LLVM 3.3 Vectorization Benchmarks

Early LLVM 3.3/3.2 SVN vectorization benchmarks for a future article on Phoronix.com.

Compare your own system(s) to this result file with the Phoronix Test Suite by running the command: phoronix-test-suite benchmark 1302189-FO-LLVM33VEC37
Jump To Table - Results

View

Do Not Show Noisy Results
Do Not Show Results With Incomplete Data
Do Not Show Results With Little Change/Spread
List Notable Results
Show Result Confidence Charts
Allow Limiting Results To Certain Suite(s)

Statistics

Show Overall Harmonic Mean(s)
Show Overall Geometric Mean
Show Wins / Losses Counts (Pie Chart)
Normalize Results
Remove Outliers Before Calculating Averages

Graph Settings

Force Line Graphs Where Applicable
Convert To Scalar Where Applicable
Prefer Vertical Bar Graphs

Multi-Way Comparison

Condense Multi-Option Tests Into Single Result Graphs

Table

Show Detailed System Result Table

Run Management

Highlight
Result
Toggle/Hide
Result
Result
Identifier
View Logs
Performance Per
Dollar
Date
Run
  Test
  Duration
-fno-vectorize
February 18 2013
 
-fvectorize
February 18 2013
 
-fvectorize -fslp-vectorize
February 18 2013
 
Invert Behavior (Only Show Selected Data)
 

Only show results where is faster than
Only show results matching title/arguments (delimit multiple options with a comma):
Do not show results matching title/arguments (delimit multiple options with a comma):

-fno-vectorize

Compiler Notes: Optimized build; Built Feb 18 2013 (10:25:08); Default target: x86_64-unknown-linux-gnu; Host CPU: bdver1
Processor Notes: Scaling Governor: ondemand

-fvectorize

-fvectorize -fslp-vectorize

Processor: AMD FX-8350 Eight-Core @ 4.00GHz (8 Cores), Motherboard: ASUS Crosshair V Formula, Chipset: AMD ATI RD890 bridge, Memory: 8192MB, Disk: 60GB OCZ VERTEX2, Graphics: AMD Radeon HD 5700 1024MB, Audio: Realtek ALC889, Monitor: DELL P2210H, Network: Intel 82583V Gigabit Connection

OS: Ubuntu 13.04, Kernel: 3.8.0-6-generic (x86_64), Desktop: Unity 6.6.0, Display Server: X Server 1.13.2, Display Driver: radeon 7.1.0, OpenGL: 2.1 Mesa 9.2-devel (git-9c1107b) Gallium 0.4, Compiler: Clang 3.3 (SVN 175449) + LLVM 3.3svn, File-System: ext4, Screen Resolution: 1920x1080

LLVM 3.3 Vectorization BenchmarksOpenBenchmarking.orgPhoronix Test SuiteAMD FX-8350 Eight-Core @ 4.00GHz (8 Cores)ASUS Crosshair V FormulaAMD ATI RD890 bridge8192MB60GB OCZ VERTEX2AMD Radeon HD 5700 1024MBRealtek ALC889DELL P2210HIntel 82583V Gigabit ConnectionUbuntu 13.043.8.0-6-generic (x86_64)Unity 6.6.0X Server 1.13.2radeon 7.1.02.1 Mesa 9.2-devel (git-9c1107b) Gallium 0.4Clang 3.3 (SVN 175449) + LLVM 3.3svnext41920x1080ProcessorMotherboardChipsetMemoryDiskGraphicsAudioMonitorNetworkOSKernelDesktopDisplay ServerDisplay DriverOpenGLCompilerFile-SystemScreen ResolutionLLVM 3.3 Vectorization Benchmarks PerformanceSystem Logs- Optimized build; Built Feb 18 2013 (10:25:08); Default target: x86_64-unknown-linux-gnu; Host CPU: bdver1 - Scaling Governor: ondemand

-fno-vectorize-fvectorize-fvectorize -fslp-vectorizeResult OverviewPhoronix Test Suite100%107%115%122%VP8 libvpx EncodingTimed PHP CompilationHimeno BenchmarkGraphicsMagickTimed HMMer SearchTimed MAFFT AlignmentGraphicsMagickGraphicsMagickGraphicsMagickSmallptC-RayGraphicsMagickvpxencTime To CompileP.P.SL.A.TP.D.SM.S.AHWB Color SpaceBlurResizingG.I.R.1.STotal TimeSharpen

LLVM 3.3 Vectorization Benchmarkshmmer: Pfam Database Searchmafft: Multiple Sequence Alignmentvpxenc: vpxencgraphics-magick: Blurgraphics-magick: Sharpengraphics-magick: Resizinggraphics-magick: HWB Color Spacegraphics-magick: Local Adaptive Thresholdinghimeno: Poisson Pressure Solverbuild-php: Time To Compilec-ray: Total Timesmallpt: Global Illumination Renderer; 100 Samples-fno-vectorize-fvectorize-fvectorize -fslp-vectorize12.906.9122.0564227210537662.6120.7732.7839712.596.8522.2664227210439640.7621.5632.7939712.537.0017.2163227110338677.4222.6932.89402OpenBenchmarking.org

Timed HMMer Search

This test searches through the Pfam database of profile hidden markov models. The search finds the domain structure of Drosophila Sevenless protein. Learn more via the OpenBenchmarking.org test page.

