28.5 Min: 28.5 / Avg: 28.5 / Max: 28.5 1. (CC) gcc options: -pthread -lz -llzma
Timed MrBayes Analysis This test performs a bayesian analysis of a set of primate genome sequences in order to estimate their phylogeny. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Timed MrBayes Analysis 3.2.7 Primate Phylogeny Analysis -O2 -O3 -march=native -O3 -march=native -flto 15 30 45 60 75 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 SE +/- 0.06, N = 3 SE +/- 0.67, N = 9 SE +/- 0.27, N = 3 SE +/- 0.07, N = 3 SE +/- 0.09, N = 3 SE +/- 0.34, N = 3 SE +/- 0.04, N = 3 67.89 63.86 63.72 67.82 63.84 63.16 66.53 64.08 63.42 1. (CC) gcc options: -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -madx -mmpx -mabm -O3 -std=c99 -pedantic -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Timed MrBayes Analysis 3.2.7 Primate Phylogeny Analysis -O2 -O3 -march=native -O3 -march=native -flto 13 26 39 52 65 Min: 67.79 / Avg: 67.89 / Max: 67.97 Min: 63.78 / Avg: 63.86 / Max: 63.93 Min: 63.6 / Avg: 63.72 / Max: 63.82 Min: 66.95 / Avg: 67.82 / Max: 73.2 Min: 63.56 / Avg: 63.84 / Max: 64.39 Min: 63.06 / Avg: 63.16 / Max: 63.3 Min: 66.41 / Avg: 66.53 / Max: 66.71 Min: 63.58 / Avg: 64.08 / Max: 64.74 Min: 63.37 / Avg: 63.42 / Max: 63.5 1. (CC) gcc options: -mmmx -msse -msse2 -msse3 -mssse3 -msse4.1 -msse4.2 -maes -mavx -mfma -mavx2 -mrdrnd -mbmi -mbmi2 -madx -mmpx -mabm -O3 -std=c99 -pedantic -lm
PostgreSQL pgbench This is a simple benchmark of PostgreSQL using pgbench. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 12.0 Scaling: Buffer Test - Test: Normal Load - Mode: Read Only -O2 -O3 -march=native -O3 -march=native -flto 70K 140K 210K 280K 350K SE +/- 100.30, N = 3 SE +/- 252.05, N = 3 SE +/- 162.53, N = 3 SE +/- 87.46, N = 3 SE +/- 123.90, N = 3 SE +/- 304.09, N = 3 SE +/- 1184.68, N = 3 SE +/- 85.04, N = 3 SE +/- 230.78, N = 3 304387.87 306636.40 306838.93 304959.24 305586.13 312444.22 193110.28 305226.21 316873.41 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 12.0 Scaling: Buffer Test - Test: Normal Load - Mode: Read Only -O2 -O3 -march=native -O3 -march=native -flto 50K 100K 150K 200K 250K Min: 304227.84 / Avg: 304387.87 / Max: 304572.63 Min: 306163.76 / Avg: 306636.4 / Max: 307024.52 Min: 306517.01 / Avg: 306838.93 / Max: 307038.89 Min: 304786.96 / Avg: 304959.24 / Max: 305071.61 Min: 305353.71 / Avg: 305586.13 / Max: 305776.78 Min: 311915.24 / Avg: 312444.22 / Max: 312968.61 Min: 191491.97 / Avg: 193110.28 / Max: 195418.17 Min: 305131.31 / Avg: 305226.21 / Max: 305395.88 Min: 316479.71 / Avg: 316873.41 / Max: 317278.89 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Stepanov Vector -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.06, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 71.11 71.34 68.61 71.04 71.20 69.00 88.73 69.31 71.48 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Stepanov Vector -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 Min: 71.09 / Avg: 71.11 / Max: 71.14 Min: 71.32 / Avg: 71.34 / Max: 71.36 Min: 68.54 / Avg: 68.61 / Max: 68.73 Min: 71.03 / Avg: 71.04 / Max: 71.06 Min: 71.19 / Avg: 71.2 / Max: 71.21 Min: 68.99 / Avg: 69 / Max: 69 Min: 88.72 / Avg: 88.73 / Max: 88.74 Min: 69.3 / Avg: 69.31 / Max: 69.31 Min: 71.48 / Avg: 71.48 / Max: 71.49 1. (CXX) g++ options: -std=c++11 -O3
Crypto++ Crypto++ is a C++ class library of cryptographic algorithms. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: Unkeyed Algorithms -O2 -O3 -march=native -O3 -march=native -flto 100 200 300 400 500 SE +/- 0.17, N = 3 SE +/- 0.61, N = 3 SE +/- 0.07, N = 3 SE +/- 0.09, N = 3 SE +/- 0.29, N = 3 SE +/- 0.49, N = 3 SE +/- 0.40, N = 3 SE +/- 0.08, N = 3 SE +/- 0.38, N = 3 407.78 441.88 423.23 397.42 429.18 421.89 393.96 446.32 428.33 1. (CXX) g++ options: -fPIC -pthread -pipe Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/second, More Is Better Crypto++ 8.2 Test: Unkeyed Algorithms -O2 -O3 -march=native -O3 -march=native -flto 80 160 240 320 400 Min: 407.44 / Avg: 407.78 / Max: 407.99 Min: 440.66 / Avg: 441.88 / Max: 442.6 Min: 423.1 / Avg: 423.23 / Max: 423.36 Min: 397.32 / Avg: 397.42 / Max: 397.59 Min: 428.63 / Avg: 429.18 / Max: 429.62 Min: 421.06 / Avg: 421.89 / Max: 422.76 Min: 393.16 / Avg: 393.96 / Max: 394.41 Min: 446.21 / Avg: 446.32 / Max: 446.48 Min: 427.72 / Avg: 428.33 / Max: 429.03 1. (CXX) g++ options: -fPIC -pthread -pipe
Tachyon This is a test of the threaded Tachyon, a parallel ray-tracing system, measuring the time to ray-trace a sample scene. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Tachyon 0.99b6 Total Time -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 SE +/- 0.05, N = 3 SE +/- 0.13, N = 3 SE +/- 0.22, N = 3 SE +/- 0.03, N = 3 SE +/- 0.05, N = 3 SE +/- 0.18, N = 3 SE +/- 0.61, N = 3 SE +/- 0.06, N = 3 SE +/- 0.09, N = 3 65.81 65.85 65.97 65.09 64.91 65.10 104.66 65.36 65.31 1. (CC) gcc options: -m64 -O3 -fomit-frame-pointer -ffast-math -ltachyon -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Tachyon 0.99b6 Total Time -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 Min: 65.7 / Avg: 65.81 / Max: 65.87 Min: 65.67 / Avg: 65.85 / Max: 66.1 Min: 65.54 / Avg: 65.97 / Max: 66.27 Min: 65.05 / Avg: 65.09 / Max: 65.15 Min: 64.85 / Avg: 64.91 / Max: 65.01 Min: 64.75 / Avg: 65.1 / Max: 65.33 Min: 103.71 / Avg: 104.66 / Max: 105.8 Min: 65.28 / Avg: 65.36 / Max: 65.48 Min: 65.2 / Avg: 65.31 / Max: 65.48 1. (CC) gcc options: -m64 -O3 -fomit-frame-pointer -ffast-math -ltachyon -lm -lpthread
Stockfish This is a test of Stockfish, an advanced C++11 chess benchmark that can scale up to 128 CPU cores. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Nodes Per Second, More Is Better Stockfish 9 Total Time -O2 -O3 -march=native -O3 -march=native -flto 8M 16M 24M 32M 40M SE +/- 326209.83, N = 3 SE +/- 212129.33, N = 3 SE +/- 355455.95, N = 3 SE +/- 257012.86, N = 3 SE +/- 196018.24, N = 3 SE +/- 271344.75, N = 3 SE +/- 141218.51, N = 3 SE +/- 169972.97, N = 3 SE +/- 192068.29, N = 3 36485480 36901232 36912963 35734166 36181380 36566683 23128174 36678583 37418623 1. (CXX) g++ options: -m64 -lpthread -fno-exceptions -std=c++11 -pedantic -O3 -msse -msse3 -mpopcnt -flto Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Nodes Per Second, More Is Better Stockfish 9 Total Time -O2 -O3 -march=native -O3 -march=native -flto 6M 12M 18M 24M 30M Min: 36138882 / Avg: 36485480.33 / Max: 37137466 Min: 36549466 / Avg: 36901231.67 / Max: 37282522 Min: 36218685 / Avg: 36912962.67 / Max: 37392505 Min: 35226293 / Avg: 35734165.67 / Max: 36056774 Min: 35844592 / Avg: 36181380.33 / Max: 36523555 Min: 36057749 / Avg: 36566683.33 / Max: 36984317 Min: 22847720 / Avg: 23128174.33 / Max: 23297333 Min: 36387309 / Avg: 36678583 / Max: 36976016 Min: 37058858 / Avg: 37418623 / Max: 37715114 1. (CXX) g++ options: -m64 -lpthread -fno-exceptions -std=c++11 -pedantic -O3 -msse -msse3 -mpopcnt -flto
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill Sync -O2 -O3 -march=native -O3 -march=native -flto 700 1400 2100 2800 3500 SE +/- 2.31, N = 3 SE +/- 30.25, N = 3 SE +/- 14.33, N = 3 SE +/- 7.51, N = 3 SE +/- 1.45, N = 3 SE +/- 25.50, N = 3 SE +/- 9.13, N = 3 SE +/- 47.79, N = 3 SE +/- 21.66, N = 3 2856 2817 2813 2829 2894 2881 3369 2843 2844 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Fill Sync -O2 -O3 -march=native -O3 -march=native -flto 600 1200 1800 2400 3000 Min: 2852 / Avg: 2856 / Max: 2860 Min: 2761 / Avg: 2816.67 / Max: 2865 Min: 2793 / Avg: 2813.33 / Max: 2841 Min: 2814 / Avg: 2829 / Max: 2837 Min: 2891 / Avg: 2893.67 / Max: 2896 Min: 2830 / Avg: 2881 / Max: 2907 Min: 3352 / Avg: 3369.33 / Max: 3383 Min: 2748 / Avg: 2843.33 / Max: 2897 Min: 2807 / Avg: 2844 / Max: 2882 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
John The Ripper This is a benchmark of John The Ripper, which is a password cracker. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Real C/S, More Is Better John The Ripper 1.9.0-jumbo-1 Test: MD5 -O2 -O3 -march=native -O3 -march=native -flto 300K 600K 900K 1200K 1500K SE +/- 2185.81, N = 3 SE +/- 666.67, N = 3 SE +/- 6173.42, N = 3 SE +/- 3179.80, N = 3 SE +/- 666.67, N = 3 SE +/- 2645.75, N = 3 SE +/- 924.38, N = 4 SE +/- 333.33, N = 3 SE +/- 6359.59, N = 3 1387667 1377667 1383667 1386333 1385667 1372000 70955 1376333 1449667 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -pthread -lm -lz -ldl -lcrypt -lbz2 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Real C/S, More Is Better John The Ripper 1.9.0-jumbo-1 Test: MD5 -O2 -O3 -march=native -O3 -march=native -flto 300K 600K 900K 1200K 1500K Min: 1385000 / Avg: 1387666.67 / Max: 1392000 Min: 1377000 / Avg: 1377666.67 / Max: 1379000 Min: 1372000 / Avg: 1383666.67 / Max: 1393000 Min: 1380000 / Avg: 1386333.33 / Max: 1390000 Min: 1385000 / Avg: 1385666.67 / Max: 1387000 Min: 1367000 / Avg: 1372000 / Max: 1376000 Min: 68219 / Avg: 70954.5 / Max: 72247 Min: 1376000 / Avg: 1376333.33 / Max: 1377000 Min: 1439000 / Avg: 1449666.67 / Max: 1461000 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -pthread -lm -lz -ldl -lcrypt -lbz2
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.33 Operation: Enhanced -O2 -O3 -march=native -O3 -march=native -flto 70 140 210 280 350 267 332 332 267 333 333 141 302 302 1. (CC) gcc options: -fopenmp -pthread -ljbig -lwebp -lwebpmux -ltiff -lfreetype -ljpeg -lXext -lSM -lICE -lX11 -llzma -lbz2 -lxml2 -lz -lzstd -lm -lpthread
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Read -O2 -O3 -march=native -O3 -march=native -flto 16M 32M 48M 64M 80M SE +/- 8446.73, N = 3 SE +/- 192421.03, N = 3 SE +/- 56983.14, N = 3 SE +/- 224015.64, N = 3 SE +/- 71110.62, N = 3 SE +/- 37867.80, N = 3 SE +/- 159320.31, N = 3 SE +/- 83522.36, N = 3 SE +/- 73778.61, N = 3 72161257 71908282 72072666 76916855 76904306 76925401 47341353 75628129 75847196 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Random Read -O2 -O3 -march=native -O3 -march=native -flto 13M 26M 39M 52M 65M Min: 72151859 / Avg: 72161256.67 / Max: 72178113 Min: 71535170 / Avg: 71908282.33 / Max: 72176496 Min: 72009045 / Avg: 72072666.33 / Max: 72186364 Min: 76524153 / Avg: 76916854.67 / Max: 77299989 Min: 76765465 / Avg: 76904306 / Max: 77000420 Min: 76849831 / Avg: 76925401 / Max: 76967521 Min: 47125555 / Avg: 47341353.33 / Max: 47652285 Min: 75514294 / Avg: 75628129.33 / Max: 75790920 Min: 75730107 / Avg: 75847196 / Max: 75983506 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
Minion Minion is an open-source constraint solver that is designed to be very scalable. This test profile uses Minion's integrated benchmarking problems to solve. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Solitaire -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 SE +/- 0.17, N = 3 SE +/- 0.16, N = 3 SE +/- 0.05, N = 3 SE +/- 0.36, N = 3 SE +/- 0.02, N = 3 SE +/- 0.06, N = 3 SE +/- 0.13, N = 3 SE +/- 0.08, N = 3 SE +/- 0.14, N = 3 59.09 57.04 58.07 58.52 55.53 57.75 79.48 54.88 58.96 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Solitaire -O2 -O3 -march=native -O3 -march=native -flto 15 30 45 60 75 Min: 58.8 / Avg: 59.09 / Max: 59.4 Min: 56.74 / Avg: 57.04 / Max: 57.31 Min: 58.01 / Avg: 58.07 / Max: 58.17 Min: 57.81 / Avg: 58.52 / Max: 58.91 Min: 55.5 / Avg: 55.53 / Max: 55.56 Min: 57.65 / Avg: 57.75 / Max: 57.85 Min: 79.24 / Avg: 79.48 / Max: 79.66 Min: 54.75 / Avg: 54.88 / Max: 55.03 Min: 58.7 / Avg: 58.96 / Max: 59.2 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic
