Speeding up sched1
Jeffrey A Law
law@cygnus.com
Tue Nov 30 23:59:00 GMT 1999
In message < 19991107012830.A15561@cygnus.com >you write:
> Probably true. I could reorg flow such that they got 0 and n-1
> respectively, and make n_basic_blocks += 2. That would kinda
> weird out things that check n_basic_blocks == 1 though...
Yea. Eventually we're going to need to do it though. But I haven't even
started to look at how many things are going to break.
> I don't have a good feel for how to fix things. What we have
> at present is gonna break with Catherine's multiple entry points.
We're in no worse shape than before my change. Previously we'd skip
the first block with a loop like
for (i = 1; i < n_basic_blocks; i++)
{
do something for block I
}
jeff
More information about the Gcc-patches
mailing list