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed HMMer Search 2.3.2Pfam Database Search-fno-vectorize-fvectorize-fvectorize -fslp-vectorize3691215SE +/- 0.17, N = 3SE +/- 0.13, N = 3SE +/- 0.19, N = 312.9012.5912.53-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pthread -lhmmer -lsquid -lm

Timed MAFFT Alignment

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed MAFFT Alignment 6.864Multiple Sequence Alignment-fno-vectorize-fvectorize-fvectorize -fslp-vectorize246810SE +/- 0.07, N = 3SE +/- 0.03, N = 3SE +/- 0.10, N = 66.916.857.001. (CC) clang options: -O3 -lm -lpthread

VP8 libvpx Encoding

OpenBenchmarking.orgFrames Per Second, More Is BetterVP8 libvpx Encoding 1.1.0vpxenc-fno-vectorize-fvectorize-fvectorize -fslp-vectorize510152025SE +/- 0.03, N = 3SE +/- 0.07, N = 3SE +/- 1.83, N = 622.0522.2617.211. (CC) clang options: -m64 -lvpx -lm -lpthread

GraphicsMagick

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.16Operation: Blur-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1428425670SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3646463-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pthread -lXext -lSM -lICE -lX11 -lbz2 -lz -lm -lpthread

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.16Operation: Sharpen-fno-vectorize-fvectorize-fvectorize -fslp-vectorize510152025SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3222222-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pthread -lXext -lSM -lICE -lX11 -lbz2 -lz -lm -lpthread

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.16Operation: Resizing-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1632486480SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3727271-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pthread -lXext -lSM -lICE -lX11 -lbz2 -lz -lm -lpthread

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.16Operation: HWB Color Space-fno-vectorize-fvectorize-fvectorize -fslp-vectorize20406080100SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3105104103-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pthread -lXext -lSM -lICE -lX11 -lbz2 -lz -lm -lpthread

OpenBenchmarking.orgIterations Per Minute, More Is BetterGraphicsMagick 1.3.16Operation: Local Adaptive Thresholding-fno-vectorize-fvectorize-fvectorize -fslp-vectorize918273645SE +/- 0.00, N = 3SE +/- 0.00, N = 3SE +/- 0.00, N = 3373938-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pthread -lXext -lSM -lICE -lX11 -lbz2 -lz -lm -lpthread

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.

OpenBenchmarking.orgMFLOPS, More Is BetterHimeno Benchmark 3.0Poisson Pressure Solver-fno-vectorize-fvectorize-fvectorize -fslp-vectorize150300450600750SE +/- 1.95, N = 3SE +/- 1.02, N = 3SE +/- 1.47, N = 3662.61640.76677.42-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -O3 -march=native

Timed PHP Compilation

OpenBenchmarking.orgSeconds, Fewer Is BetterTimed PHP Compilation 5.2.9Time To Compile-fno-vectorize-fvectorize-fvectorize -fslp-vectorize510152025SE +/- 0.03, N = 3SE +/- 0.05, N = 3SE +/- 0.15, N = 320.7721.5622.69-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -march=native -O3 -pedantic -ldl -lz -lm

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.

OpenBenchmarking.orgSeconds, Fewer Is BetterC-Ray 1.1Total Time-fno-vectorize-fvectorize-fvectorize -fslp-vectorize816243240SE +/- 0.01, N = 3SE +/- 0.02, N = 3SE +/- 0.02, N = 332.7832.7932.89-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CC) clang options: -lm -lpthread -O3 -march=native

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.

OpenBenchmarking.orgSeconds, Fewer Is BetterSmallpt 1.0Global Illumination Renderer; 100 Samples-fno-vectorize-fvectorize-fvectorize -fslp-vectorize90180270360450SE +/- 1.15, N = 3SE +/- 0.58, N = 3SE +/- 0.33, N = 3397397402-fno-vectorize-fvectorize-fvectorize -fslp-vectorize1. (CXX) clang++ options: -fopenmp -march=native -O3

-fno-vectorize

Compiler Notes: Optimized build; Built Feb 18 2013 (10:25:08); Default target: x86_64-unknown-linux-gnu; Host CPU: bdver1
Processor Notes: Scaling Governor: ondemand

Testing initiated at 18 February 2013 15:06 by user phoronix.

-fvectorize

Compiler Notes: Optimized build; Built Feb 18 2013 (10:25:08); Default target: x86_64-unknown-linux-gnu; Host CPU: bdver1
Processor Notes: Scaling Governor: ondemand

Testing initiated at 18 February 2013 14:12 by user phoronix.

-fvectorize -fslp-vectorize

Processor: AMD FX-8350 Eight-Core @ 4.00GHz (8 Cores), Motherboard: ASUS Crosshair V Formula, Chipset: AMD ATI RD890 bridge, Memory: 8192MB, Disk: 60GB OCZ VERTEX2, Graphics: AMD Radeon HD 5700 1024MB, Audio: Realtek ALC889, Monitor: DELL P2210H, Network: Intel 82583V Gigabit Connection

OS: Ubuntu 13.04, Kernel: 3.8.0-6-generic (x86_64), Desktop: Unity 6.6.0, Display Server: X Server 1.13.2, Display Driver: radeon 7.1.0, OpenGL: 2.1 Mesa 9.2-devel (git-9c1107b) Gallium 0.4, Compiler: Clang 3.3 (SVN 175449) + LLVM 3.3svn, File-System: ext4, Screen Resolution: 1920x1080

Compiler Notes: Optimized build; Built Feb 18 2013 (10:25:08); Default target: x86_64-unknown-linux-gnu; Host CPU: bdver1
Processor Notes: Scaling Governor: ondemand

Testing initiated at 18 February 2013 16:01 by user phoronix.