C-Ray This is a test of C-Ray, a simple raytracer designed to test the floating-point CPU performance. This test is multi-threaded (16 threads per core), will shoot 8 rays per pixel for anti-aliasing, and will generate a 1600 x 1200 image. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 1.1 Total Time - 4K, 16 Rays Per Pixel -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 3 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 96.90 38.00 48.26 96.91 38.47 38.22 101.33 38.13 38.56 1. (CC) gcc options: -lm -lpthread -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 1.1 Total Time - 4K, 16 Rays Per Pixel -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 Min: 96.88 / Avg: 96.9 / Max: 96.92 Min: 37.98 / Avg: 38 / Max: 38.02 Min: 48.22 / Avg: 48.26 / Max: 48.33 Min: 96.83 / Avg: 96.91 / Max: 96.99 Min: 38.44 / Avg: 38.47 / Max: 38.49 Min: 38.2 / Avg: 38.22 / Max: 38.25 Min: 101.25 / Avg: 101.33 / Max: 101.42 Min: 38.09 / Avg: 38.13 / Max: 38.15 Min: 38.55 / Avg: 38.56 / Max: 38.57 1. (CC) gcc options: -lm -lpthread -O3
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: IP Batch All - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 SE +/- 0.04, N = 3 SE +/- 0.10, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.09, N = 3 SE +/- 0.03, N = 3 SE +/- 1.00, N = 3 SE +/- 0.06, N = 3 SE +/- 0.10, N = 3 59.54 59.51 59.37 59.42 59.62 59.57 1076.94 59.65 59.56 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: IP Batch All - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 Min: 59.47 / Avg: 59.54 / Max: 59.61 Min: 59.3 / Avg: 59.51 / Max: 59.62 Min: 59.24 / Avg: 59.37 / Max: 59.48 Min: 59.32 / Avg: 59.42 / Max: 59.49 Min: 59.46 / Avg: 59.62 / Max: 59.77 Min: 59.51 / Avg: 59.57 / Max: 59.61 Min: 1075.04 / Avg: 1076.94 / Max: 1078.45 Min: 59.59 / Avg: 59.65 / Max: 59.76 Min: 59.39 / Avg: 59.56 / Max: 59.75 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
NGINX Benchmark This is a test of ab, which is the Apache Benchmark program running against nginx. This test profile measures how many requests per second a given system can sustain when carrying out 2,000,000 requests with 500 requests being carried out concurrently. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Requests Per Second, More Is Better NGINX Benchmark 1.9.9 Static Web Page Serving -O2 -O3 -march=native -O3 -march=native -flto 11K 22K 33K 44K 55K SE +/- 537.57, N = 3 SE +/- 253.18, N = 3 SE +/- 640.07, N = 3 SE +/- 444.33, N = 10 SE +/- 579.81, N = 3 SE +/- 778.35, N = 3 SE +/- 337.59, N = 3 SE +/- 405.62, N = 3 SE +/- 85.52, N = 3 45811.09 45119.62 49753.25 47213.65 48661.14 50193.24 36894.76 45061.65 49746.88 1. (CC) gcc options: -lpthread -lcrypt -lcrypto -lz -O3 -march=native Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Requests Per Second, More Is Better NGINX Benchmark 1.9.9 Static Web Page Serving -O2 -O3 -march=native -O3 -march=native -flto 9K 18K 27K 36K 45K Min: 44866.06 / Avg: 45811.09 / Max: 46727.62 Min: 44613.93 / Avg: 45119.62 / Max: 45394.98 Min: 48473.2 / Avg: 49753.25 / Max: 50406.3 Min: 45074.64 / Avg: 47213.65 / Max: 49258.28 Min: 47501.73 / Avg: 48661.14 / Max: 49260.13 Min: 48645.77 / Avg: 50193.24 / Max: 51113.51 Min: 36277.14 / Avg: 36894.76 / Max: 37439.81 Min: 44250.71 / Avg: 45061.65 / Max: 45486.04 Min: 49575.92 / Avg: 49746.88 / Max: 49836.84 1. (CC) gcc options: -lpthread -lcrypt -lcrypto -lz -O3 -march=native
Himeno Benchmark The Himeno benchmark is a linear solver of pressure Poisson using a point-Jacobi method. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MFLOPS, More Is Better Himeno Benchmark 3.0 Poisson Pressure Solver -O2 -O3 -march=native -O3 -march=native -flto 1100 2200 3300 4400 5500 SE +/- 0.54, N = 3 SE +/- 4.01, N = 3 SE +/- 0.73, N = 3 SE +/- 0.50, N = 3 SE +/- 6.67, N = 3 SE +/- 4.25, N = 3 SE +/- 1.05, N = 3 SE +/- 6.55, N = 3 SE +/- 2.10, N = 3 4363.81 4827.88 5082.93 4377.53 4990.49 5115.56 2681.55 4998.18 5193.79 1. (CC) gcc options: -O3 -mavx2 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MFLOPS, More Is Better Himeno Benchmark 3.0 Poisson Pressure Solver -O2 -O3 -march=native -O3 -march=native -flto 900 1800 2700 3600 4500 Min: 4363.05 / Avg: 4363.81 / Max: 4364.86 Min: 4820.16 / Avg: 4827.88 / Max: 4833.64 Min: 5081.79 / Avg: 5082.93 / Max: 5084.28 Min: 4376.58 / Avg: 4377.53 / Max: 4378.26 Min: 4977.16 / Avg: 4990.49 / Max: 4997.41 Min: 5107.68 / Avg: 5115.56 / Max: 5122.28 Min: 2679.87 / Avg: 2681.55 / Max: 2683.48 Min: 4985.19 / Avg: 4998.18 / Max: 5006.14 Min: 5190.07 / Avg: 5193.79 / Max: 5197.32 1. (CC) gcc options: -O3 -mavx2
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Atol -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.14, N = 3 SE +/- 0.11, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 48.45 48.54 46.97 48.55 48.54 48.60 98.74 46.89 48.43 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Atol -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 Min: 48.43 / Avg: 48.45 / Max: 48.5 Min: 48.47 / Avg: 48.54 / Max: 48.62 Min: 46.96 / Avg: 46.97 / Max: 46.97 Min: 48.54 / Avg: 48.55 / Max: 48.57 Min: 48.49 / Avg: 48.54 / Max: 48.59 Min: 48.43 / Avg: 48.6 / Max: 48.88 Min: 98.52 / Avg: 98.74 / Max: 98.91 Min: 46.89 / Avg: 46.89 / Max: 46.9 Min: 48.42 / Avg: 48.43 / Max: 48.44 1. (CXX) g++ options: -std=c++11 -O3
SQLite Speedtest This is a benchmark of SQLite's speedtest1 benchmark program with an increased problem size of 1,000. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better SQLite Speedtest 3.30 Timed Time - Size 1,000 -O2 -O3 -march=native -O3 -march=native -flto 16 32 48 64 80 SE +/- 0.27, N = 3 SE +/- 0.04, N = 3 SE +/- 0.68, N = 3 SE +/- 0.13, N = 3 SE +/- 0.08, N = 3 SE +/- 0.28, N = 3 SE +/- 0.17, N = 3 SE +/- 0.60, N = 3 SE +/- 0.14, N = 3 46.94 47.82 46.48 46.82 47.65 46.20 71.51 46.92 47.04 1. (CC) gcc options: -ldl -lz -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better SQLite Speedtest 3.30 Timed Time - Size 1,000 -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 Min: 46.65 / Avg: 46.94 / Max: 47.49 Min: 47.77 / Avg: 47.82 / Max: 47.89 Min: 45.76 / Avg: 46.48 / Max: 47.85 Min: 46.67 / Avg: 46.82 / Max: 47.08 Min: 47.52 / Avg: 47.65 / Max: 47.8 Min: 45.64 / Avg: 46.2 / Max: 46.54 Min: 71.17 / Avg: 71.51 / Max: 71.71 Min: 46.04 / Avg: 46.92 / Max: 48.06 Min: 46.9 / Avg: 47.04 / Max: 47.32 1. (CC) gcc options: -ldl -lz -lpthread
Basis Universal Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 3 -O2 -O3 -march=native -O3 -march=native -flto 16 32 48 64 80 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 SE +/- 0.05, N = 3 SE +/- 0.08, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 SE +/- 0.11, N = 3 46.14 46.32 46.16 46.45 46.33 46.43 73.46 45.86 45.82 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 3 -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 Min: 46.07 / Avg: 46.14 / Max: 46.24 Min: 46.21 / Avg: 46.32 / Max: 46.39 Min: 46.06 / Avg: 46.16 / Max: 46.21 Min: 46.3 / Avg: 46.45 / Max: 46.54 Min: 46.32 / Avg: 46.33 / Max: 46.35 Min: 46.41 / Avg: 46.42 / Max: 46.44 Min: 73.43 / Avg: 73.46 / Max: 73.52 Min: 45.75 / Avg: 45.86 / Max: 45.93 Min: 45.61 / Avg: 45.82 / Max: 45.93 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
AOM AV1 This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 0 Two-Pass -O2 -O3 -march=native -O3 -march=native -flto 0.0878 0.1756 0.2634 0.3512 0.439 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 15 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 0.38 0.38 0.38 0.38 0.38 0.38 0.25 0.38 0.39 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 0 Two-Pass -O2 -O3 -march=native -O3 -march=native -flto 1 2 3 4 5 Min: 0.37 / Avg: 0.38 / Max: 0.39 Min: 0.38 / Avg: 0.38 / Max: 0.38 Min: 0.38 / Avg: 0.38 / Max: 0.39 Min: 0.38 / Avg: 0.38 / Max: 0.39 Min: 0.38 / Avg: 0.38 / Max: 0.39 Min: 0.38 / Avg: 0.38 / Max: 0.39 Min: 0.23 / Avg: 0.25 / Max: 0.27 Min: 0.37 / Avg: 0.38 / Max: 0.39 Min: 0.39 / Avg: 0.39 / Max: 0.39 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Sequential Fill -O2 -O3 -march=native -O3 -march=native -flto 300K 600K 900K 1200K 1500K SE +/- 18510.12, N = 15 SE +/- 18378.16, N = 15 SE +/- 6066.61, N = 3 SE +/- 20649.20, N = 3 SE +/- 23355.92, N = 3 SE +/- 12160.87, N = 15 SE +/- 2724.87, N = 3 SE +/- 16768.52, N = 15 SE +/- 19584.51, N = 3 1397848 1384169 1362454 1355744 1389440 1356731 308338 1343053 1349966 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Op/s, More Is Better Facebook RocksDB 6.3.6 Test: Sequential Fill -O2 -O3 -march=native -O3 -march=native -flto 200K 400K 600K 800K 1000K Min: 1273276 / Avg: 1397848.47 / Max: 1557406 Min: 1279748 / Avg: 1384169.47 / Max: 1549202 Min: 1350336 / Avg: 1362454.33 / Max: 1369034 Min: 1328724 / Avg: 1355743.67 / Max: 1396302 Min: 1360936 / Avg: 1389440 / Max: 1435741 Min: 1290396 / Avg: 1356731.4 / Max: 1450744 Min: 303600 / Avg: 308338.33 / Max: 313039 Min: 1239975 / Avg: 1343053 / Max: 1450213 Min: 1319754 / Avg: 1349966 / Max: 1386661 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fno-builtin-memcmp -fno-rtti -rdynamic -lpthread
Basis Universal Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: ETC1S -O2 -O3 -march=native -O3 -march=native -flto 16 32 48 64 80 SE +/- 0.16, N = 3 SE +/- 0.22, N = 3 SE +/- 0.16, N = 3 SE +/- 0.23, N = 3 SE +/- 0.15, N = 3 SE +/- 0.44, N = 3 SE +/- 0.27, N = 3 SE +/- 0.50, N = 3 SE +/- 0.10, N = 3 42.83 43.71 43.51 43.34 43.38 43.32 71.69 42.13 42.79 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: ETC1S -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 Min: 42.62 / Avg: 42.83 / Max: 43.14 Min: 43.46 / Avg: 43.71 / Max: 44.14 Min: 43.25 / Avg: 43.51 / Max: 43.78 Min: 42.92 / Avg: 43.34 / Max: 43.69 Min: 43.19 / Avg: 43.38 / Max: 43.67 Min: 42.84 / Avg: 43.32 / Max: 44.21 Min: 71.27 / Avg: 71.69 / Max: 72.21 Min: 41.23 / Avg: 42.13 / Max: 42.94 Min: 42.62 / Avg: 42.79 / Max: 42.98 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
Minion Minion is an open-source constraint solver that is designed to be very scalable. This test profile uses Minion's integrated benchmarking problems to solve. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Graceful -O2 -O3 -march=native -O3 -march=native -flto 13 26 39 52 65 SE +/- 0.02, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.11, N = 3 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.46, N = 3 SE +/- 0.48, N = 3 SE +/- 0.09, N = 3 43.96 43.05 43.61 43.59 42.31 43.35 58.85 41.30 43.17 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Minion 1.8 Benchmark: Graceful -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 Min: 43.93 / Avg: 43.96 / Max: 44 Min: 42.98 / Avg: 43.05 / Max: 43.15 Min: 43.54 / Avg: 43.61 / Max: 43.68 Min: 43.37 / Avg: 43.59 / Max: 43.76 Min: 42.29 / Avg: 42.31 / Max: 42.34 Min: 43.27 / Avg: 43.35 / Max: 43.4 Min: 58.34 / Avg: 58.85 / Max: 59.77 Min: 40.75 / Avg: 41.3 / Max: 42.26 Min: 43.07 / Avg: 43.17 / Max: 43.35 1. (CXX) g++ options: -std=gnu++11 -O3 -fomit-frame-pointer -rdynamic
Basis Universal Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 2 -O2 -O3 -march=native -O3 -march=native -flto 9 18 27 36 45 SE +/- 0.10, N = 3 SE +/- 0.06, N = 3 SE +/- 0.03, N = 3 SE +/- 0.06, N = 3 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.36, N = 15 SE +/- 0.02, N = 3 SE +/- 0.08, N = 3 25.08 25.24 25.14 25.32 25.15 25.19 38.45 24.85 24.91 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 2 -O2 -O3 -march=native -O3 -march=native -flto 8 16 24 32 40 Min: 24.98 / Avg: 25.08 / Max: 25.28 Min: 25.18 / Avg: 25.24 / Max: 25.35 Min: 25.1 / Avg: 25.14 / Max: 25.21 Min: 25.19 / Avg: 25.32 / Max: 25.39 Min: 25.13 / Avg: 25.15 / Max: 25.16 Min: 25.08 / Avg: 25.19 / Max: 25.25 Min: 37.01 / Avg: 38.45 / Max: 40.38 Min: 24.83 / Avg: 24.85 / Max: 24.89 Min: 24.79 / Avg: 24.91 / Max: 25.06 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
Apache Benchmark This is a test of ab, which is the Apache benchmark program. This test profile measures how many requests per second a given system can sustain when carrying out 1,000,000 requests with 100 requests being carried out concurrently. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Requests Per Second, More Is Better Apache Benchmark 2.4.29 Static Web Page Serving -O2 -O3 -march=native -O3 -march=native -flto 9K 18K 27K 36K 45K SE +/- 35.38, N = 3 SE +/- 246.05, N = 3 SE +/- 640.04, N = 3 SE +/- 455.16, N = 3 SE +/- 37.31, N = 3 SE +/- 247.64, N = 3 SE +/- 320.59, N = 15 SE +/- 189.14, N = 3 40820.72 40794.45 38913.10 40288.65 40876.30 39804.02 34551.32 40544.78 1. (CC) gcc options: -shared -fPIC -pthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Requests Per Second, More Is Better Apache Benchmark 2.4.29 Static Web Page Serving -O2 -O3 -march=native -O3 -march=native -flto 7K 14K 21K 28K 35K Min: 40772.4 / Avg: 40820.72 / Max: 40889.64 Min: 40302.99 / Avg: 40794.45 / Max: 41062.07 Min: 37865.69 / Avg: 38913.1 / Max: 40074.11 Min: 39490.44 / Avg: 40288.65 / Max: 41066.78 Min: 40837.73 / Avg: 40876.3 / Max: 40950.91 Min: 39359.01 / Avg: 39804.02 / Max: 40214.82 Min: 32787.61 / Avg: 34551.32 / Max: 37500.83 Min: 40166.78 / Avg: 40544.78 / Max: 40746.59 1. (CC) gcc options: -shared -fPIC -pthread
AOM AV1 This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 6 Two-Pass -O2 -O3 -march=native -O3 -march=native -flto 1.062 2.124 3.186 4.248 5.31 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.04, N = 15 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 4.72 4.60 4.62 4.66 4.61 4.70 3.07 4.67 4.70 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 6 Two-Pass -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 Min: 4.69 / Avg: 4.72 / Max: 4.76 Min: 4.57 / Avg: 4.6 / Max: 4.64 Min: 4.62 / Avg: 4.62 / Max: 4.62 Min: 4.65 / Avg: 4.66 / Max: 4.66 Min: 4.6 / Avg: 4.61 / Max: 4.62 Min: 4.68 / Avg: 4.7 / Max: 4.71 Min: 2.83 / Avg: 3.07 / Max: 3.45 Min: 4.63 / Avg: 4.67 / Max: 4.73 Min: 4.68 / Avg: 4.7 / Max: 4.73 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread
7-Zip Compression This is a test of 7-Zip using p7zip with its integrated benchmark feature or upstream 7-Zip for the Windows x64 build. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 16.02 Compress Speed Test -O2 -O3 -march=native -O3 -march=native -flto 13K 26K 39K 52K 65K SE +/- 117.40, N = 3 SE +/- 68.12, N = 3 SE +/- 111.41, N = 3 SE +/- 77.51, N = 3 SE +/- 77.59, N = 3 SE +/- 145.26, N = 3 SE +/- 118.60, N = 3 SE +/- 148.12, N = 3 SE +/- 118.80, N = 3 59969 59969 60368 59931 59966 59890 39892 59283 59799 1. (CXX) g++ options: -pipe -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MIPS, More Is Better 7-Zip Compression 16.02 Compress Speed Test -O2 -O3 -march=native -O3 -march=native -flto 10K 20K 30K 40K 50K Min: 59845 / Avg: 59969.33 / Max: 60204 Min: 59894 / Avg: 59969 / Max: 60105 Min: 60203 / Avg: 60367.67 / Max: 60580 Min: 59820 / Avg: 59930.67 / Max: 60080 Min: 59878 / Avg: 59966.33 / Max: 60121 Min: 59649 / Avg: 59890 / Max: 60151 Min: 39679 / Avg: 39891.67 / Max: 40089 Min: 58988 / Avg: 59283 / Max: 59454 Min: 59666 / Avg: 59799 / Max: 60036 1. (CXX) g++ options: -pipe -lpthread
SMHasher SMHasher is a hash function tester. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: t1ha2_atonce -O2 -O3 -march=native -O3 -march=native -flto 8 16 24 32 40 SE +/- 0.00, N = 15 SE +/- 0.00, N = 3 SE +/- 0.00, N = 15 SE +/- 0.00, N = 15 SE +/- 0.00, N = 15 SE +/- 0.00, N = 15 SE +/- 0.00, N = 3 SE +/- 0.01, N = 4 SE +/- 0.00, N = 15 32.08 31.07 32.57 31.07 32.07 32.58 32.07 30.85 32.07 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: t1ha2_atonce -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 Min: 32.07 / Avg: 32.08 / Max: 32.11 Min: 31.07 / Avg: 31.07 / Max: 31.08 Min: 32.56 / Avg: 32.57 / Max: 32.58 Min: 31.07 / Avg: 31.07 / Max: 31.1 Min: 32.07 / Avg: 32.07 / Max: 32.07 Min: 32.58 / Avg: 32.58 / Max: 32.61 Min: 32.07 / Avg: 32.07 / Max: 32.07 Min: 30.84 / Avg: 30.85 / Max: 30.89 Min: 32.07 / Avg: 32.07 / Max: 32.08 1. (CXX) g++ options: -march=native -O3 -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: t1ha2_atonce -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K SE +/- 278.22, N = 15 SE +/- 143.38, N = 3 SE +/- 297.98, N = 15 SE +/- 281.78, N = 15 SE +/- 225.72, N = 15 SE +/- 196.72, N = 15 SE +/- 138.52, N = 3 SE +/- 242.60, N = 4 SE +/- 236.14, N = 15 15753.82 17491.59 16350.17 16578.46 16775.87 16358.06 16512.21 16868.68 15028.14 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: t1ha2_atonce -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K Min: 13911 / Avg: 15753.82 / Max: 17283.63 Min: 17340.99 / Avg: 17491.59 / Max: 17778.24 Min: 14463.2 / Avg: 16350.17 / Max: 17593.98 Min: 15383.15 / Avg: 16578.46 / Max: 18173.79 Min: 14561.5 / Avg: 16775.87 / Max: 17585.49 Min: 14902.93 / Avg: 16358.06 / Max: 17409.15 Min: 16235.2 / Avg: 16512.21 / Max: 16653.97 Min: 16148.31 / Avg: 16868.68 / Max: 17166.28 Min: 13665.73 / Avg: 15028.14 / Max: 16142.63 1. (CXX) g++ options: -march=native -O3 -lpthread
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Ctype -O2 -O3 -march=native -O3 -march=native -flto 9 18 27 36 45 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.41, N = 7 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 25.92 25.03 24.75 25.19 25.97 23.78 38.41 24.93 25.24 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Ctype -O2 -O3 -march=native -O3 -march=native -flto 8 16 24 32 40 Min: 25.91 / Avg: 25.92 / Max: 25.92 Min: 25.03 / Avg: 25.03 / Max: 25.04 Min: 24.73 / Avg: 24.75 / Max: 24.78 Min: 25.18 / Avg: 25.19 / Max: 25.19 Min: 25.97 / Avg: 25.97 / Max: 25.98 Min: 23.77 / Avg: 23.78 / Max: 23.79 Min: 37.95 / Avg: 38.41 / Max: 40.88 Min: 24.93 / Avg: 24.93 / Max: 24.94 Min: 25.24 / Avg: 25.24 / Max: 25.25 1. (CXX) g++ options: -std=c++11 -O3
John The Ripper This is a benchmark of John The Ripper, which is a password cracker. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Real C/S, More Is Better John The Ripper 1.9.0-jumbo-1 Test: Blowfish -O2 -O3 -march=native -O3 -march=native -flto 5K 10K 15K 20K 25K SE +/- 3.53, N = 3 SE +/- 7.86, N = 3 SE +/- 3.71, N = 3 SE +/- 0.67, N = 3 SE +/- 2.67, N = 3 SE +/- 42.16, N = 3 SE +/- 4.00, N = 3 SE +/- 2.91, N = 3 23561 22794 23509 23091 22655 22765 5078 22892 22463 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -pthread -lm -lz -ldl -lcrypt -lbz2 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Real C/S, More Is Better John The Ripper 1.9.0-jumbo-1 Test: Blowfish -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K Min: 23554 / Avg: 23560.67 / Max: 23566 Min: 23494 / Avg: 23509.33 / Max: 23520 Min: 23086 / Avg: 23090.67 / Max: 23098 Min: 22654 / Avg: 22655.33 / Max: 22656 Min: 22762 / Avg: 22764.67 / Max: 22770 Min: 5006 / Avg: 5078 / Max: 5152 Min: 22884 / Avg: 22892 / Max: 22896 Min: 22458 / Avg: 22462.67 / Max: 22468 1. (CC) gcc options: -m64 -lssl -lcrypto -fopenmp -lgmp -pthread -lm -lz -ldl -lcrypt -lbz2
AOM AV1 This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 4 Two-Pass -O2 -O3 -march=native -O3 -march=native -flto 0.6728 1.3456 2.0184 2.6912 3.364 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 15 SE +/- 0.02, N = 3 SE +/- 0.03, N = 3 2.99 2.93 2.93 2.99 2.98 2.99 2.01 2.98 2.96 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 4 Two-Pass -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 Min: 2.98 / Avg: 2.99 / Max: 3 Min: 2.9 / Avg: 2.93 / Max: 2.97 Min: 2.92 / Avg: 2.93 / Max: 2.94 Min: 2.98 / Avg: 2.99 / Max: 3.01 Min: 2.97 / Avg: 2.98 / Max: 3.01 Min: 2.98 / Avg: 2.99 / Max: 3.01 Min: 1.82 / Avg: 2.01 / Max: 2.23 Min: 2.94 / Avg: 2.98 / Max: 3.01 Min: 2.93 / Avg: 2.96 / Max: 3.01 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Stepanov Abstraction -O2 -O3 -march=native -O3 -march=native -flto 8 16 24 32 40 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 28.04 28.09 27.17 27.80 27.93 27.06 33.47 27.16 28.05 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Stepanov Abstraction -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 Min: 28.03 / Avg: 28.04 / Max: 28.04 Min: 28.08 / Avg: 28.09 / Max: 28.09 Min: 27.17 / Avg: 27.17 / Max: 27.18 Min: 27.79 / Avg: 27.8 / Max: 27.81 Min: 27.93 / Avg: 27.93 / Max: 27.94 Min: 27.05 / Avg: 27.05 / Max: 27.06 Min: 33.47 / Avg: 33.47 / Max: 33.48 Min: 27.15 / Avg: 27.16 / Max: 27.16 Min: 28.05 / Avg: 28.05 / Max: 28.06 1. (CXX) g++ options: -std=c++11 -O3
SciMark This test runs the ANSI C version of SciMark 2.0, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This test is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Composite -O2 -O3 -march=native -O3 -march=native -flto 800 1600 2400 3200 4000 SE +/- 16.70, N = 3 SE +/- 22.11, N = 3 SE +/- 4.76, N = 3 SE +/- 10.87, N = 3 SE +/- 38.42, N = 3 SE +/- 4.31, N = 3 SE +/- 1.49, N = 3 SE +/- 3.27, N = 3 SE +/- 0.74, N = 3 2346.95 3425.82 2454.76 2445.14 3510.66 2560.09 1564.88 3641.48 2701.45 1. (CC) gcc options: -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Composite -O2 -O3 -march=native -O3 -march=native -flto 600 1200 1800 2400 3000 Min: 2328.86 / Avg: 2346.95 / Max: 2380.3 Min: 3386.84 / Avg: 3425.82 / Max: 3463.38 Min: 2448.83 / Avg: 2454.76 / Max: 2464.18 Min: 2423.5 / Avg: 2445.14 / Max: 2457.71 Min: 3433.85 / Avg: 3510.66 / Max: 3550.88 Min: 2551.49 / Avg: 2560.09 / Max: 2564.88 Min: 1561.92 / Avg: 1564.88 / Max: 1566.57 Min: 3635.04 / Avg: 3641.48 / Max: 3645.69 Min: 2699.98 / Avg: 2701.45 / Max: 2702.37 1. (CC) gcc options: -lm
AOBench AOBench is a lightweight ambient occlusion renderer, written in C. The test profile is using a size of 2048 x 2048. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better AOBench Size: 2048 x 2048 - Total Time -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.07, N = 3 SE +/- 0.05, N = 3 SE +/- 0.04, N = 3 SE +/- 0.10, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 28.19 24.51 26.06 28.28 24.63 25.09 43.26 23.97 23.88 1. (CC) gcc options: -lm -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better AOBench Size: 2048 x 2048 - Total Time -O2 -O3 -march=native -O3 -march=native -flto 9 18 27 36 45 Min: 28.17 / Avg: 28.18 / Max: 28.2 Min: 24.48 / Avg: 24.51 / Max: 24.57 Min: 26.04 / Avg: 26.06 / Max: 26.07 Min: 28.18 / Avg: 28.28 / Max: 28.4 Min: 24.57 / Avg: 24.63 / Max: 24.73 Min: 25.03 / Avg: 25.09 / Max: 25.15 Min: 43.12 / Avg: 43.26 / Max: 43.44 Min: 23.95 / Avg: 23.97 / Max: 24.01 Min: 23.87 / Avg: 23.88 / Max: 23.9 1. (CC) gcc options: -lm -O3
