Performance regressions in HEAD
simonmarhaskell at gmail.com
Fri Jul 6 04:39:43 EDT 2007
Just to be clear here: you're saying that
- 6.6.1 and 6.7 generate exactly the same Cmm
- but their performance differs significantly,
- with -fasm or -fvia-C?
If -fasm, did you look at -ddump-opt-cmm? (we have a cmm-to-cmm optimiser pass
that runs before the NCG, that flag shows its output. We don't bother running
this pass for -fvia-C, gcc will do these optimisations anyway.
If -fvia-C, are you using the same version of gcc in both cases?
Neil Mitchell wrote:
> Yet more information. Both were compiled using -O2, and for the
> ASM/CMM dumps with -fasm as well.
> On 7/6/07, Neil Mitchell <ndmitchell at gmail.com> wrote:
>> Further information:
>> Focusing on 4.hs (I will have another email in detail about the
>> differences between the two), GHC 6.7 gives the slower time even with
>> -fvia-C turned on. Taking a closer look at the STG and CMM, GHC 6.6.1
>> and 6.7 both give the same results. It appears the difference comes
>> after the CMM.
>> I have attached:
>> oldlog.txt - the optimised CMM and ASM with GHC 6.6.1
>> headlog.txt - the same, but with GHC 6.7
>> On 7/5/07, Neil Mitchell <ndmitchell at gmail.com> wrote:
>> > Hi
>> > Running my benchmark of character counting, with the files
>> > ExampleSingle.hs (hand written) and 4.hs (generated by Supero), I get
>> > vastly superior performance using GHC 6.6.1 compared to
>> > ghc-6.7.20070626.
>> > GHC 6.6.1
>> > GHC: Elapsed time = 8.047 seconds
>> > Supero: Elapsed time = 2.641 seconds
>> > GHC 6.7.20070626
>> > GHC: Elapsed time = 11.328 seconds
>> > Supero: Elapsed time = 6.375 seconds
>> > Any idea why?
>> > Thanks
>> > Neil
> Cvs-ghc mailing list
> Cvs-ghc at haskell.org
More information about the Cvs-ghc