loop optimizations, mark 12[a]
Jeffrey A Law
law@cygnus.com
Mon Jul 6 17:41:00 GMT 1998
In message < 19980706091616.A12625@dot.cygnus.com >you write:
> On Thu, Jul 02, 1998 at 09:50:50AM -0600, Jeffrey A Law wrote:
> > Could be a double alignment issue -- I did run them multiple times and th
> e
> > triggering module was main000.f, which seemed a little weird since twldrv
> > and fppp are the known hog routines.
>
> Nope. You probably didn't get to my message to egcs that showed the
> half-speed regression from 980506 to 980702 on Alpha as well. I'm
> just now getting around to checking out old versions to see what's
> changed in the output.
Nope. I believe I was at 30000ft when I wrote that message and thus
it drained when I returned to SLC last night and plugged back into
my home net -- before I read email.
Odd as I thought I'd done the test correctly -- ie test without the
patch, apply the patch and retest.
jeff
More information about the Gcc-patches
mailing list