Re: libm problem

2003-03-23 Thread Michael Nottebrock
On Sunday 23 March 2003 20:22, Till Riedel wrote: > I think no one should simply use the sse2 stuff at the moment, > because you will notice problems too late. My > mp3s sounded horrible and gv didn't display pdf files correctly. > if there hadn't been libvorbis though that had a division by zero,

Re: libm problem

2003-03-23 Thread Julian St.
Am So, 2003-03-23 um 18.56 schrieb Bruce Cran: > On Sun, Mar 23, 2003 at 06:43:16PM +0100, Michael Nottebrock wrote: > > > > Just out of curiousity, have you tried using -mfpmath=sse? I remember someone > > on this list claiming that the SSE fpa-code works much better than the i387 > > code whic

Re: libm problem

2003-03-23 Thread Till Riedel
On Sun, Mar 23, 2003 at 07:39:06PM +0100, Michael Nottebrock wrote: Content-Description: signed data > Actually, when I said "better", I really meant "doesn't break -O(ptimized) > code" rather than "faster". :) ok, got you now. it is neither faster nor better. I tested it with e_pow.c. funnily tho

Re: libm problem

2003-03-23 Thread Michael Nottebrock
On Sunday 23 March 2003 19:24, Till Riedel wrote: > On Sun, Mar 23, 2003 at 06:43:16PM +0100, Michael Nottebrock wrote: > > Just out of curiousity, have you tried using -mfpmath=sse? I remember > > someone on this list claiming that the SSE fpa-code works much better > > than the i387 code which is

Re: libm problem

2003-03-23 Thread Till Riedel
On Sun, Mar 23, 2003 at 06:43:16PM +0100, Michael Nottebrock wrote: Content-Description: signed data > On Sunday 23 March 2003 18:02, Till Riedel wrote: > > why not > > +_CPUCFLAGS = -march=pentium4 -mno-sse2 > > > > > choose, and in the case of pentium4 producing broken code the > > > obvious fall

Re: libm problem

2003-03-23 Thread Bruce Cran
On Sun, Mar 23, 2003 at 06:43:16PM +0100, Michael Nottebrock wrote: > > Just out of curiousity, have you tried using -mfpmath=sse? I remember someone > on this list claiming that the SSE fpa-code works much better than the i387 > code which is used by default (even with -march=pentium4). > I do

Re: libm problem

2003-03-23 Thread Michael Nottebrock
On Sunday 23 March 2003 18:02, Till Riedel wrote: > why not > +_CPUCFLAGS = -march=pentium4 -mno-sse2 > > > choose, and in the case of pentium4 producing broken code the > > obvious fallback would be pentium3... > > above would be in fact the same because only the SSE2 code differs from > march=pen

Re: libm problem

2003-03-23 Thread Till Riedel
On Sun, Mar 23, 2003 at 07:08:08AM -0600, Anti wrote: > On Sat, 22 Mar 2003 23:41:14 -0800 > David Schultz <[EMAIL PROTECTED]> wrote: > > > Thus spake Anti <[EMAIL PROTECTED]>: > > > On Sat, 22 Mar 2003 10:28:46 -0800 > > > Steve Kargl <[EMAIL PROTECTED]> wrote: > > > > > > > Pentium 4 is definit

Re: libm problem

2003-03-23 Thread Anti
On Sat, 22 Mar 2003 23:41:14 -0800 David Schultz <[EMAIL PROTECTED]> wrote: > Thus spake Anti <[EMAIL PROTECTED]>: > > On Sat, 22 Mar 2003 10:28:46 -0800 > > Steve Kargl <[EMAIL PROTECTED]> wrote: > > > > > Pentium 4 is definitely broken on 5.x. Perhaps, we should remove > > > the footshooting.

Re: libm problem

2003-03-22 Thread David Schultz
Thus spake Anti <[EMAIL PROTECTED]>: > On Sat, 22 Mar 2003 10:28:46 -0800 > Steve Kargl <[EMAIL PROTECTED]> wrote: > > > Pentium 4 is definitely broken on 5.x. Perhaps, we should remove > > the footshooting. > > > > --- bsd.cpu.mk.orig Sat Mar 22 10:23:42 2003 > > +++ bsd.cpu.mk Sat Mar 22

Re: libm problem

2003-03-22 Thread Dan Nelson
In the last episode (Mar 22), Steve Kargl said: > On Sat, Mar 22, 2003 at 12:39:51PM -0600, Anti wrote: > > pentium3 would be better than pentiumpro on a p4 i think... > > I know pentium4 produces some strange code and I know my > system is quite stable with pentiumpro. I've never tested > what h