Zstd Compression This test measures the time needed to compress a sample file (an Ubuntu ISO) using Zstd compression. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.5 Compression Level: 3 -O2 -O3 -march=native -O3 -march=native -flto 600 1200 1800 2400 3000 SE +/- 1.03, N = 3 SE +/- 11.14, N = 3 SE +/- 24.64, N = 3 SE +/- 7.32, N = 3 SE +/- 13.83, N = 3 SE +/- 9.66, N = 3 SE +/- 7.60, N = 3 SE +/- 21.98, N = 3 SE +/- 15.30, N = 3 2559.1 2538.8 2534.3 2575.3 2543.2 2560.3 2571.2 2533.0 2551.7 1. (CC) gcc options: -pthread -lz -llzma Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better Zstd Compression 1.4.5 Compression Level: 3 -O2 -O3 -march=native -O3 -march=native -flto 400 800 1200 1600 2000 Min: 2557.8 / Avg: 2559.07 / Max: 2561.1 Min: 2516.8 / Avg: 2538.83 / Max: 2552.7 Min: 2490.8 / Avg: 2534.27 / Max: 2576.1 Min: 2567 / Avg: 2575.3 / Max: 2589.9 Min: 2515.8 / Avg: 2543.2 / Max: 2560.2 Min: 2541.3 / Avg: 2560.3 / Max: 2572.8 Min: 2562 / Avg: 2571.23 / Max: 2586.3 Min: 2492.3 / Avg: 2533.03 / Max: 2567.7 Min: 2533.7 / Avg: 2551.67 / Max: 2582.1 1. (CC) gcc options: -pthread -lz -llzma
Darmstadt Automotive Parallel Heterogeneous Suite DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Test Cases Per Minute, More Is Better Darmstadt Automotive Parallel Heterogeneous Suite Backend: OpenMP - Kernel: Euclidean Cluster -O2 -O3 -march=native -O3 -march=native -flto 300 600 900 1200 1500 SE +/- 0.37, N = 3 SE +/- 1.12, N = 3 SE +/- 1.13, N = 3 SE +/- 0.18, N = 3 SE +/- 0.25, N = 3 SE +/- 0.96, N = 3 SE +/- 1.14, N = 3 SE +/- 2.34, N = 3 SE +/- 0.48, N = 3 1457.34 1551.81 1492.07 1462.41 1544.70 1506.61 173.68 1582.78 1470.91 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Test Cases Per Minute, More Is Better Darmstadt Automotive Parallel Heterogeneous Suite Backend: OpenMP - Kernel: Euclidean Cluster -O2 -O3 -march=native -O3 -march=native -flto 300 600 900 1200 1500 Min: 1456.79 / Avg: 1457.34 / Max: 1458.05 Min: 1549.68 / Avg: 1551.81 / Max: 1553.5 Min: 1489.86 / Avg: 1492.07 / Max: 1493.61 Min: 1462.07 / Avg: 1462.41 / Max: 1462.66 Min: 1544.19 / Avg: 1544.7 / Max: 1545.01 Min: 1504.93 / Avg: 1506.61 / Max: 1508.25 Min: 172.25 / Avg: 173.68 / Max: 175.93 Min: 1578.13 / Avg: 1582.78 / Max: 1585.56 Min: 1469.99 / Avg: 1470.91 / Max: 1471.59 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp
XZ Compression This test measures the time needed to compress a sample file (an Ubuntu file-system image) using XZ compression. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better XZ Compression 5.2.4 Compressing ubuntu-16.04.3-server-i386.img, Compression Level 9 -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.01, N = 3 SE +/- 0.07, N = 3 SE +/- 0.17, N = 3 SE +/- 0.17, N = 3 SE +/- 0.04, N = 3 SE +/- 0.11, N = 3 25.66 25.40 25.20 25.56 25.60 25.38 30.46 25.47 25.05 1. (CC) gcc options: -pthread -fvisibility=hidden Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better XZ Compression 5.2.4 Compressing ubuntu-16.04.3-server-i386.img, Compression Level 9 -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 Min: 25.64 / Avg: 25.66 / Max: 25.69 Min: 25.36 / Avg: 25.4 / Max: 25.42 Min: 25.12 / Avg: 25.2 / Max: 25.25 Min: 25.55 / Avg: 25.56 / Max: 25.58 Min: 25.46 / Avg: 25.6 / Max: 25.68 Min: 25.21 / Avg: 25.38 / Max: 25.72 Min: 30.14 / Avg: 30.46 / Max: 30.72 Min: 25.41 / Avg: 25.47 / Max: 25.54 Min: 24.93 / Avg: 25.05 / Max: 25.27 1. (CC) gcc options: -pthread -fvisibility=hidden
Coremark This is a test of EEMBC CoreMark processor benchmark. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Iterations/Sec, More Is Better Coremark 1.0 CoreMark Size 666 - Iterations Per Second -O2 -O3 -march=native -O3 -march=native -flto 120K 240K 360K 480K 600K SE +/- 1093.60, N = 3 SE +/- 5023.84, N = 3 SE +/- 145.06, N = 3 SE +/- 431.91, N = 3 SE +/- 2378.13, N = 3 SE +/- 1551.88, N = 3 SE +/- 1498.61, N = 3 SE +/- 1594.98, N = 3 SE +/- 655.75, N = 3 526975.19 537687.29 554306.73 517390.78 541993.71 539052.14 323826.26 520000.55 559563.83 1. (CC) gcc options: -O2 -lrt" -lrt Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Iterations/Sec, More Is Better Coremark 1.0 CoreMark Size 666 - Iterations Per Second -O2 -O3 -march=native -O3 -march=native -flto 100K 200K 300K 400K 500K Min: 524911.42 / Avg: 526975.19 / Max: 528634.36 Min: 532197.98 / Avg: 537687.29 / Max: 547720.12 Min: 554016.62 / Avg: 554306.73 / Max: 554451.79 Min: 516662.36 / Avg: 517390.78 / Max: 518157.09 Min: 537417.71 / Avg: 541993.71 / Max: 545404.96 Min: 536336.82 / Avg: 539052.14 / Max: 541711.81 Min: 320962.89 / Avg: 323826.26 / Max: 326024.94 Min: 517085.36 / Avg: 520000.55 / Max: 522579.8 Min: 558451.23 / Avg: 559563.83 / Max: 560721.46 1. (CC) gcc options: -O2 -lrt" -lrt
Darmstadt Automotive Parallel Heterogeneous Suite DAPHNE is the Darmstadt Automotive Parallel HeterogeNEous Benchmark Suite with OpenCL / CUDA / OpenMP test cases for these automotive benchmarks for evaluating programming models in context to vehicle autonomous driving capabilities. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Test Cases Per Minute, More Is Better Darmstadt Automotive Parallel Heterogeneous Suite Backend: OpenMP - Kernel: NDT Mapping -O2 -O3 -march=native -O3 -march=native -flto 150 300 450 600 750 SE +/- 1.02, N = 3 SE +/- 0.87, N = 3 SE +/- 0.91, N = 3 SE +/- 0.87, N = 3 SE +/- 0.76, N = 3 SE +/- 0.93, N = 3 SE +/- 2.05, N = 3 SE +/- 1.09, N = 3 SE +/- 0.92, N = 3 708.57 709.70 711.52 708.20 702.61 708.35 231.02 704.66 708.57 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Test Cases Per Minute, More Is Better Darmstadt Automotive Parallel Heterogeneous Suite Backend: OpenMP - Kernel: NDT Mapping -O2 -O3 -march=native -O3 -march=native -flto 130 260 390 520 650 Min: 706.55 / Avg: 708.57 / Max: 709.81 Min: 708.05 / Avg: 709.7 / Max: 711 Min: 709.7 / Avg: 711.52 / Max: 712.52 Min: 706.47 / Avg: 708.2 / Max: 709.18 Min: 701.19 / Avg: 702.61 / Max: 703.8 Min: 706.53 / Avg: 708.35 / Max: 709.58 Min: 227.8 / Avg: 231.02 / Max: 234.84 Min: 702.49 / Avg: 704.66 / Max: 705.91 Min: 706.74 / Avg: 708.57 / Max: 709.51 1. (CXX) g++ options: -O3 -std=c++11 -fopenmp
AOM AV1 This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 6 Realtime -O2 -O3 -march=native -O3 -march=native -flto 6 12 18 24 30 SE +/- 0.07, N = 3 SE +/- 0.24, N = 3 SE +/- 0.13, N = 3 SE +/- 0.21, N = 3 SE +/- 0.13, N = 3 SE +/- 0.22, N = 3 SE +/- 0.06, N = 3 SE +/- 0.05, N = 3 SE +/- 0.24, N = 3 25.94 25.79 25.89 25.72 25.79 25.96 16.71 25.68 26.26 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 6 Realtime -O2 -O3 -march=native -O3 -march=native -flto 6 12 18 24 30 Min: 25.82 / Avg: 25.94 / Max: 26.06 Min: 25.32 / Avg: 25.79 / Max: 26.05 Min: 25.7 / Avg: 25.89 / Max: 26.15 Min: 25.33 / Avg: 25.72 / Max: 26.05 Min: 25.57 / Avg: 25.79 / Max: 26.01 Min: 25.54 / Avg: 25.96 / Max: 26.26 Min: 16.63 / Avg: 16.71 / Max: 16.82 Min: 25.63 / Avg: 25.68 / Max: 25.78 Min: 25.79 / Avg: 26.26 / Max: 26.52 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Sequential Fill -O2 -O3 -march=native -O3 -march=native -flto 13 26 39 52 65 SE +/- 0.25, N = 3 SE +/- 0.05, N = 3 SE +/- 0.33, N = 3 SE +/- 0.05, N = 3 SE +/- 0.21, N = 3 SE +/- 0.25, N = 3 SE +/- 0.11, N = 3 SE +/- 0.10, N = 3 SE +/- 0.18, N = 3 47.38 46.59 47.12 46.83 46.86 46.23 59.75 47.03 46.26 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Sequential Fill -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 Min: 47.13 / Avg: 47.38 / Max: 47.87 Min: 46.5 / Avg: 46.59 / Max: 46.65 Min: 46.65 / Avg: 47.12 / Max: 47.77 Min: 46.77 / Avg: 46.83 / Max: 46.93 Min: 46.45 / Avg: 46.86 / Max: 47.17 Min: 45.75 / Avg: 46.23 / Max: 46.52 Min: 59.61 / Avg: 59.75 / Max: 59.97 Min: 46.89 / Avg: 47.03 / Max: 47.22 Min: 45.96 / Avg: 46.26 / Max: 46.59 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Sequential Fill -O2 -O3 -march=native -O3 -march=native -flto 11 22 33 44 55 SE +/- 0.23, N = 3 SE +/- 0.07, N = 3 SE +/- 0.33, N = 3 SE +/- 0.07, N = 3 SE +/- 0.21, N = 3 SE +/- 0.25, N = 3 SE +/- 0.06, N = 3 SE +/- 0.12, N = 3 SE +/- 0.17, N = 3 46.7 47.5 46.9 47.2 47.2 47.8 37.0 47.0 47.8 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Sequential Fill -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 Min: 46.2 / Avg: 46.67 / Max: 46.9 Min: 47.4 / Avg: 47.47 / Max: 47.6 Min: 46.3 / Avg: 46.93 / Max: 47.4 Min: 47.1 / Avg: 47.23 / Max: 47.3 Min: 46.9 / Avg: 47.2 / Max: 47.6 Min: 47.5 / Avg: 47.8 / Max: 48.3 Min: 36.9 / Avg: 37 / Max: 37.1 Min: 46.8 / Avg: 47.03 / Max: 47.2 Min: 47.5 / Avg: 47.8 / Max: 48.1 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Delete -O2 -O3 -march=native -O3 -march=native -flto 13 26 39 52 65 SE +/- 0.19, N = 3 SE +/- 0.14, N = 3 SE +/- 0.16, N = 3 SE +/- 0.14, N = 3 SE +/- 0.35, N = 3 SE +/- 0.24, N = 3 SE +/- 0.04, N = 3 SE +/- 0.14, N = 3 SE +/- 0.21, N = 3 46.52 46.65 46.30 45.98 46.95 45.97 58.72 46.26 46.15 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Delete -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 Min: 46.15 / Avg: 46.52 / Max: 46.72 Min: 46.48 / Avg: 46.65 / Max: 46.93 Min: 45.99 / Avg: 46.3 / Max: 46.48 Min: 45.77 / Avg: 45.98 / Max: 46.25 Min: 46.56 / Avg: 46.95 / Max: 47.65 Min: 45.64 / Avg: 45.97 / Max: 46.45 Min: 58.65 / Avg: 58.72 / Max: 58.8 Min: 45.99 / Avg: 46.26 / Max: 46.45 Min: 45.72 / Avg: 46.15 / Max: 46.38 1. (CXX) g++ options: -O3 -lsnappy -lpthread
FLAC Audio Encoding This test times how long it takes to encode a sample WAV file to FLAC format five times. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better FLAC Audio Encoding 1.3.2 WAV To FLAC -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 SE +/- 0.095, N = 6 SE +/- 0.047, N = 5 SE +/- 0.061, N = 14 SE +/- 0.070, N = 5 SE +/- 0.061, N = 5 SE +/- 0.082, N = 7 SE +/- 0.072, N = 5 SE +/- 0.056, N = 25 SE +/- 0.080, N = 7 7.763 7.943 7.804 8.091 8.025 7.661 11.243 7.385 7.483 1. (CXX) g++ options: -fvisibility=hidden -logg -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better FLAC Audio Encoding 1.3.2 WAV To FLAC -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 Min: 7.48 / Avg: 7.76 / Max: 8.07 Min: 7.87 / Avg: 7.94 / Max: 8.13 Min: 7.29 / Avg: 7.8 / Max: 8.16 Min: 7.89 / Avg: 8.09 / Max: 8.23 Min: 7.86 / Avg: 8.03 / Max: 8.19 Min: 7.3 / Avg: 7.66 / Max: 7.95 Min: 11.04 / Avg: 11.24 / Max: 11.42 Min: 6.9 / Avg: 7.38 / Max: 7.98 Min: 7.29 / Avg: 7.48 / Max: 7.79 1. (CXX) g++ options: -fvisibility=hidden -logg -lm
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Deconvolution Batch deconv_1d - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 80 160 240 320 400 SE +/- 0.00425, N = 3 SE +/- 0.00152, N = 3 SE +/- 0.00421, N = 3 SE +/- 0.00392, N = 3 SE +/- 0.00272, N = 3 SE +/- 0.00307, N = 3 SE +/- 0.64545, N = 3 SE +/- 0.00372, N = 3 SE +/- 0.00625, N = 3 2.97255 2.97413 2.97576 2.99286 2.97665 2.97503 350.88700 2.93195 2.95134 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Deconvolution Batch deconv_1d - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 60 120 180 240 300 Min: 2.97 / Avg: 2.97 / Max: 2.98 Min: 2.97 / Avg: 2.97 / Max: 2.98 Min: 2.97 / Avg: 2.98 / Max: 2.98 Min: 2.99 / Avg: 2.99 / Max: 3 Min: 2.97 / Avg: 2.98 / Max: 2.98 Min: 2.97 / Avg: 2.98 / Max: 2.98 Min: 349.92 / Avg: 350.89 / Max: 352.11 Min: 2.93 / Avg: 2.93 / Max: 2.94 Min: 2.94 / Avg: 2.95 / Max: 2.96 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
