GCC vs. LLVM Clang 3.8 3.9 Compiler Benchmarking Benchmarks by Michael Larabel for a future article on phoronix looking at early GCC 7 compiler performance compared to GCC 6 and GCC 5 and then LLVM Clang 3.8 and Clang 3.9.
HTML result view exported from: https://openbenchmarking.org/result/1609139-LO-GCCCLANG151&grs&sro .
GCC vs. LLVM Clang 3.8 3.9 Compiler Benchmarking Processor Motherboard Chipset Memory Disk Graphics Audio Network OS Kernel Desktop Display Server Display Driver OpenGL Compiler File-System Screen Resolution GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 Clang 3.8.0 Clang 3.9.0 Intel Xeon E5-2609 v4 @ 1.70GHz (8 Cores) MSI X99A WORKSTATION (MS-7A54) v1.0 Intel Xeon E7 v4/Xeon 16384MB 3 x 120GB TOSHIBA-TR150 LLVMpipe Realtek ALC1150 Intel Connection Ubuntu 16.04 4.8.0-999-generic (x86_64) 20160908 Unity 7.4.0 X Server 1.18.3 modesetting 1.18.3 3.3 Mesa 11.2.0 Gallium 0.4 GCC 5.4.0 ext4 1024x768 GCC 6.2.0 GCC 7.0.0 20160904 Clang 3.8.0-2ubuntu4 Clang 3.9.0-svn279689-1~exp1 OpenBenchmarking.org Environment Details - LIBGL_ALWAYS_SOFTWARE=1 Compiler Details - GCC 5.4.0, GCC 6.2.0, GCC 7.0.0 20160904: --disable-multilib --enable-checking=release --enable-languages=c,c++,fortran Disk Details - CFQ / data=ordered,errors=remount-ro,relatime,rw Processor Details - Scaling Governor: intel_pstate powersave
GCC vs. LLVM Clang 3.8 3.9 Compiler Benchmarking scimark2: Monte Carlo scimark2: Dense LU Matrix Factorization c-ray: Total Time scimark2: Composite smallpt: Global Illumination Renderer; 100 Samples scimark2: Jacobi Successive Over-Relaxation himeno: Poisson Pressure Solver build-imagemagick: Time To Compile scimark2: Sparse Matrix Multiply build-php: Time To Compile hint: FLOAT encode-mp3: WAV To MP3 lammps: Rhodopsin Protein hmmer: Pfam Database Search ebizzy: Phoronix Test Suite v6.6.0 openssl: RSA 4096-bit Performance fftw: Float + SSE - 2D FFT Size 4096 n-queens: Elapsed Time fhourstones: Complex Connect-4 Solving bullet: 3000 Fall apache: Static Web Page Serving bullet: 136 Ragdolls scimark2: Fast Fourier Transform redis: GET fftw: Float + SSE - 1D FFT Size 4096 redis: SET bullet: Raytests pgbench: Buffer Test - Normal Load - Read Write encode-flac: WAV To FLAC pgbench: Buffer Test - Single Thread - Read Write mafft: Multiple Sequence Alignment GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 Clang 3.8.0 Clang 3.9.0 304.10 1327.15 21.30 695.70 47 571.65 893.82 67.21 1047.40 35.31 177738053.32 22.84 72.00 12.18 155360 569.37 7410.50 57.53 6954.67 9.13 27564.72 6.45 228.18 1434042.92 10966 969029.10 5.94 4468.82 13.45 552.55 5.99 304.07 1302.74 21.25 738.21 47 572.05 1091.32 94.51 1280.02 36.25 166219905.04 22.42 12.18 155273 569.27 7515.98 56.21 6884.93 9.11 28383.22 6.46 232.19 1408567.41 10936 972159.31 5.98 4411.75 13.44 546.21 6.41 304.03 1596.90 23.36 800.64 30 572.06 1095.90 81.27 1290.94 36.52 165733740.33 22.03 12.22 155692 570.20 7517.94 52.54 6905.30 9.01 29512.13 6.40 239.27 1393442.96 10951 965269.60 5.96 4437.25 13.45 549.49 6.45 126.46 2936.85 39.34 1111.78 857.42 748.09 74.34 1414.48 28.45 141990405.69 26.94 63.22 12.37 152747 566.80 6841.34 7164.37 9.57 29812.18 6.89 223.71 1342892.96 10357 925939.11 6.20 4411.72 13.26 538.44 7.74 126.43 2917.96 44.42 1115.00 856.35 841.55 78.45 1448.22 35.25 143252876.20 27.46 62.30 13.76 138671 511.07 6923.64 6563.47 9.83 27344.42 6.92 226.01 1348061.04 10331 933433.87 6.21 4356.38 13.15 531.79 7.27 OpenBenchmarking.org