Re: libm problem

2003-03-22 Thread Steve Kargl
On Sat, Mar 22, 2003 at 12:39:51PM -0600, Anti wrote: > On Sat, 22 Mar 2003 10:28:46 -0800 > Steve Kargl <[EMAIL PROTECTED]> wrote: > > > Pentium 4 is definitely broken on 5.x. Perhaps, we should remove > > the footshooting. > > > > --- bsd.cpu.mk.orig Sat Mar 22 10:23:42 2003 > > +++ bsd.cp

Re: libm problem

2003-03-22 Thread Anti
On Sat, 22 Mar 2003 10:28:46 -0800 Steve Kargl <[EMAIL PROTECTED]> wrote: > Pentium 4 is definitely broken on 5.x. Perhaps, we should remove > the footshooting. > > --- bsd.cpu.mk.orig Sat Mar 22 10:23:42 2003 > +++ bsd.cpu.mk Sat Mar 22 10:27:11 2003 > @@ -62,7 +62,9 @@ > . elif ${CPUTYP

Re: libm problem

2003-03-22 Thread Steve Kargl
On Fri, Mar 21, 2003 at 09:14:43PM -0800, David Schultz wrote: > Thus spake Till Riedel <[EMAIL PROTECTED]>: > > I now know the thing that makes it break. > > > > cc -O2 -pipe -mcpu=pentiumpro -c /usr/src/lib/msun/src/e_pow.c > > works fine! > > cc -O0 -pipe -march=pentium4 -c /usr/src/lib/msun/sr

Re: libm problem

2003-03-21 Thread David Schultz
Thus spake Till Riedel <[EMAIL PROTECTED]>: > I now know the thing that makes it break. > > cc -O2 -pipe -mcpu=pentiumpro -c /usr/src/lib/msun/src/e_pow.c > works fine! > cc -O0 -pipe -march=pentium4 -c /usr/src/lib/msun/src/e_pow.c > ... works > but... > cc -O -pipe -march=pentium4 -c /usr/src/li

Re: libm problem

2003-03-21 Thread Till Riedel
On Sat, Mar 22, 2003 at 01:54:35AM +0100, Till Riedel wrote: > cc -O -pipe -march=pentium4 -c /usr/src/lib/msun/src/e_pow.c OK, I found PR 43299. Why do I never find them in first place. till To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the mess

Re: libm problem

2003-03-21 Thread Till Riedel
I now know the thing that makes it break. cc -O2 -pipe -mcpu=pentiumpro -c /usr/src/lib/msun/src/e_pow.c works fine! cc -O0 -pipe -march=pentium4 -c /usr/src/lib/msun/src/e_pow.c ... works but... cc -O -pipe -march=pentium4 -c /usr/src/lib/msun/src/e_pow.c breaks it. Hey its only gcc :-), nothing

Re: libm problem

2003-03-21 Thread Bruce Evans
On Sat, 22 Mar 2003, Till Riedel wrote: > > > res=pow((float)base,(float)dim); > this is actually not a smart thing. it was cut and paste from libvorbis. > pow is a function for doubles. if you i use powf everything works fine. This should work OK. The casts to float should have no effect exce

Re: libm problem

2003-03-21 Thread Till Riedel
> > res=pow((float)base,(float)dim); this is actually not a smart thing. it was cut and paste from libvorbis. pow is a function for doubles. if you i use powf everything works fine. res=pow((double)base,(double)dim) however still gives 1 > the output of 'print/x {int}&res' right after the call to

Re: libm problem

2003-03-19 Thread David Schultz
Thus spake Till Riedel <[EMAIL PROTECTED]>: > [EMAIL PROTECTED] $ cat test.c > #include "math.h" > > int main() > { > int base=8; > int dim=2; > float res; > res=pow((float)base,(float)dim); > printf("%f\n",res); > return 0; > } > [EMAIL PROTECTED] $ gcc -lm test.c > [EMAIL PROTECTED]

Re: libm problem

2003-03-18 Thread Till Riedel
On Tue, Mar 18, 2003 at 08:37:50PM +0300, Sergey A. Osokin wrote: > On Tue, Mar 18, 2003 at 06:30:51PM +0100, Till Riedel wrote: > > Hi, > > since installed a new world this weekend strange things happen to me: > > > > [EMAIL PROTECTED] $ cat test.c > > #include "math.h" > > > > int main() > > {

Re: libm problem

2003-03-18 Thread Sergey A. Osokin
On Tue, Mar 18, 2003 at 06:30:51PM +0100, Till Riedel wrote: > Hi, > since installed a new world this weekend strange things happen to me: > > [EMAIL PROTECTED] $ cat test.c > #include "math.h" > > int main() > { > int base=8; > int dim=2; > float res; > res=pow((float)base,(float)dim); >