On Nov 5, 2010, at 8:56 PM, Brandon S Allbery KF8NH wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 11/5/10 19:22 , David Peixotto wrote:
Probably there are some wins to be had by choosing a good optimization sequence for the code generated from GHC, rather than just using `-O1`, `-O2`, etc. I believe It should be possible to find a good optimization sequence that would work well for Haskell codes.
Didn't someone (dons?) already make a start on this?
Searching for good compiler sequences is certainly not a new idea. Most work that I know focuses on finding good sequences for a particular program. I think an interesting opportunity here would be to search for good sequences that generally work well for Haskell programs to replace the standard -O1, -O2 used by LLVM. I would think the code generated from GHC is different enough that we should be able to find standard sequences that we could use to replace the ones currently used by LLVM.
- -- brandon s. allbery [linux,solaris,freebsd,perl] allbery@kf8nh.com system administrator [openafs,heimdal,too many hats] allbery@ece.cmu.edu electrical and computer engineering, carnegie mellon university KF8NH -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.10 (Darwin) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAkzUtdQACgkQIn7hlCsL25VG6ACeJ3sSXoI4YLbXW3KIFVMqKqdK oTsAn23bxl0mvfdl3up69xM4qWPnklGj =TXBk -----END PGP SIGNATURE----- _______________________________________________ Glasgow-haskell-users mailing list Glasgow-haskell-users@haskell.org http://www.haskell.org/mailman/listinfo/glasgow-haskell-users