This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: SPEC 2000 results on PowerPC

From: Dale Johannesen <dalej at apple dot com>
To: Toon Moene <toon at moene dot indiv dot nluug dot nl>
Cc: gcc at gcc dot gnu dot org, Mark Mitchell <mark at codesourcery dot com>, Dale Johannesen <dalej at apple dot com>
Date: Tue, 17 Aug 2004 16:47:08 -0700
Subject: Re: SPEC 2000 results on PowerPC
References: <41223D86.5000708@codesourcery.com> <518E35ED-F079-11D8-8E24-000A95D7CD40@apple.com> <412269DA.2040205@moene.indiv.nluug.nl>

On Aug 17, 2004, at 1:26 PM, Toon Moene wrote:

Dale Johannesen wrote:

The hot code in wupwise is zgemm.f, loops ending 70 and 100, and zaxpy.f, loop 30. Multiply-add insns are very important.
Which you won't get unless using -ffast-math.

It could in all these cases under the F77 "mathematically equivalent" rule. I suppose we don't because it can't tell there were no parentheses?

References:
- SPEC 2000 results on PowerPC
  - From: Mark Mitchell
- Re: SPEC 2000 results on PowerPC
  - From: Dale Johannesen
- Re: SPEC 2000 results on PowerPC
  - From: Toon Moene

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]