kr.angelov: > 2005/10/19, Simon Marlow <[EMAIL PROTECTED]>: > > On 19 October 2005 01:08, Donald Bruce Stewart wrote: > > > Ah!! So what's going on on Linux, I wonder. Could it be something > > > about > > > 6.4.1? Are we seeing the difference between ForeignPtrs from 6.4 to > > > 6.5? I will investigate. > > > > I bet that's the difference between ForeignPtrs in 6.4.1 and 6.5. I'm > > very tempted to merge the ForeignPtr optimisations into 6.4.2, so we can > > benefit from this earlier. The changes are all confined to one module > > (GHC/ForeignPtr.hs). > > I made my benchmarking with ghc-6.5 and I am getting relatively the > same result as with ghc-6.4.1. I checked that the optimized ForeignPtr > implementation is in my source tree. Any thoughts?
Ok, rerunning the test you posted with only the compiler version varying (so, just the FPS ForeignPtr code on OpenBSD x86, Pentium M 1.6G). With ghc 6.4.1 test1: 121.10s test2: 252.11s And with ghc 6.5 test1: 2.006s test2: 6.112s Wow, ForeignPtr was slow with 6.4.1! -- Don _______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe