darcs patch: Don't charge for int2Word# and word2Int# --
simonpj at microsoft.com
Tue Oct 31 12:56:06 EST 2006
| Hmm. I think this needs work. If I'm not greatly mistaken, it slows a
| benchmarks down. At least, I don't think the "RULES" patch did it...
I appreciate the work you've been doing on inlining etc.
I have not applied any of the inlining or rules related ones (as you'll
have (not) seen), partly because I've been terribly snowed under, but
also because, as you have been discovering, when you nail down the
carpet in one place it tends to pop up in another, so it can take a
while to get a patch that improves performance consistently.
How to progress? Rather than try to follow the detailed path of your
investigation, I wonder if you might do the following. When you achieve
a stable situation where you think you have a collection of
modifications that improve at least some programs, without making any
significantly worse (you can negotiate about exceptions) send a patch or
patches (to GHC and the libraries) that implements your proposal, along
with a summary of what they do (unless that's all clear from the patch
messages themselves). Preferably without patches that do X and later
That way Simon and Ian and I can review and test one thing. Does that
More information about the Cvs-ghc