Re: Compiler slowdowns (was Re: [PATCH] Optimize manual byte swap implementations v2)

[Bradley Lucier <lucier at math dot purdue dot edu>]

On Feb 6, 2009, at 7:07 PM, Bernd Schmidt wrote:

> Bradley Lucier wrote:
>
> > Well, I'm seeing an -O1 compile (with some extra passes thrown in,  
> > like
> > -fschedule-insn2) go from 168 seconds to 228.17 seconds in four  
> > months.
> > Without a nail to pound down.
>
> So, did this happen gradually or can it be pinpointed to a smaller set
> of revisions by bisecting it?

Well, I noted a 13% compile time regression between September 19 and  
September 26 here:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26854#c76

but I didn't get much encouragement.  (Last September I decided I'd  
try to measure compiler times regularly to catch problems soon after  
they happened, but I stopped when I got no response to the 13%  
problem.  I'm not saying I was right to stop, just saying what  
happened then.)

I do have a faster machine now so I can try to find out more  
precisely where the problems are if people are interested.  I did  
spend two hours doing that for another problem I was having here:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33928#c39

but I didn't get much encouragement from that, either.

Brad