Simon L Peyton Jones <[EMAIL PROTECTED]> writes:
> Here is another powerful motivation for doing simple common subexpression
> analysis. (The duplications arise from inlinings, mostly of prelude
> functions.)
Not all of them! (see my other message) I'd be interested to see the
results of this experiment with an up-to-date compiler (actually just
the file ghc/driver/ghc-split.prl needs to be updated).
i386-unknown-linux libHS.a from 2.10:
-rw-r--r-- 1 simonm grasp 8757766 Dec 22 12:01 libHS.a
libHS.a from proto-3.02:
-rw-r--r-- 1 simonm grasp 6750106 May 19 00:27 libHS.a
I also added some stuff to strip out local symbols from the
libraries, which might improve linking time.
Cheers,
Simon
--
Simon Marlow [EMAIL PROTECTED]
University of Glasgow http://www.dcs.gla.ac.uk/~simonm/
finger for PGP public key
