Changes between Version 73 and Version 74 of DataParallel/BenchmarkStatus


Ignore:
Timestamp:
Dec 3, 2010 4:25:26 AM (5 years ago)
Author:
benl
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • DataParallel/BenchmarkStatus

    v73 v74  
    181181These programs also use user defined algebraic data types. Vectorization of these programs is still a work in progress. 
    182182 
    183  [http://darcs.haskell.org/libraries/dph/dph-examples/real/NBody/ BarnesHut] '''(BROKEN)''':: 
     183 [http://darcs.haskell.org/libraries/dph/dph-examples/real/NBody/ BarnesHut] '''(SLOWLORIS)''':: 
    184184  This benchmark implements the Barnes-Hut algorithm to solve the ''n''-body problem in two dimensions. There is a naive O(n^2^) version in the same package. 
    185185 
    186   '''Status''': Core-lint error due to bug in GHC's rule matcher. If we compile without -dcore-lint it segfaults when run. 
     186  || '''name''' || '''runtime''' || '''speedup''' || '''efficiency''' || '''notes''' || 
     187  || dph.nbody.vector.seq.N4 || 100ms || 1 || || A || 
     188  || dph.nbody.vectorised.seq.N4 || 4681ms || ~50x slower   ||  || || 
     189  || dph.nbody.vectorised.par.N1 || 2381ms ||  ~25x slower || || || 
     190  A : Time stated is end-to-end, not just for the kernel. 
     191 
     192  '''Status''':  Compiles, but fusion hasn't worked so is very slow. [[br]] 
     193  '''!ToDo''': Make the vectorised way give the same output as the vector way. The benchmark setup is a bit different. Fixing this won't cause a 50x speed difference though. 
     194 
    187195 
    188196[[br]]