OpenSSL OpenSSL is an open-source toolkit that implements SSL (Secure Sockets Layer) and TLS (Transport Layer Security) protocols. This test measures the RSA 4096-bit performance of OpenSSL. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Signs Per Second, More Is Better OpenSSL 1.1.1 RSA 4096-bit Performance -O2 -O3 -march=native -O3 -march=native -flto 800 1600 2400 3200 4000 SE +/- 6.08, N = 3 SE +/- 7.20, N = 3 SE +/- 8.30, N = 3 SE +/- 2.34, N = 3 SE +/- 6.38, N = 3 SE +/- 9.30, N = 3 SE +/- 12.64, N = 3 SE +/- 7.35, N = 3 SE +/- 6.80, N = 3 3716.0 3738.0 3749.8 3739.1 3747.2 3741.0 2500.2 3745.1 3741.0 1. (CC) gcc options: -pthread -m64 -lssl -lcrypto -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Signs Per Second, More Is Better OpenSSL 1.1.1 RSA 4096-bit Performance -O2 -O3 -march=native -O3 -march=native -flto 700 1400 2100 2800 3500 Min: 3707.1 / Avg: 3715.97 / Max: 3727.6 Min: 3723.6 / Avg: 3738 / Max: 3745.3 Min: 3733.6 / Avg: 3749.77 / Max: 3761.1 Min: 3736.2 / Avg: 3739.07 / Max: 3743.7 Min: 3736 / Avg: 3747.17 / Max: 3758.1 Min: 3731.4 / Avg: 3741 / Max: 3759.6 Min: 2482.8 / Avg: 2500.23 / Max: 2524.8 Min: 3733.3 / Avg: 3745.13 / Max: 3758.6 Min: 3728.8 / Avg: 3740.97 / Max: 3752.3 1. (CC) gcc options: -pthread -m64 -lssl -lcrypto -ldl
SMHasher SMHasher is a hash function tester. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: Spooky32 -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 SE +/- 0.01, N = 3 SE +/- 0.01, N = 8 SE +/- 0.00, N = 15 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 41.25 39.71 42.47 39.72 41.21 42.47 40.88 39.29 40.94 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: Spooky32 -O2 -O3 -march=native -O3 -march=native -flto 9 18 27 36 45 Min: 41.24 / Avg: 41.25 / Max: 41.26 Min: 39.69 / Avg: 39.71 / Max: 39.74 Min: 42.46 / Avg: 42.47 / Max: 42.48 Min: 39.7 / Avg: 39.72 / Max: 39.73 Min: 41.2 / Avg: 41.21 / Max: 41.21 Min: 42.45 / Avg: 42.47 / Max: 42.48 Min: 40.87 / Avg: 40.88 / Max: 40.88 Min: 39.27 / Avg: 39.29 / Max: 39.3 Min: 40.93 / Avg: 40.94 / Max: 40.95 1. (CXX) g++ options: -march=native -O3 -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: Spooky32 -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K SE +/- 140.52, N = 3 SE +/- 161.29, N = 8 SE +/- 133.15, N = 15 SE +/- 3.82, N = 3 SE +/- 95.35, N = 3 SE +/- 106.46, N = 3 SE +/- 91.74, N = 3 SE +/- 119.67, N = 3 SE +/- 63.60, N = 3 15326.29 15989.72 14997.87 16096.57 15135.86 15766.13 15317.40 15807.06 15247.55 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: Spooky32 -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K Min: 15183.66 / Avg: 15326.29 / Max: 15607.32 Min: 15268.25 / Avg: 15989.72 / Max: 16492.27 Min: 14340.24 / Avg: 14997.87 / Max: 15850.46 Min: 16090.1 / Avg: 16096.57 / Max: 16103.31 Min: 14972.68 / Avg: 15135.86 / Max: 15302.92 Min: 15654.92 / Avg: 15766.13 / Max: 15978.98 Min: 15134.93 / Avg: 15317.4 / Max: 15425.34 Min: 15570.47 / Avg: 15807.06 / Max: 15956.73 Min: 15124.64 / Avg: 15247.55 / Max: 15337.35 1. (CXX) g++ options: -march=native -O3 -lpthread
Timed Apache Compilation This test times how long it takes to build the Apache HTTPD web server. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Timed Apache Compilation 2.4.41 Time To Compile -O2 -O3 -march=native -O3 -march=native -flto 5 10 15 20 25 SE +/- 0.02, N = 3 SE +/- 0.07, N = 3 SE +/- 0.04, N = 3 SE +/- 0.04, N = 3 SE +/- 0.08, N = 3 SE +/- 0.03, N = 3 SE +/- 0.26, N = 3 SE +/- 0.02, N = 3 14.79 15.91 17.29 15.11 16.54 17.65 22.57 16.60 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Timed Apache Compilation 2.4.41 Time To Compile -O2 -O3 -march=native -O3 -march=native -flto 5 10 15 20 25 Min: 14.74 / Avg: 14.79 / Max: 14.81 Min: 15.84 / Avg: 15.91 / Max: 16.05 Min: 17.22 / Avg: 17.29 / Max: 17.36 Min: 15.07 / Avg: 15.11 / Max: 15.19 Min: 16.43 / Avg: 16.54 / Max: 16.7 Min: 17.62 / Avg: 17.65 / Max: 17.71 Min: 22.11 / Avg: 22.57 / Max: 23.02 Min: 16.57 / Avg: 16.6 / Max: 16.63
SMHasher SMHasher is a hash function tester. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: t1ha0_aes_avx2 -O2 -O3 -march=native -O3 -march=native -flto 8 16 24 32 40 SE +/- 0.00, N = 8 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 8 SE +/- 0.00, N = 15 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.02, N = 3 32.73 31.11 32.81 31.07 32.21 32.07 32.07 31.07 32.12 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: t1ha0_aes_avx2 -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 Min: 32.7 / Avg: 32.73 / Max: 32.74 Min: 31.1 / Avg: 31.11 / Max: 31.11 Min: 32.81 / Avg: 32.81 / Max: 32.81 Min: 31.07 / Avg: 31.07 / Max: 31.07 Min: 32.15 / Avg: 32.21 / Max: 32.26 Min: 32.07 / Avg: 32.07 / Max: 32.07 Min: 32.07 / Avg: 32.07 / Max: 32.07 Min: 31.07 / Avg: 31.07 / Max: 31.07 Min: 32.08 / Avg: 32.12 / Max: 32.15 1. (CXX) g++ options: -march=native -O3 -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: t1ha0_aes_avx2 -O2 -O3 -march=native -O3 -march=native -flto 11K 22K 33K 44K 55K SE +/- 489.67, N = 8 SE +/- 233.96, N = 3 SE +/- 557.14, N = 3 SE +/- 134.72, N = 3 SE +/- 452.92, N = 8 SE +/- 467.28, N = 15 SE +/- 196.41, N = 3 SE +/- 359.74, N = 3 SE +/- 396.49, N = 3 47344.61 49050.29 45701.27 47949.93 44548.22 44735.32 43894.25 48098.68 49065.21 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: t1ha0_aes_avx2 -O2 -O3 -march=native -O3 -march=native -flto 9K 18K 27K 36K 45K Min: 45342.15 / Avg: 47344.61 / Max: 49096.63 Min: 48620.77 / Avg: 49050.29 / Max: 49425.8 Min: 45023.03 / Avg: 45701.27 / Max: 46806.03 Min: 47681.95 / Avg: 47949.93 / Max: 48108.16 Min: 43313.86 / Avg: 44548.22 / Max: 46518.49 Min: 41642.71 / Avg: 44735.32 / Max: 48011.44 Min: 43503.5 / Avg: 43894.25 / Max: 44124.56 Min: 47392.75 / Avg: 48098.68 / Max: 48572 Min: 48316.41 / Avg: 49065.21 / Max: 49665.64 1. (CXX) g++ options: -march=native -O3 -lpthread
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: IP Batch 1D - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 50 100 150 200 250 SE +/- 0.00891, N = 3 SE +/- 0.00239, N = 3 SE +/- 0.04819, N = 4 SE +/- 0.00218, N = 3 SE +/- 0.00956, N = 3 SE +/- 0.01673, N = 3 SE +/- 0.72309, N = 3 SE +/- 0.00901, N = 3 SE +/- 0.00850, N = 3 3.41291 3.41282 3.48994 3.45347 3.41167 3.39303 245.24900 3.40571 3.40923 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: IP Batch 1D - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 40 80 120 160 200 Min: 3.4 / Avg: 3.41 / Max: 3.42 Min: 3.41 / Avg: 3.41 / Max: 3.42 Min: 3.43 / Avg: 3.49 / Max: 3.63 Min: 3.45 / Avg: 3.45 / Max: 3.46 Min: 3.4 / Avg: 3.41 / Max: 3.43 Min: 3.37 / Avg: 3.39 / Max: 3.43 Min: 243.81 / Avg: 245.25 / Max: 246.12 Min: 3.39 / Avg: 3.41 / Max: 3.42 Min: 3.4 / Avg: 3.41 / Max: 3.43 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
AOM AV1 This is a simple test of the AOMedia AV1 encoder run on the CPU with a sample video file. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 8 Realtime -O2 -O3 -march=native -O3 -march=native -flto 11 22 33 44 55 SE +/- 0.21, N = 3 SE +/- 0.03, N = 3 SE +/- 0.04, N = 3 SE +/- 0.10, N = 3 SE +/- 0.05, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.09, N = 3 SE +/- 0.04, N = 3 48.30 48.28 49.02 48.16 48.49 49.43 31.95 48.78 49.69 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better AOM AV1 2.0 Encoder Mode: Speed 8 Realtime -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 Min: 47.99 / Avg: 48.3 / Max: 48.69 Min: 48.23 / Avg: 48.28 / Max: 48.33 Min: 48.95 / Avg: 49.02 / Max: 49.09 Min: 47.98 / Avg: 48.16 / Max: 48.31 Min: 48.39 / Avg: 48.49 / Max: 48.56 Min: 49.39 / Avg: 49.43 / Max: 49.45 Min: 31.94 / Avg: 31.95 / Max: 31.96 Min: 48.6 / Avg: 48.78 / Max: 48.88 Min: 49.61 / Avg: 49.69 / Max: 49.75 1. (CXX) g++ options: -O3 -std=c++11 -U_FORTIFY_SOURCE -lm -lpthread
SMHasher SMHasher is a hash function tester. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: wyhash -O2 -O3 -march=native -O3 -march=native -flto 6 12 18 24 30 SE +/- 0.31, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 7 SE +/- 0.01, N = 4 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.05, N = 15 23.66 23.36 24.26 23.65 24.49 24.53 24.49 23.55 24.60 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: wyhash -O2 -O3 -march=native -O3 -march=native -flto 6 12 18 24 30 Min: 23.36 / Avg: 23.66 / Max: 24.27 Min: 23.36 / Avg: 23.36 / Max: 23.36 Min: 24.26 / Avg: 24.26 / Max: 24.26 Min: 23.64 / Avg: 23.65 / Max: 23.66 Min: 24.48 / Avg: 24.49 / Max: 24.51 Min: 24.52 / Avg: 24.53 / Max: 24.53 Min: 24.49 / Avg: 24.49 / Max: 24.51 Min: 23.54 / Avg: 23.55 / Max: 23.55 Min: 24.55 / Avg: 24.6 / Max: 25.32 1. (CXX) g++ options: -march=native -O3 -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: wyhash -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K SE +/- 187.08, N = 3 SE +/- 26.79, N = 3 SE +/- 179.33, N = 3 SE +/- 200.17, N = 7 SE +/- 283.95, N = 4 SE +/- 102.55, N = 3 SE +/- 251.37, N = 3 SE +/- 41.75, N = 3 SE +/- 155.20, N = 15 19811.86 19612.78 19206.41 18758.21 19005.25 19361.06 17794.48 19417.84 18121.35 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: wyhash -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K Min: 19470.23 / Avg: 19811.86 / Max: 20114.83 Min: 19572.7 / Avg: 19612.78 / Max: 19663.61 Min: 19010.23 / Avg: 19206.41 / Max: 19564.53 Min: 18162.65 / Avg: 18758.21 / Max: 19605.26 Min: 18425.78 / Avg: 19005.25 / Max: 19732.59 Min: 19203.5 / Avg: 19361.06 / Max: 19553.55 Min: 17454.17 / Avg: 17794.48 / Max: 18285.11 Min: 19339.48 / Avg: 19417.84 / Max: 19481.99 Min: 17455.15 / Avg: 18121.35 / Max: 19061.75 1. (CXX) g++ options: -march=native -O3 -lpthread
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Read -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 SE +/- 0.106, N = 5 SE +/- 0.117, N = 4 SE +/- 0.076, N = 7 SE +/- 0.105, N = 5 SE +/- 0.084, N = 3 SE +/- 0.092, N = 4 SE +/- 0.115, N = 4 SE +/- 0.113, N = 3 SE +/- 0.068, N = 9 8.316 8.039 6.976 8.179 8.024 6.814 8.424 8.320 6.914 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Read -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 Min: 8.01 / Avg: 8.32 / Max: 8.64 Min: 7.72 / Avg: 8.04 / Max: 8.22 Min: 6.71 / Avg: 6.98 / Max: 7.34 Min: 7.81 / Avg: 8.18 / Max: 8.37 Min: 7.86 / Avg: 8.02 / Max: 8.13 Min: 6.58 / Avg: 6.81 / Max: 7.02 Min: 8.16 / Avg: 8.42 / Max: 8.69 Min: 8.11 / Avg: 8.32 / Max: 8.49 Min: 6.62 / Avg: 6.91 / Max: 7.2 1. (CXX) g++ options: -O3 -lsnappy -lpthread
SMHasher SMHasher is a hash function tester. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: MeowHash -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.24, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 52.47 50.48 52.43 49.72 51.48 51.78 51.48 49.48 51.49 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: MeowHash -O2 -O3 -march=native -O3 -march=native -flto 11 22 33 44 55 Min: 52.46 / Avg: 52.47 / Max: 52.47 Min: 50.47 / Avg: 50.48 / Max: 50.49 Min: 52.43 / Avg: 52.43 / Max: 52.44 Min: 49.48 / Avg: 49.72 / Max: 50.2 Min: 51.48 / Avg: 51.48 / Max: 51.48 Min: 51.78 / Avg: 51.78 / Max: 51.79 Min: 51.48 / Avg: 51.48 / Max: 51.48 Min: 49.48 / Avg: 49.48 / Max: 49.48 Min: 51.48 / Avg: 51.49 / Max: 51.49 1. (CXX) g++ options: -march=native -O3 -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: MeowHash -O2 -O3 -march=native -O3 -march=native -flto 9K 18K 27K 36K 45K SE +/- 181.83, N = 3 SE +/- 127.85, N = 3 SE +/- 52.15, N = 3 SE +/- 146.62, N = 3 SE +/- 104.32, N = 3 SE +/- 169.57, N = 3 SE +/- 113.93, N = 3 SE +/- 63.04, N = 3 SE +/- 59.97, N = 3 41279.80 42778.81 41310.68 42817.01 40824.30 41398.08 41503.35 43465.99 41815.99 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: MeowHash -O2 -O3 -march=native -O3 -march=native -flto 8K 16K 24K 32K 40K Min: 40917.34 / Avg: 41279.8 / Max: 41486.68 Min: 42584.9 / Avg: 42778.81 / Max: 43020.1 Min: 41214.28 / Avg: 41310.68 / Max: 41393.36 Min: 42570.07 / Avg: 42817.01 / Max: 43077.44 Min: 40623.71 / Avg: 40824.3 / Max: 40974.31 Min: 41061.35 / Avg: 41398.08 / Max: 41601.37 Min: 41279.57 / Avg: 41503.35 / Max: 41652.45 Min: 43341.6 / Avg: 43465.99 / Max: 43545.98 Min: 41733.77 / Avg: 41815.99 / Max: 41932.73 1. (CXX) g++ options: -march=native -O3 -lpthread