SciMark Computational Test: Monte Carlo OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Monte Carlo Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 70 140 210 280 350 SE +/- 0.02, N = 4 SE +/- 0.15, N = 4 SE +/- 0.00, N = 4 SE +/- 0.00, N = 4 SE +/- 0.00, N = 4 126.46 126.43 304.10 304.07 304.03 1. (CXX) g++ options: -O3 -march=native
SciMark Computational Test: Dense LU Matrix Factorization OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Dense LU Matrix Factorization Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 600 1200 1800 2400 3000 SE +/- 0.55, N = 4 SE +/- 1.74, N = 4 SE +/- 9.92, N = 4 SE +/- 0.23, N = 4 SE +/- 0.14, N = 4 2936.85 2917.96 1327.15 1302.74 1596.90 1. (CXX) g++ options: -O3 -march=native
C-Ray Total Time OpenBenchmarking.org Seconds, Fewer Is Better C-Ray 1.1 Total Time Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 10 20 30 40 50 SE +/- 0.08, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 39.34 44.42 21.30 21.25 23.36 1. (CC) gcc options: -lm -lpthread -O3 -march=native
SciMark Computational Test: Composite OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Composite Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 200 400 600 800 1000 SE +/- 0.09, N = 4 SE +/- 0.69, N = 4 SE +/- 2.13, N = 4 SE +/- 0.14, N = 4 SE +/- 0.09, N = 4 1111.78 1115.00 695.70 738.21 800.64 1. (CXX) g++ options: -O3 -march=native
Smallpt Global Illumination Renderer; 100 Samples OpenBenchmarking.org Seconds, Fewer Is Better Smallpt 1.0 Global Illumination Renderer; 100 Samples GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 11 22 33 44 55 SE +/- 0.33, N = 3 SE +/- 0.88, N = 3 SE +/- 0.00, N = 3 47 47 30 1. (CXX) g++ options: -fopenmp -O3 -march=native
SciMark Computational Test: Jacobi Successive Over-Relaxation OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Jacobi Successive Over-Relaxation Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 200 400 600 800 1000 SE +/- 0.01, N = 4 SE +/- 0.04, N = 4 SE +/- 0.41, N = 4 SE +/- 0.00, N = 4 SE +/- 0.01, N = 4 857.42 856.35 571.65 572.05 572.06 1. (CXX) g++ options: -O3 -march=native
Himeno Benchmark Poisson Pressure Solver OpenBenchmarking.org MFLOPS, More Is Better Himeno Benchmark 3.0 Poisson Pressure Solver Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 200 400 600 800 1000 SE +/- 0.07, N = 3 SE +/- 1.07, N = 3 SE +/- 0.49, N = 3 SE +/- 1.34, N = 3 SE +/- 0.65, N = 3 748.09 841.55 893.82 1091.32 1095.90 1. (CC) gcc options: -O3 -march=native -mavx2
Timed ImageMagick Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed ImageMagick Compilation 6.9.0 Time To Compile Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 20 40 60 80 100 SE +/- 0.05, N = 3 SE +/- 0.22, N = 3 SE +/- 0.01, N = 3 SE +/- 0.17, N = 3 SE +/- 0.14, N = 3 74.34 78.45 67.21 94.51 81.27
