Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-07 Thread Simon Marlow
On 06/03/2012 18:16, Johan Tibell wrote: On Tue, Mar 6, 2012 at 8:54 AM, Simon Marlow marlo...@gmail.com mailto:marlo...@gmail.com wrote: what loop unrolling? In LLVM you mean? I added loop unrolling for the thawArray# etc primops both to the NCG and the LLVM backend. Ah yes, I'd

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-07 Thread David Terei
On 6 March 2012 08:54, Simon Marlow marlo...@gmail.com wrote: On 06/03/2012 16:02, Johan Tibell wrote: On Tue, Mar 6, 2012 at 7:23 AM, Simon Marlow marlo...@gmail.com mailto:marlo...@gmail.com wrote:    However it's possible that we'll replace the mini-inliner entirely.      It has always

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-07 Thread Simon Marlow
On 07/03/2012 09:08, David Terei wrote: On 6 March 2012 08:54, Simon Marlowmarlo...@gmail.com wrote: On 06/03/2012 16:02, Johan Tibell wrote: On Tue, Mar 6, 2012 at 7:23 AM, Simon Marlowmarlo...@gmail.com mailto:marlo...@gmail.com wrote: However it's possible that we'll replace the

[commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-06 Thread Simon Marlow
Repository : ssh://darcs.haskell.org//srv/darcs/ghc On branch : newcg http://hackage.haskell.org/trac/ghc/changeset/9a32e71d912985a6fd8e3491518ac357f2e8686b --- commit 9a32e71d912985a6fd8e3491518ac357f2e8686b Author: Simon Marlow

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-06 Thread Johan Tibell
Is this something that should be ported to master? ___ Cvs-ghc mailing list Cvs-ghc@haskell.org http://www.haskell.org/mailman/listinfo/cvs-ghc

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-06 Thread Simon Marlow
On 06/03/2012 14:00, Johan Tibell wrote: Is this something that should be ported to master? I made a couple of improvements to the mini-inliner on the newcg branch which turned out to be wrong, and this commit fixed it. If the improvements get merged over, then the fixes should too.

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-06 Thread Johan Tibell
On Tue, Mar 6, 2012 at 7:23 AM, Simon Marlow marlo...@gmail.com wrote: However it's possible that we'll replace the mini-inliner entirely. It has always been a bit of a hack, and the new code generator is quite effective at exposing its shortcomings. Edward's CmmRewriteAssignments pass was

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-06 Thread Simon Marlow
On 06/03/2012 16:02, Johan Tibell wrote: On Tue, Mar 6, 2012 at 7:23 AM, Simon Marlow marlo...@gmail.com mailto:marlo...@gmail.com wrote: However it's possible that we'll replace the mini-inliner entirely. It has always been a bit of a hack, and the new code generator is quite

Re: [commit: ghc] newcg: fixes to the mini-inliner (fixes stage2 crashes) (9a32e71)

2012-03-06 Thread Johan Tibell
On Tue, Mar 6, 2012 at 8:54 AM, Simon Marlow marlo...@gmail.com wrote: what loop unrolling? In LLVM you mean? I added loop unrolling for the thawArray# etc primops both to the NCG and the LLVM backend. Eventually hoopl should buy us the ability to do some great optimisations at the C--