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 SE +/- 0.00793, N = 3 SE +/- 0.01172, N = 3 SE +/- 0.00564, N = 3 SE +/- 0.00850, N = 3 SE +/- 0.00367, N = 3 SE +/- 0.00383, N = 3 SE +/- 0.29325, N = 3 SE +/- 0.00334, N = 3 SE +/- 0.00469, N = 3 4.44237 4.44588 4.41308 4.47609 4.43760 4.42921 95.53010 4.41646 4.43864 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Matrix Multiply Batch Shapes Transformer - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 20 40 60 80 100 Min: 4.43 / Avg: 4.44 / Max: 4.46 Min: 4.42 / Avg: 4.45 / Max: 4.46 Min: 4.4 / Avg: 4.41 / Max: 4.42 Min: 4.46 / Avg: 4.48 / Max: 4.49 Min: 4.43 / Avg: 4.44 / Max: 4.44 Min: 4.42 / Avg: 4.43 / Max: 4.44 Min: 95.07 / Avg: 95.53 / Max: 96.08 Min: 4.41 / Avg: 4.42 / Max: 4.42 Min: 4.43 / Avg: 4.44 / Max: 4.45 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
CppPerformanceBenchmarks CppPerformanceBenchmarks is a set of C++ compiler performance benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Function Objects -O2 -O3 -march=native -O3 -march=native -flto 4 8 12 16 20 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 11.69 11.45 11.29 11.62 11.35 11.11 13.91 10.55 10.90 1. (CXX) g++ options: -std=c++11 -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better CppPerformanceBenchmarks 9 Test: Function Objects -O2 -O3 -march=native -O3 -march=native -flto 4 8 12 16 20 Min: 11.69 / Avg: 11.69 / Max: 11.7 Min: 11.44 / Avg: 11.45 / Max: 11.46 Min: 11.29 / Avg: 11.29 / Max: 11.3 Min: 11.62 / Avg: 11.62 / Max: 11.62 Min: 11.35 / Avg: 11.35 / Max: 11.36 Min: 11.1 / Avg: 11.11 / Max: 11.11 Min: 13.9 / Avg: 13.91 / Max: 13.93 Min: 10.54 / Avg: 10.55 / Max: 10.55 Min: 10.9 / Avg: 10.9 / Max: 10.9 1. (CXX) g++ options: -std=c++11 -O3
Basis Universal Basis Universal is a GPU texture codoec. This test times how long it takes to convert sRGB PNGs into Basis Univeral assets with various settings. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 0 -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 SE +/- 0.013, N = 3 SE +/- 0.046, N = 3 SE +/- 0.004, N = 3 SE +/- 0.087, N = 3 SE +/- 0.035, N = 3 SE +/- 0.071, N = 3 SE +/- 0.334, N = 15 SE +/- 0.089, N = 3 SE +/- 0.069, N = 3 6.592 6.711 6.526 6.655 6.552 6.651 9.746 6.552 6.671 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Basis Universal 1.12 Settings: UASTC Level 0 -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 Min: 6.57 / Avg: 6.59 / Max: 6.62 Min: 6.62 / Avg: 6.71 / Max: 6.76 Min: 6.52 / Avg: 6.53 / Max: 6.53 Min: 6.57 / Avg: 6.66 / Max: 6.83 Min: 6.49 / Avg: 6.55 / Max: 6.6 Min: 6.52 / Avg: 6.65 / Max: 6.75 Min: 8.54 / Avg: 9.75 / Max: 11.41 Min: 6.37 / Avg: 6.55 / Max: 6.65 Min: 6.54 / Avg: 6.67 / Max: 6.76 1. (CXX) g++ options: -std=c++11 -fvisibility=hidden -fPIC -fno-strict-aliasing -O3 -rdynamic -lm -lpthread
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Seek Random -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 SE +/- 0.043, N = 3 SE +/- 0.068, N = 3 SE +/- 0.093, N = 3 SE +/- 0.083, N = 3 SE +/- 0.057, N = 3 SE +/- 0.105, N = 6 SE +/- 0.172, N = 3 SE +/- 0.144, N = 3 SE +/- 0.065, N = 3 9.954 9.761 8.936 10.056 9.953 8.992 10.250 10.062 8.810 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Seek Random -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 Min: 9.88 / Avg: 9.95 / Max: 10.03 Min: 9.65 / Avg: 9.76 / Max: 9.88 Min: 8.76 / Avg: 8.94 / Max: 9.08 Min: 9.93 / Avg: 10.06 / Max: 10.21 Min: 9.86 / Avg: 9.95 / Max: 10.05 Min: 8.73 / Avg: 8.99 / Max: 9.41 Min: 9.96 / Avg: 10.25 / Max: 10.56 Min: 9.78 / Avg: 10.06 / Max: 10.25 Min: 8.71 / Avg: 8.81 / Max: 8.93 1. (CXX) g++ options: -O3 -lsnappy -lpthread
SMHasher SMHasher is a hash function tester. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: fasthash32 -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 31.94 30.72 31.96 30.35 31.68 31.88 31.98 30.82 31.91 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org cycles/hash, Fewer Is Better SMHasher 2020-02-29 Hash: fasthash32 -O2 -O3 -march=native -O3 -march=native -flto 7 14 21 28 35 Min: 31.94 / Avg: 31.94 / Max: 31.94 Min: 30.71 / Avg: 30.72 / Max: 30.74 Min: 31.96 / Avg: 31.96 / Max: 31.96 Min: 30.35 / Avg: 30.35 / Max: 30.36 Min: 31.66 / Avg: 31.68 / Max: 31.7 Min: 31.87 / Avg: 31.88 / Max: 31.88 Min: 31.98 / Avg: 31.98 / Max: 31.98 Min: 30.82 / Avg: 30.82 / Max: 30.82 Min: 31.91 / Avg: 31.91 / Max: 31.92 1. (CXX) g++ options: -march=native -O3 -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: fasthash32 -O2 -O3 -march=native -O3 -march=native -flto 2K 4K 6K 8K 10K SE +/- 31.07, N = 3 SE +/- 18.24, N = 3 SE +/- 8.50, N = 3 SE +/- 26.16, N = 3 SE +/- 3.34, N = 3 SE +/- 1.13, N = 3 SE +/- 8.85, N = 3 SE +/- 4.63, N = 3 SE +/- 18.54, N = 3 7790.76 8127.47 7833.84 8131.33 7839.73 7868.37 7847.09 8164.02 7803.94 1. (CXX) g++ options: -march=native -O3 -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MiB/sec, More Is Better SMHasher 2020-02-29 Hash: fasthash32 -O2 -O3 -march=native -O3 -march=native -flto 1400 2800 4200 5600 7000 Min: 7754.35 / Avg: 7790.76 / Max: 7852.57 Min: 8094.95 / Avg: 8127.47 / Max: 8158.04 Min: 7817.28 / Avg: 7833.84 / Max: 7845.45 Min: 8085.75 / Avg: 8131.33 / Max: 8176.37 Min: 7835.67 / Avg: 7839.73 / Max: 7846.35 Min: 7866.23 / Avg: 7868.37 / Max: 7870.09 Min: 7830.82 / Avg: 7847.09 / Max: 7861.28 Min: 8155.43 / Avg: 8164.02 / Max: 8171.32 Min: 7770.99 / Avg: 7803.94 / Max: 7835.14 1. (CXX) g++ options: -march=native -O3 -lpthread
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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Hot Read -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 SE +/- 0.057, N = 3 SE +/- 0.029, N = 3 SE +/- 0.099, N = 3 SE +/- 0.089, N = 7 SE +/- 0.102, N = 4 SE +/- 0.028, N = 3 SE +/- 0.026, N = 3 SE +/- 0.077, N = 3 SE +/- 0.091, N = 3 8.214 8.177 6.812 8.205 8.084 6.785 8.355 8.353 6.698 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Hot Read -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 Min: 8.11 / Avg: 8.21 / Max: 8.31 Min: 8.12 / Avg: 8.18 / Max: 8.21 Min: 6.67 / Avg: 6.81 / Max: 7.01 Min: 7.79 / Avg: 8.21 / Max: 8.54 Min: 7.78 / Avg: 8.08 / Max: 8.22 Min: 6.74 / Avg: 6.79 / Max: 6.84 Min: 8.3 / Avg: 8.36 / Max: 8.39 Min: 8.23 / Avg: 8.35 / Max: 8.5 Min: 6.52 / Avg: 6.7 / Max: 6.8 1. (CXX) g++ options: -O3 -lsnappy -lpthread
x265 This is a simple test of the x265 encoder run on the CPU with a sample 1080p video file. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better x265 3.1.2 H.265 1080p Video Encoding -O2 -O3 -march=native -O3 -march=native -flto 16 32 48 64 80 SE +/- 0.11, N = 3 SE +/- 0.52, N = 3 SE +/- 0.18, N = 3 SE +/- 0.25, N = 3 SE +/- 0.60, N = 3 SE +/- 0.31, N = 3 SE +/- 0.11, N = 3 SE +/- 0.21, N = 3 SE +/- 0.38, N = 3 71.67 71.80 72.61 71.45 71.07 72.19 41.78 71.97 72.19 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Frames Per Second, More Is Better x265 3.1.2 H.265 1080p Video Encoding -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 Min: 71.47 / Avg: 71.67 / Max: 71.83 Min: 71 / Avg: 71.8 / Max: 72.78 Min: 72.35 / Avg: 72.61 / Max: 72.95 Min: 71.12 / Avg: 71.45 / Max: 71.93 Min: 70.26 / Avg: 71.07 / Max: 72.25 Min: 71.62 / Avg: 72.19 / Max: 72.68 Min: 41.57 / Avg: 41.78 / Max: 41.92 Min: 71.59 / Avg: 71.97 / Max: 72.3 Min: 71.47 / Avg: 72.19 / Max: 72.74 1. (CXX) g++ options: -O3 -rdynamic -lpthread -lrt -ldl -lnuma
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 2D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 13K 26K 39K 52K 65K SE +/- 466.51, N = 7 SE +/- 560.71, N = 15 SE +/- 265.59, N = 3 SE +/- 201.62, N = 3 SE +/- 227.32, N = 3 SE +/- 865.65, N = 4 SE +/- 755.18, N = 15 SE +/- 205.61, N = 3 SE +/- 873.98, N = 15 43568 48916 47923 58281 55856 58881 56951 59445 57625 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 2D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 10K 20K 30K 40K 50K Min: 40801 / Avg: 43568.29 / Max: 44195 Min: 42795 / Avg: 48916.27 / Max: 50322 Min: 47466 / Avg: 47923.33 / Max: 48386 Min: 57920 / Avg: 58281.33 / Max: 58617 Min: 55407 / Avg: 55856.33 / Max: 56141 Min: 56337 / Avg: 58880.75 / Max: 60220 Min: 49867 / Avg: 56951.13 / Max: 59536 Min: 59083 / Avg: 59444.67 / Max: 59795 Min: 47377 / Avg: 57625.27 / Max: 61044 1. (CC) gcc options: -pthread -lm
Smallpt Smallpt is a C++ global illumination renderer written in less than 100 lines of code. Global illumination is done via unbiased Monte Carlo path tracing and there is multi-threading support via the OpenMP library. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Smallpt 1.0 Global Illumination Renderer; 128 Samples -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 SE +/- 0.022, N = 3 SE +/- 0.005, N = 3 SE +/- 0.007, N = 3 SE +/- 0.009, N = 3 SE +/- 0.005, N = 3 SE +/- 0.003, N = 3 SE +/- 0.138, N = 3 SE +/- 0.008, N = 3 SE +/- 0.007, N = 3 7.719 7.165 7.163 7.804 7.189 7.180 12.360 7.157 7.195 1. (CXX) g++ options: -fopenmp -O3 Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Seconds, Fewer Is Better Smallpt 1.0 Global Illumination Renderer; 128 Samples -O2 -O3 -march=native -O3 -march=native -flto 4 8 12 16 20 Min: 7.68 / Avg: 7.72 / Max: 7.75 Min: 7.16 / Avg: 7.17 / Max: 7.18 Min: 7.15 / Avg: 7.16 / Max: 7.18 Min: 7.79 / Avg: 7.8 / Max: 7.82 Min: 7.18 / Avg: 7.19 / Max: 7.2 Min: 7.17 / Avg: 7.18 / Max: 7.19 Min: 12.14 / Avg: 12.36 / Max: 12.61 Min: 7.14 / Avg: 7.16 / Max: 7.17 Min: 7.18 / Avg: 7.2 / Max: 7.21 1. (CXX) g++ options: -fopenmp -O3
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Fill Sync -O2 -O3 -march=native -O3 -march=native -flto 0.09 0.18 0.27 0.36 0.45 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 0.3 0.3 0.3 0.3 0.3 0.3 0.4 0.3 0.3 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Fill Sync -O2 -O3 -march=native -O3 -march=native -flto 1 2 3 4 5 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.4 / Avg: 0.4 / Max: 0.4 Min: 0.3 / Avg: 0.3 / Max: 0.3 Min: 0.3 / Avg: 0.3 / Max: 0.3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 1D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 14K 28K 42K 56K 70K SE +/- 946.90, N = 3 SE +/- 377.54, N = 3 SE +/- 239.80, N = 3 SE +/- 511.72, N = 3 SE +/- 617.23, N = 3 SE +/- 362.48, N = 3 SE +/- 791.01, N = 3 SE +/- 910.81, N = 3 SE +/- 539.26, N = 3 60106 54200 53111 63386 52670 53345 60671 63303 60764 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 1D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 11K 22K 33K 44K 55K Min: 58709 / Avg: 60106.33 / Max: 61912 Min: 53527 / Avg: 54200 / Max: 54833 Min: 52651 / Avg: 53110.67 / Max: 53459 Min: 62828 / Avg: 63386 / Max: 64408 Min: 51742 / Avg: 52670 / Max: 53839 Min: 52654 / Avg: 53345.33 / Max: 53880 Min: 59134 / Avg: 60671 / Max: 61764 Min: 61605 / Avg: 63303.33 / Max: 64723 Min: 59876 / Avg: 60763.67 / Max: 61738 1. (CC) gcc options: -pthread -lm