SciMark Computational Test: Sparse Matrix Multiply OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Sparse Matrix Multiply Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 300 600 900 1200 1500 SE +/- 0.32, N = 4 SE +/- 2.31, N = 4 SE +/- 2.22, N = 4 SE +/- 0.17, N = 4 SE +/- 0.35, N = 4 1414.48 1448.22 1047.40 1280.02 1290.94 1. (CXX) g++ options: -O3 -march=native
Timed PHP Compilation Time To Compile OpenBenchmarking.org Seconds, Fewer Is Better Timed PHP Compilation 5.2.9 Time To Compile Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 8 16 24 32 40 SE +/- 0.17, N = 3 SE +/- 0.40, N = 3 SE +/- 0.06, N = 3 SE +/- 0.09, N = 3 SE +/- 0.05, N = 3 28.45 35.25 35.31 36.25 36.52 1. (CC) gcc options: -O3 -march=native -pedantic -ldl -lz -lm
Hierarchical INTegration Test: FLOAT OpenBenchmarking.org QUIPs, More Is Better Hierarchical INTegration 1.0 Test: FLOAT Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 40M 80M 120M 160M 200M SE +/- 15968.48, N = 3 SE +/- 131098.46, N = 3 SE +/- 355658.03, N = 3 SE +/- 165686.10, N = 3 SE +/- 291268.73, N = 3 141990405.69 143252876.20 177738053.32 166219905.04 165733740.33 1. (CC) gcc options: -O3 -march=native -lm
LAME MP3 Encoding WAV To MP3 OpenBenchmarking.org Seconds, Fewer Is Better LAME MP3 Encoding 3.99.3 WAV To MP3 Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 6 12 18 24 30 SE +/- 0.04, N = 5 SE +/- 0.04, N = 5 SE +/- 0.01, N = 5 SE +/- 0.03, N = 5 SE +/- 0.01, N = 5 26.94 27.46 22.84 22.42 22.03 1. (CC) gcc options: -O3 -ffast-math -funroll-loops -fschedule-insns2 -fbranch-count-reg -fforce-addr -pipe -march=native -lm
LAMMPS Molecular Dynamics Simulator Test: Rhodopsin Protein OpenBenchmarking.org Loop Time, Fewer Is Better LAMMPS Molecular Dynamics Simulator 1.0 Test: Rhodopsin Protein Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 16 32 48 64 80 SE +/- 0.14, N = 3 SE +/- 0.09, N = 3 SE +/- 0.12, N = 3 63.22 62.30 72.00 1. (CXX) g++ options: -lfftw -lmpich
Timed HMMer Search Pfam Database Search OpenBenchmarking.org Seconds, Fewer Is Better Timed HMMer Search 2.3.2 Pfam Database Search Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 4 8 12 16 20 SE +/- 0.02, N = 3 SE +/- 0.04, N = 3 SE +/- 0.03, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 12.37 13.76 12.18 12.18 12.22 1. (CC) gcc options: -O3 -march=native -pthread -lhmmer -lsquid -lm
ebizzy Phoronix Test Suite v6.6.0 OpenBenchmarking.org Records/s, More Is Better ebizzy 0.3 Phoronix Test Suite v6.6.0 Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 30K 60K 90K 120K 150K SE +/- 2308.58, N = 4 SE +/- 392.88, N = 3 SE +/- 87.83, N = 3 SE +/- 68.34, N = 3 SE +/- 360.05, N = 3 152747 138671 155360 155273 155692 1. (CC) gcc options: -pthread -lpthread -O3 -march=native
OpenSSL RSA 4096-bit Performance OpenBenchmarking.org Signs Per Second, More Is Better OpenSSL 1.0.1g RSA 4096-bit Performance Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 120 240 360 480 600 SE +/- 1.57, N = 3 SE +/- 2.68, N = 3 SE +/- 0.58, N = 3 SE +/- 1.31, N = 3 SE +/- 0.25, N = 3 566.80 511.07 569.37 569.27 570.20 1. (CC) gcc options: -m64 -O3 -lssl -lcrypto -ldl
