Re: Performance of Integer Multiplication on PIII (Results forgcc-2.95 & Athlon)

[pete at ltoi dot iap dot physik dot tu-darmstadt dot de]

Hi,

> >
> > Hi,
> Hi,
> could you please try the patch I've attached if it solves the
> slowdown?  it should IMO.
> > gcc-2.95.2,3 (haifa):
> >  Loop: 1.01, Code: 2.57
> >  Clocks: 19.15
> >
> > gcc-2.95.2: (normal scheduler)
> >  Loop: 1.01, Code: 3.02
> >  Clocks: 22.50
> >
> > gcc-3.0.2 & gcc-3.0-haifa
> >  Loop: 0.90, Code: 3.35
> >  Clocks: 24.96
> >
> > gcc-295.2 (haifa) -O2 -march=i686 -s -o imul imul.c read_empty.c read_asm.s
> >   Loop: 1.01, Code: 3.02
> >   Clocks: 22.50
> >
> > gcc-295.2 (haifa) -O2 -march=i686 -s -o imul imul.c read_empty.c read_iasm.s
>
> Note that with -march=i686 new gcc often perofrms worse on Athlon, but it is
> mainly because it do more Athlon specific stuff.
>
> > And now we outperform (again) the famous Intel compilers with gcc!
> >
> > -ffplfj=2:= -fomit-frame-pointer -malign-loops=2 -malign-functions=2
> >             -malign-jumps=2
> >
> > gcc3.0.2 -O2 -march=athlon -ffplfj=2
> >  Loop: 2.35, Code: 0.90
> >  Clocks: 6.70
>
> How much of this speedup is accounted to fomit-frame-pointer and how
> much to the alignment changes?

Unfortuately the Athlon box is located 15 minutes (one way) from here,
with no internet conection nor LAN conection ...

I will check this at home. But i could say, that -fomit-frame-pointer
without alignment don't helps. binary runs even slower ...

As you can see in my prevoius mail, the problem seems to be more
involved (?), insofar, as this "trick" did not work on the PII ...

Only one point come out so far: there are optimization problems in the new x86
backend ...

>
>
> I am attaching the assembly of pent.c (w/o midification) I get from
> -march=athlon -O2 -fomit-frame-pointer on the mainline after my patch
> so you can try how it works for you.
>
 ... homework ... (will cause delay)

> Honza
>

Peter