[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [e-devel] [patch] edje jittery with sse



Le Sun, 5 Mar 2006 00:25:03 +1000, dans son message intitulé Re:
[e-devel] [patch] edje jittery with sse, "Imran Syed"
<freakabcd@gmail.com> a eu le courage de nous raconter :

> Hello ilLogict,
> 
> I can't replicate what you have shown (atleast without your exact
> compile strings)
> 
> $ cat floaty.c
> #include <stdio.h>
> int main( void )
> {
>     double c0=.6;
>     double k1;
>     int t0,t1;
>     int i=1,j=6;
> 
>     t0=(1.0-c0)*i+c0*j;
>     k1=(1.0-c0)*i+c0*j;
>     t1=k1;
>     printf("t0=%d\n",t0);
>     printf("k1=%e\n",k1);
>     printf("t1=%d\n",t1);
>     c0=k1-4;
>     printf("k1-4=%e\n",c0);
> 
>     return 0;
> }
> $ gcc -march=athlon-xp -mfpmath=sse floaty.c -o floaty
> $ ./floaty
> t0=3
> k1=4.000000e+00
> t1=4
> k1-4=0.000000e+00
> $ gcc -march=athlon-xp -mfpmath=387 floaty.c -o floaty
> $ ./floaty
> t0=3
> k1=4.000000e+00
> t1=4
> k1-4=0.000000e+00
> 
> I think the athlon-xp sse doesn't have support for double and it is
> probably using 387 here. Do you happen to have a P4 or newer AMD chip
> with sse2? Also, with regards to 80bits in 387 and 64 in sse, you
> want to look at -ffloat-store. Last time I checked, it was supposed
> to help in these kind of situations.
> 
> Regards,
> Imran

 Hello!

 Yes, I have an Athlon64, so with sse2. -ffloat-store doesn't help
(tested), as it is used to make calculations with 64bits instead of 80
when using 387, no?

 Cheers!

                                                     ilLogict