Re: optimization/6010: sse code appears to have issues with some instructions

[Jan Hubicka <jh@suse.cz>]

The following reply was made to PR optimization/6010; it has been noted by GNATS.

From: Jan Hubicka <jh@suse.cz>
To: Jan Hubicka <jh@suse.cz>
Cc: jakub@gcc.gnu.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org,
	haganc@cray.com, nobody@gcc.gnu.org, gcc-gnats@gcc.gnu.org
Subject: Re: optimization/6010: sse code appears to have issues with some instructions
Date: Thu, 18 Apr 2002 16:15:06 +0200

 > > Synopsis: sse code appears to have issues with some instructions
 > 
 > Cau,
 > tak jsem zjistil, ze na to mam patch a dokonce uz zapprovovany (mam v tom
 > vetsi binec, nez bych cekal).  Az dobootstrapuje, tak ho commitnu, takze
 > asi muzes bugreport uzavrit.
 
 Oops, I apologize for Czech Language, I tought I am sending email only to
 Jakub.  I have patch approved to fix the testcase, so I will install it once
 re-bootstrapping passes.
 
 Honza
 > 
 > Honza
 > > 
 > > State-Changed-From-To: open->analyzed
 > > State-Changed-By: jakub
 > > State-Changed-When: Thu Apr 18 05:56:37 2002
 > > State-Changed-Why:
 > >     Simplified into:
 > >     /* PR optimization/6010 */
 > >     /* { dg-do compile } */
 > >     /* { dg-options "-O2 -funroll-all-loops" } */
 > >     /* { dg-options "-O2 -funroll-all-loops -march=pentium3" { target i?86-*-* } } */
 > >     
 > >     void bar (float);
 > >     
 > >     void foo (float y, unsigned long z)
 > >     {
 > >       int b;
 > >       float c = y;
 > >     
 > >       for (b = 0; b < z; b++)
 > >         {
 > >           bar (c);
 > >           if (c == y)
 > >             c = -y;
 > >           else
 > >             c = y;
 > >         }
 > >     }
 > > 
 > > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=6010