FFTW Build: Float + SSE - Size: 2D FFT Size 4096 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.4 Build: Float + SSE - Size: 2D FFT Size 4096 Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 1600 3200 4800 6400 8000 SE +/- 36.82, N = 5 SE +/- 16.42, N = 5 SE +/- 43.96, N = 5 SE +/- 22.70, N = 5 SE +/- 15.43, N = 5 6841.34 6923.64 7410.50 7515.98 7517.94 1. (CC) gcc options: -O3 -march=native -lm
N-Queens Elapsed Time OpenBenchmarking.org Seconds, Fewer Is Better N-Queens 1.0 Elapsed Time GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 13 26 39 52 65 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 57.53 56.21 52.54 1. (CC) gcc options: -static -fopenmp -O3 -march=native
Fhourstones Complex Connect-4 Solving OpenBenchmarking.org Kpos / sec, More Is Better Fhourstones 3.1 Complex Connect-4 Solving Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 1500 3000 4500 6000 7500 SE +/- 0.87, N = 3 SE +/- 2.59, N = 3 SE +/- 5.88, N = 3 SE +/- 11.54, N = 3 SE +/- 10.00, N = 3 7164.37 6563.47 6954.67 6884.93 6905.30 1. (CC) gcc options: -O3
Bullet Physics Engine Test: 3000 Fall OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: 3000 Fall Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 3 6 9 12 15 SE +/- 0.02, N = 3 SE +/- 0.02, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 SE +/- 0.00, N = 3 9.57 9.83 9.13 9.11 9.01 1. (CXX) g++ options: -O3 -march=native -rdynamic -lglut -lGL -lGLU
Apache Benchmark Static Web Page Serving OpenBenchmarking.org Requests Per Second, More Is Better Apache Benchmark 2.4.7 Static Web Page Serving Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 6K 12K 18K 24K 30K SE +/- 52.16, N = 3 SE +/- 518.63, N = 3 SE +/- 80.36, N = 3 SE +/- 236.47, N = 3 SE +/- 81.85, N = 3 29812.18 27344.42 27564.72 28383.22 29512.13 1. (CC) gcc options: -shared -fPIC -pthread -O3 -march=native
Bullet Physics Engine Test: 136 Ragdolls OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: 136 Ragdolls Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 2 4 6 8 10 SE +/- 0.01, N = 3 SE +/- 0.01, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 6.89 6.92 6.45 6.46 6.40 1. (CXX) g++ options: -O3 -march=native -rdynamic -lglut -lGL -lGLU
SciMark Computational Test: Fast Fourier Transform OpenBenchmarking.org Mflops, More Is Better SciMark 2.0 Computational Test: Fast Fourier Transform Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 50 100 150 200 250 SE +/- 0.36, N = 4 SE +/- 1.03, N = 4 SE +/- 0.13, N = 4 SE +/- 0.38, N = 4 SE +/- 0.16, N = 4 223.71 226.01 228.18 232.19 239.27 1. (CXX) g++ options: -O3 -march=native
Redis Test: GET OpenBenchmarking.org Requests Per Second, More Is Better Redis 3.0.1 Test: GET Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 300K 600K 900K 1200K 1500K SE +/- 2621.93, N = 3 SE +/- 15364.63, N = 3 SE +/- 2474.75, N = 3 SE +/- 9042.38, N = 3 SE +/- 5158.54, N = 3 1342892.96 1348061.04 1434042.92 1408567.41 1393442.96 -std=gnu99 -pipe -g3 -O3 -funroll-loops -march=native 1. (CC) gcc options: -ggdb -rdynamic -lm -pthread -ldl
