X-pop3-spooler: POP3MAIL 2.1.0 b 4 980420 -bs- Message-ID: <19980910025449.48146@cerebro.laendle> Date: Thu, 10 Sep 1998 02:54:49 +0200 From: Marc Lehmann To: Vincent Diepeveen , beastium-list AT Desk DOT nl Subject: Re: -O1 -ffast-math bug on Pentium Pro Mail-Followup-To: Vincent Diepeveen , beastium-list AT Desk DOT nl References: <3 DOT 0 DOT 32 DOT 19980908180857 DOT 009d99b0 AT xs4all DOT nl> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <3.0.32.19980908180857.009d99b0@xs4all.nl>; from Vincent Diepeveen on Tue, Sep 08, 1998 at 06:09:05PM +0000 X-Operating-System: Linux version 2.1.120 (root AT cerebro) (gcc version pgcc-2.91.57 19980901 (egcs-1.1 release)) Status: RO Content-Length: 968 Lines: 22 On Tue, Sep 08, 1998 at 06:09:05PM +0000, Vincent Diepeveen wrote: > Or does this not matter for the bug? > > >Here is the asembler output of pgcc-1.1a (fmbug, with fast-math) > > Who can read such unreadable assembler outputs? hmm, how do you identify the problem? finding a bug in a diff (correct -> incorrect) seems easier to me than wading through every line of assembly code ;) For example, in the diff you can easily see that there are missing cc0-setting instructions.. (although I admit I can't read x86-fpu code very well). -----==- | ----==-- _ | ---==---(_)__ __ ____ __ Marc Lehmann +-- --==---/ / _ \/ // /\ \/ / pcg AT goof DOT com |e| -=====/_/_//_/\_,_/ /_/\_\ --+ The choice of a GNU generation | |