[PATCH] Re: Rewrite i386 string operation expansion

Jan Hubicka jh@suse.cz
Thu Nov 30 09:19:00 GMT 2006

It would be probably good idea to fill in the pentium4 code generation
descriptor with correct values, since the code is probably just picking
up the wrong alternatives.  If you really do have pentium4 core (and nod
nocona or prescott), could you please run the attached benchmark with
./test_stringop 32 640000000 ./xgcc -B ./ -march=pentium4
With properlly filed descriptor, GCC ought to do best withtout
-minline-all-stringops (or with -minline-stringops-dynamically) on the
benchmark.  Is povray memcpy/memset bound? Where I find the sources?

> Jan,
> my proposed patch was more or less a band-aid for povray compilation
> failure (-march=pentium4 -minline-all-stringops). However, I've
> noticed that this proposed "solution" actually introduced a bit
> (~1.4%) longer execution times on my povray test, perhaps due to the
> reason you outlined above.
> There is a fixup code for TARGET_INLINE_ALL_STRINGOPS below, but it is
> never reached due to the assert that triggers in pentium4 case. I
> guess that the correct solution is more complex than the one I
> proposed.
> Uros.

More information about the Gcc-patches mailing list