FFTW Build: Float + SSE - Size: 1D FFT Size 4096 OpenBenchmarking.org Mflops, More Is Better FFTW 3.3.4 Build: Float + SSE - Size: 1D FFT Size 4096 Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 2K 4K 6K 8K 10K SE +/- 40.65, N = 5 SE +/- 39.29, N = 5 SE +/- 68.86, N = 5 SE +/- 44.65, N = 5 SE +/- 50.75, N = 5 10357 10331 10966 10936 10951 1. (CC) gcc options: -O3 -march=native -lm
Redis Test: SET OpenBenchmarking.org Requests Per Second, More Is Better Redis 3.0.1 Test: SET Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 200K 400K 600K 800K 1000K SE +/- 2474.98, N = 3 SE +/- 2866.48, N = 3 SE +/- 4216.02, N = 3 SE +/- 3624.82, N = 3 SE +/- 2999.32, N = 3 925939.11 933433.87 969029.10 972159.31 965269.60 -std=gnu99 -pipe -g3 -O3 -funroll-loops -march=native 1. (CC) gcc options: -ggdb -rdynamic -lm -pthread -ldl
Bullet Physics Engine Test: Raytests OpenBenchmarking.org Seconds, Fewer Is Better Bullet Physics Engine 2.81 Test: Raytests Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 2 4 6 8 10 SE +/- 0.01, N = 3 SE +/- 0.05, N = 3 SE +/- 0.00, N = 3 SE +/- 0.01, N = 3 SE +/- 0.02, N = 3 6.20 6.21 5.94 5.98 5.96 1. (CXX) g++ options: -O3 -march=native -rdynamic -lglut -lGL -lGLU
PostgreSQL pgbench Scaling: Buffer Test - Test: Normal Load - Mode: Read Write OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 9.4.3 Scaling: Buffer Test - Test: Normal Load - Mode: Read Write Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 1000 2000 3000 4000 5000 SE +/- 78.63, N = 3 SE +/- 68.11, N = 3 SE +/- 33.84, N = 3 SE +/- 67.37, N = 5 SE +/- 85.01, N = 3 4411.72 4356.38 4468.82 4411.75 4437.25 -pthreads -mthreads -pthreads -mthreads 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O3 -march=native -pthread -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm
FLAC Audio Encoding WAV To FLAC OpenBenchmarking.org Seconds, Fewer Is Better FLAC Audio Encoding 1.3.1 WAV To FLAC Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 3 6 9 12 15 SE +/- 0.01, N = 5 SE +/- 0.01, N = 5 SE +/- 0.01, N = 5 SE +/- 0.01, N = 5 SE +/- 0.01, N = 5 13.26 13.15 13.45 13.44 13.45 -fvisibility=hidden -fvisibility=hidden -fvisibility=hidden 1. (CXX) g++ options: -O3 -march=native -lm
PostgreSQL pgbench Scaling: Buffer Test - Test: Single Thread - Mode: Read Write OpenBenchmarking.org TPS, More Is Better PostgreSQL pgbench 9.4.3 Scaling: Buffer Test - Test: Single Thread - Mode: Read Write Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 120 240 360 480 600 SE +/- 14.62, N = 6 SE +/- 1.42, N = 3 SE +/- 5.89, N = 3 SE +/- 1.53, N = 3 SE +/- 0.66, N = 3 538.44 531.79 552.55 546.21 549.49 -pthreads -mthreads -pthreads -mthreads 1. (CC) gcc options: -fno-strict-aliasing -fwrapv -O3 -march=native -pthread -lpgcommon -lpgport -lpq -lpthread -lrt -lcrypt -ldl -lm
Timed MAFFT Alignment Multiple Sequence Alignment OpenBenchmarking.org Seconds, Fewer Is Better Timed MAFFT Alignment 6.864 Multiple Sequence Alignment Clang 3.8.0 Clang 3.9.0 GCC 5.4.0 GCC 6.2.0 GCC 7.0.0 20160904 2 4 6 8 10 SE +/- 0.14, N = 6 SE +/- 0.17, N = 6 SE +/- 0.03, N = 3 SE +/- 0.19, N = 6 SE +/- 0.10, N = 6 7.74 7.27 5.99 6.41 6.45 1. (CC) gcc options: -O3 -lm -lpthread
Phoronix Test Suite v10.8.5