Looks like calculation of % diff is wrong. repa-mmult takes about 2x as much time as before. Still, not too bad though.
Ben. On 16/10/2010, at 4:36 PM, [email protected] wrote: > Repa Performance Test Succeeded > > Environment > Platform > host: limitingfactor.cse.unsw.EDU.AU > arch: i386 > processor: i386 > system: Darwin 10.4.0 > Versions > GHC The Glorious Glasgow Haskell Compilation System, version 6.12.1 > GCC i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5659) > > aspect min ref% avg ref% max ref% > spread% > --------- ---------- ---------- ---------- > ------- > repa-mmult elapsed 4.202(+54) 4.203(+54) 4.203(+54) 0 > k.elapsed 4.073(+56) 4.082(+56) 4.094(+56) 0 > > repa-laplace elapsed 3.502(+17) 3.502(+14) 3.502(+14) 0 > k.elapsed 2.553(+22) 2.585(+22) 2.601(+22) 1 > > repa-fft2d elapsed 4.303(+13) 4.423(+16) 4.504(+17) 4 > k.elapsed 3.994(+17) 4.057(+18) 4.107(+19) 2 > > repa-fft3d elapsed 11.108(+27) 11.148(+28) 11.208(+27) 0 > k.elapsed 9.274(+33) 9.327(+33) 9.370(+33) 1 > > > _______________________________________________ > Cvs-ghc mailing list > [email protected] > http://www.haskell.org/mailman/listinfo/cvs-ghc _______________________________________________ Cvs-ghc mailing list [email protected] http://www.haskell.org/mailman/listinfo/cvs-ghc