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.22, N = 3 SE +/- 0.03, N = 3 SE +/- 0.02, N = 3 23.17 23.16 23.16 23.25 23.16 23.17 62.82 23.20 23.17 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Convolution Batch Shapes Auto - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 Min: 23.14 / Avg: 23.17 / Max: 23.21 Min: 23.15 / Avg: 23.16 / Max: 23.17 Min: 23.15 / Avg: 23.16 / Max: 23.19 Min: 23.22 / Avg: 23.25 / Max: 23.31 Min: 23.14 / Avg: 23.16 / Max: 23.2 Min: 23.15 / Avg: 23.17 / Max: 23.19 Min: 62.39 / Avg: 62.82 / Max: 63.08 Min: 23.14 / Avg: 23.2 / Max: 23.24 Min: 23.14 / Avg: 23.17 / Max: 23.2 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -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.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Fill -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 SE +/- 0.13, N = 3 SE +/- 0.13, N = 3 SE +/- 0.09, N = 3 SE +/- 0.14, N = 3 SE +/- 0.15, N = 3 SE +/- 0.04, N = 3 SE +/- 0.16, N = 3 SE +/- 0.08, N = 3 SE +/- 0.07, N = 3 48.84 48.83 48.60 48.56 48.62 47.87 61.69 48.99 48.01 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Random Fill -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 Min: 48.58 / Avg: 48.84 / Max: 49.01 Min: 48.64 / Avg: 48.83 / Max: 49.09 Min: 48.44 / Avg: 48.6 / Max: 48.74 Min: 48.33 / Avg: 48.56 / Max: 48.81 Min: 48.37 / Avg: 48.62 / Max: 48.88 Min: 47.82 / Avg: 47.87 / Max: 47.95 Min: 61.36 / Avg: 61.69 / Max: 61.91 Min: 48.84 / Avg: 48.99 / Max: 49.11 Min: 47.93 / Avg: 48.01 / Max: 48.14 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Random Fill -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 SE +/- 0.12, N = 3 SE +/- 0.12, N = 3 SE +/- 0.06, N = 3 SE +/- 0.12, N = 3 SE +/- 0.15, N = 3 SE +/- 0.03, N = 3 SE +/- 0.09, N = 3 SE +/- 0.06, N = 3 SE +/- 0.07, N = 3 45.3 45.2 45.5 45.5 45.5 46.2 35.8 45.1 46.0 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Random Fill -O2 -O3 -march=native -O3 -march=native -flto 9 18 27 36 45 Min: 45.1 / Avg: 45.27 / Max: 45.5 Min: 45 / Avg: 45.23 / Max: 45.4 Min: 45.4 / Avg: 45.5 / Max: 45.6 Min: 45.3 / Avg: 45.5 / Max: 45.7 Min: 45.2 / Avg: 45.47 / Max: 45.7 Min: 46.1 / Avg: 46.17 / Max: 46.2 Min: 35.7 / Avg: 35.83 / Max: 36 Min: 45 / Avg: 45.1 / Max: 45.2 Min: 45.9 / Avg: 46.03 / Max: 46.1 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Overwrite -O2 -O3 -march=native -O3 -march=native -flto 14 28 42 56 70 SE +/- 0.09, N = 3 SE +/- 0.10, N = 3 SE +/- 0.02, N = 3 SE +/- 0.14, N = 3 SE +/- 0.18, N = 3 SE +/- 0.08, N = 3 SE +/- 0.23, N = 3 SE +/- 0.05, N = 3 SE +/- 0.37, N = 3 48.83 48.71 48.26 48.28 48.96 47.94 62.02 48.88 48.16 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Microseconds Per Op, Fewer Is Better LevelDB 1.22 Benchmark: Overwrite -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 Min: 48.73 / Avg: 48.83 / Max: 49.02 Min: 48.53 / Avg: 48.71 / Max: 48.89 Min: 48.21 / Avg: 48.26 / Max: 48.29 Min: 47.99 / Avg: 48.27 / Max: 48.47 Min: 48.73 / Avg: 48.96 / Max: 49.32 Min: 47.79 / Avg: 47.94 / Max: 48.08 Min: 61.58 / Avg: 62.02 / Max: 62.34 Min: 48.78 / Avg: 48.88 / Max: 48.97 Min: 47.7 / Avg: 48.16 / Max: 48.89 1. (CXX) g++ options: -O3 -lsnappy -lpthread
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Overwrite -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 SE +/- 0.09, N = 3 SE +/- 0.09, N = 3 SE +/- 0.00, N = 3 SE +/- 0.12, N = 3 SE +/- 0.19, N = 3 SE +/- 0.06, N = 3 SE +/- 0.12, N = 3 SE +/- 0.06, N = 3 SE +/- 0.34, N = 3 45.3 45.3 45.8 45.8 45.2 46.1 35.6 45.2 45.9 1. (CXX) g++ options: -O3 -lsnappy -lpthread Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org MB/s, More Is Better LevelDB 1.22 Benchmark: Overwrite -O2 -O3 -march=native -O3 -march=native -flto 9 18 27 36 45 Min: 45.1 / Avg: 45.27 / Max: 45.4 Min: 45.2 / Avg: 45.33 / Max: 45.5 Min: 45.8 / Avg: 45.8 / Max: 45.8 Min: 45.6 / Avg: 45.77 / Max: 46 Min: 44.8 / Avg: 45.17 / Max: 45.4 Min: 46 / Avg: 46.1 / Max: 46.2 Min: 35.4 / Avg: 35.57 / Max: 35.8 Min: 45.1 / Avg: 45.2 / Max: 45.3 Min: 45.2 / Avg: 45.87 / Max: 46.3 1. (CXX) g++ options: -O3 -lsnappy -lpthread
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 1D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K SE +/- 42.57, N = 3 SE +/- 107.82, N = 3 SE +/- 224.46, N = 13 SE +/- 255.86, N = 3 SE +/- 104.64, N = 3 SE +/- 292.69, N = 15 SE +/- 184.00, N = 3 SE +/- 311.73, N = 3 17895 19590 18953 20559 20299 19725 19239 20257 19467 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Float + SSE - Size: 1D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K Min: 17828 / Avg: 17895 / Max: 17974 Min: 19375 / Avg: 19590.33 / Max: 19708 Min: 18020 / Avg: 20558.69 / Max: 21016 Min: 19963 / Avg: 20298.67 / Max: 20801 Min: 19516 / Avg: 19725 / Max: 19839 Min: 16352 / Avg: 19239.27 / Max: 20323 Min: 20072 / Avg: 20257 / Max: 20625 Min: 19093 / Avg: 19467 / Max: 20086 1. (CC) gcc options: -pthread -lm
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 1D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K SE +/- 42.83, N = 3 SE +/- 104.46, N = 3 SE +/- 57.50, N = 3 SE +/- 19.70, N = 3 SE +/- 20.00, N = 3 SE +/- 68.34, N = 3 SE +/- 10.37, N = 3 SE +/- 56.62, N = 3 SE +/- 122.97, N = 3 9740.0 11966.0 12186.0 10162.0 11608.0 12095.0 9744.9 11430.0 12317.0 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 1D FFT Size 4096 -O2 -O3 -march=native -O3 -march=native -flto 2K 4K 6K 8K 10K Min: 9669.5 / Avg: 9740 / Max: 9817.4 Min: 11759 / Avg: 11966 / Max: 12094 Min: 12071 / Avg: 12186 / Max: 12244 Min: 10123 / Avg: 10162.33 / Max: 10184 Min: 11584 / Avg: 11608.33 / Max: 11648 Min: 11959 / Avg: 12095.33 / Max: 12172 Min: 9733.1 / Avg: 9744.93 / Max: 9765.6 Min: 11341 / Avg: 11429.67 / Max: 11535 Min: 12078 / Avg: 12317.33 / Max: 12486 1. (CC) gcc options: -pthread -lm
LAMMPS Molecular Dynamics Simulator LAMMPS is a classical molecular dynamics code, and an acronym for Large-scale Atomic/Molecular Massively Parallel Simulator. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ns/day, More Is Better LAMMPS Molecular Dynamics Simulator 9Jan2020 Model: Rhodopsin Protein -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 SE +/- 0.116, N = 3 SE +/- 0.100, N = 3 SE +/- 0.101, N = 5 SE +/- 0.128, N = 3 SE +/- 0.066, N = 3 SE +/- 0.063, N = 3 SE +/- 0.088, N = 3 SE +/- 0.219, N = 12 SE +/- 0.052, N = 3 8.326 8.449 8.639 8.511 8.830 8.806 8.217 8.698 8.818 1. (CXX) g++ options: -O3 -rdynamic -ljpeg -lpng -lz -lfftw3 -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ns/day, More Is Better LAMMPS Molecular Dynamics Simulator 9Jan2020 Model: Rhodopsin Protein -O2 -O3 -march=native -O3 -march=native -flto 3 6 9 12 15 Min: 8.1 / Avg: 8.33 / Max: 8.47 Min: 8.31 / Avg: 8.45 / Max: 8.64 Min: 8.28 / Avg: 8.64 / Max: 8.85 Min: 8.26 / Avg: 8.51 / Max: 8.69 Min: 8.71 / Avg: 8.83 / Max: 8.93 Min: 8.71 / Avg: 8.81 / Max: 8.92 Min: 8.07 / Avg: 8.22 / Max: 8.37 Min: 6.31 / Avg: 8.7 / Max: 9.08 Min: 8.76 / Avg: 8.82 / Max: 8.92 1. (CXX) g++ options: -O3 -rdynamic -ljpeg -lpng -lz -lfftw3 -lm
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 1D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K SE +/- 13.91, N = 3 SE +/- 69.22, N = 3 SE +/- 17.34, N = 3 SE +/- 237.03, N = 3 SE +/- 177.60, N = 3 SE +/- 30.23, N = 3 SE +/- 226.75, N = 4 SE +/- 158.55, N = 3 13388 16506 13087 13122 15594 13282 12530 16126 12811 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 1D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K Min: 16478 / Avg: 16505.67 / Max: 16522 Min: 12949 / Avg: 13087.33 / Max: 13161 Min: 13087 / Avg: 13121.67 / Max: 13140 Min: 15123 / Avg: 15594.33 / Max: 15874 Min: 12927 / Avg: 13281.67 / Max: 13476 Min: 12470 / Avg: 12529.67 / Max: 12568 Min: 15452 / Avg: 16125.75 / Max: 16397 Min: 12541 / Avg: 12811 / Max: 13090 1. (CC) gcc options: -pthread -lm
oneDNN This is a test of the Intel oneDNN as an Intel-optimized library for Deep Neural Networks and making use of its built-in benchdnn functionality. The result is the total perf time reported. Intel oneDNN was formerly known as DNNL (Deep Neural Network Library) and MKL-DNN before being rebranded as part of the oneAPI initiative. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Deconvolution Batch deconv_3d - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 12 24 36 48 60 SE +/- 0.00536, N = 3 SE +/- 0.00272, N = 3 SE +/- 0.00941, N = 3 SE +/- 0.00826, N = 3 SE +/- 0.00535, N = 3 SE +/- 0.01120, N = 3 SE +/- 0.17447, N = 3 SE +/- 0.01344, N = 3 SE +/- 0.00098, N = 3 4.77958 4.77613 4.80401 4.78777 4.78115 4.79115 52.37190 4.78543 4.78818 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org ms, Fewer Is Better oneDNN 1.5 Harness: Deconvolution Batch deconv_3d - Data Type: f32 - Engine: CPU -O2 -O3 -march=native -O3 -march=native -flto 10 20 30 40 50 Min: 4.77 / Avg: 4.78 / Max: 4.79 Min: 4.77 / Avg: 4.78 / Max: 4.78 Min: 4.79 / Avg: 4.8 / Max: 4.82 Min: 4.77 / Avg: 4.79 / Max: 4.8 Min: 4.77 / Avg: 4.78 / Max: 4.79 Min: 4.78 / Avg: 4.79 / Max: 4.81 Min: 52.11 / Avg: 52.37 / Max: 52.7 Min: 4.77 / Avg: 4.79 / Max: 4.81 Min: 4.79 / Avg: 4.79 / Max: 4.79 1. (CXX) g++ options: -O3 -march=native -std=c++11 -fopenmp -msse4.1 -fPIC -pie -lpthread -ldl
FFTW FFTW is a C subroutine library for computing the discrete Fourier transform (DFT) in one or more dimensions. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 2D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 4K 8K 12K 16K 20K SE +/- 61.06, N = 3 SE +/- 53.20, N = 3 SE +/- 78.45, N = 3 SE +/- 6.33, N = 3 SE +/- 72.51, N = 3 SE +/- 14.74, N = 3 SE +/- 14.77, N = 3 SE +/- 70.00, N = 3 SE +/- 12.49, N = 3 13413 16680 16223 13661 16046 16314 12990 16885 16146 1. (CC) gcc options: -pthread -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.6 Build: Stock - Size: 2D FFT Size 32 -O2 -O3 -march=native -O3 -march=native -flto 3K 6K 9K 12K 15K Min: 13340 / Avg: 13412.67 / Max: 13534 Min: 16585 / Avg: 16680 / Max: 16769 Min: 16076 / Avg: 16223 / Max: 16344 Min: 13648 / Avg: 13660.67 / Max: 13667 Min: 15918 / Avg: 16046.33 / Max: 16169 Min: 16286 / Avg: 16314 / Max: 16336 Min: 12961 / Avg: 12990.33 / Max: 13008 Min: 16814 / Avg: 16885 / Max: 17025 Min: 16122 / Avg: 16146 / Max: 16164 1. (CC) gcc options: -pthread -lm
BLAKE2 This is a benchmark of BLAKE2 using the blake2s binary. BLAKE2 is a high-performance crypto alternative to MD5 and SHA-2/3. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Cycles Per Byte, Fewer Is Better BLAKE2 20170307 -O2 -O3 -march=native -O3 -march=native -flto 0.9428 1.8856 2.8284 3.7712 4.714 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 4.19 4.19 4.19 4.19 4.19 4.03 4.17 4.17 4.02 1. (CC) gcc options: -O3 -march=native -lcrypto -lz Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Cycles Per Byte, Fewer Is Better BLAKE2 20170307 -O2 -O3 -march=native -O3 -march=native -flto 2 4 6 8 10 Min: 4.18 / Avg: 4.19 / Max: 4.19 Min: 4.18 / Avg: 4.19 / Max: 4.19 Min: 4.18 / Avg: 4.19 / Max: 4.2 Min: 4.18 / Avg: 4.19 / Max: 4.19 Min: 4.18 / Avg: 4.19 / Max: 4.2 Min: 4.03 / Avg: 4.03 / Max: 4.03 Min: 4.17 / Avg: 4.17 / Max: 4.17 Min: 4.17 / Avg: 4.17 / Max: 4.18 Min: 4.01 / Avg: 4.02 / Max: 4.02 1. (CC) gcc options: -O3 -march=native -lcrypto -lz
SciMark This test runs the ANSI C version of SciMark 2.0, which is a benchmark for scientific and numerical computing developed by programmers at the National Institute of Standards and Technology. This test is made up of Fast Foruier Transform, Jacobi Successive Over-relaxation, Monte Carlo, Sparse Matrix Multiply, and dense LU matrix factorization benchmarks. Learn more via the OpenBenchmarking.org test page.
