Article View: gnu.gcc.bug
Article #6024bsfl
From: Joe Keane
Date: Sun, 05 Jan 1997 00:00
Date: Sun, 05 Jan 1997 00:00
4 lines
675 bytes
675 bytes
On the Intel x86, the "bsfl" instruction should never be used, because it's just way too slow. The generic code is much faster, even adding some extra function calls. I think this is true across processors.
Technical Details
Message-ID:
<199701060431.UAA30733@rocky.jgk.org>Path:
arc1-ca.pugleaf.net!unknown!not-for-mailShow Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 113d02,7ec0b449b79ebe72,start X-Google-Attributes: gid113d02,public From: jgk@jgk.org (Joe Keane) Subject: bsfl Date: 1997/01/05 Message-ID: <199701060431.UAA30733@rocky.jgk.org> X-Deja-AN: 208008751 sender: bug-gcc-request@prep.ai.mit.edu x-gateway: relay1.UU.NET from bug-gcc to gnu.gcc.bug; Mon, 6 Jan 1997 01:51:03 EST x-moon-phase: Waning Crescent (19% of Full) newsgroups: gnu.gcc.bug