Changes between Version 71 and Version 72 of DataParallel/BenchmarkStatus


Ignore:
Timestamp:
Dec 3, 2010 3:20:39 AM (5 years ago)
Author:
benl
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • DataParallel/BenchmarkStatus

    v71 v72  
    5959  Applies a Gaussian blur filter to a 2D image. Size=512x512.
    6060
    61   '''ToDo:''' Runs ok, but need to add other versions for comparison.
     61  || '''name''' || '''runtime''' || '''speedup''' || '''efficiency''' || '''notes''' ||
     62  || repa.blur.par.N1 || 2.620s || 1 || - || ||
     63  || repa.blur.par.N4 || 0.717s || 3.65 || - || ||
     64  || repa.blur.par.N8 || 0.414s || 6.33 || - || ||
     65
     66  '''ToDo:''' Runs ok, but need other versions for comparison.
    6267
    6368
     
    6671  Performs Canny edge detection to a 2D image. Size=512x512.
    6772
    68   '''ToDo:''' Runs ok, but need to add other versions for comparison.
     73  || '''name''' || '''runtime''' || '''speedup''' || '''efficiency''' || '''notes''' ||
     74  || repa.edgedetect.par.N1 || 206ms || 1 || - || ||
     75  || repa.edgedetect.par.N4 || 79ms || 2.6 || - || ||
     76  || repa.edgedetect.par.N8 || 55ms || 3.75 || - || ||
     77
     78  '''ToDo:''' Runs ok, but need other versions for comparison.
    6979
    7080[[br]]
     
    7282  Performs high-pass filtering using 2D and 3D FFTs. These are naive benchmarks used for regression testing only. They divide right down to (rank generalise) two-point vectors and construct the result using copying append. Using an inplace algorithm (like with FFTW) would be significantly faster.
    7383
    74   '''ToDo:''' Runs ok, but need to add other versions for comparison.
     84  '''ToDo:''' Runs ok, but need other versions for comparison.
    7585
    7686
     
    113123
    114124  || '''name''' || '''runtime''' || '''speedup''' || '''efficiency''' || '''notes''' ||
    115   || dph.evens.vectorised.seq.N4 || 1.075s || 1 || 1 || ||
    116   || dph.evens.vectorised.par.N1 || 736ms ||  1.46 || 1.46 || ||
    117   || dph.evens.vectorised.par.N2 || 768ms ||  1.40 || 0.70 || ||
    118   || dph.evens.vectorised.par.N4 || 859ms ||  1.25 || 0.31 || ||
     125  || dph.evens.vectorised.seq.N4 || 1.075s || 1 || - || ||
     126  || dph.evens.vectorised.par.N1 || 736ms ||  1.46 || - || ||
     127  || dph.evens.vectorised.par.N2 || 768ms ||  1.40 || - || ||
     128  || dph.evens.vectorised.par.N4 || 859ms ||  1.25 || - || ||
    119129
    120130  '''Status''': Benchmark runs slower when number of threads increases. This benchmark invokes {{{packByTag}}} due to the filtering operation. This is probably affecting Quickhull as it also uses filtering. [[br]]