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Jacobi Successive Over-Relaxation -O2 -O3 -march=native -O3 -march=native -flto 600 1200 1800 2400 3000 SE +/- 0.26, N = 3 SE +/- 0.58, N = 3 SE +/- 1.02, N = 3 SE +/- 0.22, N = 3 SE +/- 8.81, N = 3 SE +/- 0.38, N = 3 SE +/- 0.52, N = 3 SE +/- 1.11, N = 3 SE +/- 0.61, N = 3 1471.97 2580.59 2577.65 1472.00 2572.51 2661.24 1359.41 2473.80 2553.09 1. (CC) gcc options: -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Jacobi Successive Over-Relaxation -O2 -O3 -march=native -O3 -march=native -flto 500 1000 1500 2000 2500 Min: 1471.62 / Avg: 1471.97 / Max: 1472.47 Min: 2579.45 / Avg: 2580.59 / Max: 2581.32 Min: 2575.61 / Avg: 2577.65 / Max: 2578.69 Min: 1471.56 / Avg: 1472 / Max: 1472.24 Min: 2554.88 / Avg: 2572.51 / Max: 2581.43 Min: 2660.79 / Avg: 2661.24 / Max: 2662 Min: 1358.88 / Avg: 1359.41 / Max: 1360.45 Min: 2471.59 / Avg: 2473.8 / Max: 2475 Min: 2552.42 / Avg: 2553.09 / Max: 2554.32 1. (CC) gcc options: -lm
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Dense LU Matrix Factorization -O2 -O3 -march=native -O3 -march=native -flto 2K 4K 6K 8K 10K SE +/- 78.81, N = 3 SE +/- 104.73, N = 3 SE +/- 15.57, N = 3 SE +/- 30.81, N = 3 SE +/- 1.61, N = 3 SE +/- 10.95, N = 3 SE +/- 2.02, N = 3 SE +/- 1.48, N = 3 SE +/- 4.00, N = 3 4633.35 8723.22 8278.96 5041.29 9203.70 8438.95 2709.63 9893.07 8604.08 1. (CC) gcc options: -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Dense LU Matrix Factorization -O2 -O3 -march=native -O3 -march=native -flto 2K 4K 6K 8K 10K Min: 4475.78 / Avg: 4633.35 / Max: 4715.66 Min: 8513.78 / Avg: 8723.22 / Max: 8829.42 Min: 8254.23 / Avg: 8278.96 / Max: 8307.71 Min: 4979.68 / Avg: 5041.29 / Max: 5073.3 Min: 9200.49 / Avg: 9203.7 / Max: 9205.44 Min: 8417.93 / Avg: 8438.95 / Max: 8454.75 Min: 2707.46 / Avg: 2709.63 / Max: 2713.66 Min: 9890.41 / Avg: 9893.07 / Max: 9895.51 Min: 8597.26 / Avg: 8604.08 / Max: 8611.11 1. (CC) gcc options: -lm
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Sparse Matrix Multiply -O2 -O3 -march=native -O3 -march=native -flto 900 1800 2700 3600 4500 SE +/- 2.80, N = 3 SE +/- 21.07, N = 3 SE +/- 10.21, N = 3 SE +/- 23.38, N = 3 SE +/- 117.60, N = 3 SE +/- 10.64, N = 3 SE +/- 3.50, N = 3 SE +/- 13.33, N = 3 SE +/- 10.45, N = 3 3764.14 3879.29 2995.23 3772.37 3841.79 3387.40 2436.74 3858.10 4034.24 1. (CC) gcc options: -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Sparse Matrix Multiply -O2 -O3 -march=native -O3 -march=native -flto 700 1400 2100 2800 3500 Min: 3758.61 / Avg: 3764.14 / Max: 3767.64 Min: 3855.41 / Avg: 3879.29 / Max: 3921.31 Min: 2980.49 / Avg: 2995.23 / Max: 3014.85 Min: 3725.84 / Avg: 3772.37 / Max: 3799.58 Min: 3607.2 / Avg: 3841.79 / Max: 3973.86 Min: 3370.34 / Avg: 3387.4 / Max: 3406.95 Min: 2431.16 / Avg: 2436.74 / Max: 2443.2 Min: 3831.53 / Avg: 3858.1 / Max: 3873.34 Min: 4015.08 / Avg: 4034.24 / Max: 4051.05 1. (CC) gcc options: -lm
Result
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Fast Fourier Transform -O2 -O3 -march=native -O3 -march=native -flto 200 400 600 800 1000 SE +/- 80.06, N = 3 SE +/- 40.98, N = 3 SE +/- 1.02, N = 3 SE +/- 3.18, N = 3 SE +/- 66.66, N = 3 SE +/- 3.27, N = 3 SE +/- 3.50, N = 3 SE +/- 4.09, N = 3 SE +/- 6.50, N = 3 773.56 823.32 880.16 845.68 815.13 897.85 675.51 857.59 900.25 1. (CC) gcc options: -lm Result Confidence
GCC 8.4 GCC 9.3 GCC 10.2 OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Fast Fourier Transform -O2 -O3 -march=native -O3 -march=native -flto 160 320 480 640 800 Min: 613.73 / Avg: 773.56 / Max: 861.85 Min: 741.45 / Avg: 823.32 / Max: 867.43 Min: 878.25 / Avg: 880.16 / Max: 881.74 Min: 840.32 / Avg: 845.68 / Max: 851.32 Min: 681.85 / Avg: 815.13 / Max: 884.92 Min: 893.81 / Avg: 897.85 / Max: 904.32 Min: 668.57 / Avg: 675.51 / Max: 679.72 Min: 851.49 / Avg: 857.59 / Max: 865.35 Min: 892.59 / Avg: 900.25 / Max: 913.17 1. (CC) gcc options: -lm
GCC 8.4: -O2 Environment Notes: CXXFLAGS=-O2 CFLAGS=-O2Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 20 July 2020 17:25 by user system.
GCC 8.4: -O3 -march=native Environment Notes: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 21 July 2020 06:10 by user system.
GCC 8.4: -O3 -march=native -flto Processor: Intel Core i9-10900K @ 5.30GHz (10 Cores / 20 Threads), Motherboard: Gigabyte Z490 AORUS MASTER (F3 BIOS), Chipset: Intel Comet Lake PCH, Memory: 16GB, Disk: Samsung SSD 970 EVO 250GB, Graphics: Gigabyte AMD Radeon RX 5500/5500M / Pro 5500M 8GB (1890/875MHz), Audio: Realtek ALC1220, Monitor: DELL P2415Q, Network: Intel Device 15f3 + Intel Wi-Fi 6 AX201
OS: Ubuntu 20.04, Kernel: 5.8.0-050800daily20200707-generic (x86_64), Desktop: GNOME Shell 3.36.3, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-ef67218 2020-07-07 focal-oibaf-ppa) (LLVM 10.0.1), OpenCL: OpenCL 1.1 Mesa 20.2.0-devel (git-c977567db6), Vulkan: 1.2.131, Compiler: GCC 8.4.0, File-System: ext4, Screen Resolution: 3840x2160
Environment Notes: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 21 July 2020 18:31 by user system.
GCC 9.3: -O2 Environment Notes: CXXFLAGS=-O2 CFLAGS=-O2Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 19 July 2020 06:02 by user system.
GCC 9.3: -O3 -march=native Environment Notes: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 19 July 2020 17:45 by user system.
GCC 9.3: -O3 -march=native -flto Processor: Intel Core i9-10900K @ 5.30GHz (10 Cores / 20 Threads), Motherboard: Gigabyte Z490 AORUS MASTER (F3 BIOS), Chipset: Intel Comet Lake PCH, Memory: 16GB, Disk: Samsung SSD 970 EVO 250GB, Graphics: Gigabyte AMD Radeon RX 5500/5500M / Pro 5500M 8GB (1890/875MHz), Audio: Realtek ALC1220, Monitor: DELL P2415Q, Network: Intel Device 15f3 + Intel Wi-Fi 6 AX201
OS: Ubuntu 20.04, Kernel: 5.8.0-050800daily20200707-generic (x86_64), Desktop: GNOME Shell 3.36.3, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-ef67218 2020-07-07 focal-oibaf-ppa) (LLVM 10.0.1), OpenCL: OpenCL 1.1 Mesa 20.2.0-devel (git-c977567db6), Vulkan: 1.2.131, Compiler: GCC 9.3.0, File-System: ext4, Screen Resolution: 3840x2160
Environment Notes: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 20 July 2020 07:02 by user system.
GCC 10.2: -O2 Environment Notes: CXXFLAGS=-O2 CFLAGS=-O2Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 16 July 2020 17:04 by user system.
GCC 10.2: -O3 -march=native Environment Notes: CXXFLAGS="-O3 -march=native" CFLAGS="-O3 -march=native"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 17 July 2020 21:04 by user system.
GCC 10.2: -O3 -march=native -flto Processor: Intel Core i9-10900K @ 5.30GHz (10 Cores / 20 Threads), Motherboard: Gigabyte Z490 AORUS MASTER (F3 BIOS), Chipset: Intel Comet Lake PCH, Memory: 16GB, Disk: Samsung SSD 970 EVO 250GB, Graphics: Gigabyte AMD Radeon RX 5500/5500M / Pro 5500M 8GB (1890/875MHz), Audio: Realtek ALC1220, Monitor: DELL P2415Q, Network: Intel Device 15f3 + Intel Wi-Fi 6 AX201
OS: Ubuntu 20.04, Kernel: 5.8.0-050800daily20200707-generic (x86_64), Desktop: GNOME Shell 3.36.3, Display Server: X Server 1.20.8, Display Driver: modesetting 1.20.8, OpenGL: 4.6 Mesa 20.2.0-devel (git-ef67218 2020-07-07 focal-oibaf-ppa) (LLVM 10.0.1), OpenCL: OpenCL 1.1 Mesa 20.2.0-devel (git-c977567db6), Vulkan: 1.2.131, Compiler: GCC 10.1.1 20200715, File-System: ext4, Screen Resolution: 3840x2160
Environment Notes: CXXFLAGS="-O3 -march=native -flto" CFLAGS="-O3 -march=native -flto"Compiler Notes: --disable-multilib --enable-checking-releaseProcessor Notes: Scaling Governor: intel_pstate powersave - CPU Microcode: 0xc8Python Notes: Python 3.8.2Security Notes: itlb_multihit: KVM: Mitigation of Split huge pages + l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of usercopy/swapgs barriers and __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling + srbds: Not affected + tsx_async_abort: Not affected
Testing initiated at 18 July 2020 10:16 by user system.