AMD Bulldozer Compiler Optimizations Testing GCC 4.6, GCC 4.7 snapshot, and Open64 4.2.4 compilers on an AMD FX-8150 Bulldozer CPU. Testing with stock compiler support and then when the CFLAGS/CXXFLAGS with having mtune/march options for Bulldozer (bdver1) set. Benchmarking by Michael Larabel for a future article on Phoronix.com.
HTML result view exported from: https://openbenchmarking.org/result/1110219-AR-BDVER1OPT06&grt .
Processor Motherboard Chipset Memory Disk Graphics Audio Network OS Kernel Desktop Display Server Display Driver OpenGL Compiler File-System Screen Resolution GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 Stock bdver1 Stock bdver1 Stock bdver1 AMD FX-8150 Eight-Core @ 3.61GHz (8 Cores) ASUS Crosshair V Formula ATI RD890 PCI to PCI bridge 8192MB 60GB OCZ VERTEX2 NVIDIA GeForce GTX 460 768MB (675/1800MHz) Realtek ALC889 Intel 82583V Gigabit Connection Ubuntu 11.10 3.1.0-rc10-ic-aliasing-patch+ (x86_64) Unity 4.22.0 X Server 1.10.4 NVIDIA 285.05.09 4.2.0 NVIDIA 285.05.09 GCC 4.6.1 ext4 1920x1080 GCC 4.7.0 Open64 4.2.4 OpenBenchmarking.org System Details - Compiz was running on this system.
gcrypt: CAMELLIA256-ECB Cipher graphics-magick: HWB Color Space graphics-magick: Blur graphics-magick: Local Adaptive Thresholding graphics-magick: Resizing graphics-magick: Sharpen povray: Total Time GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 Stock bdver1 Stock bdver1 Stock bdver1 2263 162 98 68 131 65 822 12723 83 59 31 66 39 636 2503 163 98 68 127 65 851 10977 83 60 30 67 40 636 1960 86 59 20 69 35 823 1993 88 60 19 70 35 705 OpenBenchmarking.org
Gcrypt Library CAMELLIA256-ECB Cipher GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Microseconds, Fewer Is Better Gcrypt Library 1.4.4 CAMELLIA256-ECB Cipher Stock bdver1 3K 6K 9K 12K 15K SE +/- 18.56, N = 3 SE +/- 893.08, N = 6 SE +/- 3.33, N = 3 SE +/- 139.80, N = 3 SE +/- 10.00, N = 3 SE +/- 14.53, N = 3 2263 12723 2503 10977 1960 1993
GraphicsMagick Operation: HWB Color Space GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.12 Operation: HWB Color Space Stock bdver1 40 80 120 160 200 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 162 83 163 83 86 88
GraphicsMagick Operation: Blur GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.12 Operation: Blur Stock bdver1 20 40 60 80 100 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 +/- 0.33, N = 3 98 59 98 60 59 60
GraphicsMagick Operation: Local Adaptive Thresholding GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.12 Operation: Local Adaptive Thresholding Stock bdver1 15 30 45 60 75 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 68 31 68 30 20 19
GraphicsMagick Operation: Resizing GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.12 Operation: Resizing Stock bdver1 30 60 90 120 150 SE +/- 0.67, N = 3 SE +/- 0.67, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 131 66 127 67 69 70
GraphicsMagick Operation: Sharpen GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Iterations Per Minute, More Is Better GraphicsMagick 1.3.12 Operation: Sharpen Stock bdver1 15 30 45 60 75 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.00, N = 3 SE +/- 0.33, N = 3 SE +/- 0.33, N = 3 SE +/- 0.00, N = 3 65 39 65 40 35 35
POV-Ray Total Time GCC 4.6.1 GCC 4.7-20111015 Open64 4.2.4 OpenBenchmarking.org Seconds, Fewer Is Better POV-Ray 3.6.1 Total Time Stock bdver1 200 400 600 800 1000 822 636 851 636 823 705
Phoronix Test Suite v10.8.4