On Feb 28, 2006, at 11:31 PM, Zdenek Dvorak wrote:
I have a patch that decreases number of update_stmt calls in tree alias
analysis to 46525; still, is it really that useful to run
pass_may_alias
*six* times during compilation? Obviously, we need the initial one,
and
there are comments after pass_sra and pass_fold_builtins that indicate
that the following pass_may_alias cannot be avoided (which seems
doubtful to me, at the very least in the later case).
They can be avoided but just not today. The one after PRE is the same
issue as the SRA and FAB. I filed PR 23659 for the problem with FAB,
it is similar issue with the other two.
Thanks,
Andrew Pinski
