[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: subscribe

	From owner-atlas-comm@cs.utk.edu Tue Nov 14 05:37:11 2000
	Return-Path: <owner-atlas-comm@cs.utk.edu>
	Received: from cs.utk.edu (CS.UTK.EDU [])
		by cs.utexas.edu (8.9.3/8.9.3) with ESMTP id FAA18093
		for <rvdg@cs.utexas.edu>; Tue, 14 Nov 2000 05:37:09 -0600 (CST)
	Received: from localhost (daemon@localhost) 
	        by cs.utk.edu with SMTP (cf v2.9s-UTK)
		id GAA06632; Tue, 14 Nov 2000 06:31:36 -0500 (EST)
	Received: by cs.utk.edu (bulk_mailer v1.13); Tue, 14 Nov 2000 06:31:36 -0500
	Received: from serenity.mcc.ac.uk (marvin@localhost) 
	        by cs.utk.edu with ESMTP (cf v2.9s-UTK)
		id GAA06619; Tue, 14 Nov 2000 06:31:33 -0500 (EST)
	Received: from serenity.mcc.ac.uk ( -> serenity.mcc.ac.uk)
	 by cs.utk.edu (smtpshim v1.0); Tue, 14 Nov 2000 06:31:34 -0500
	Received: from wallace.mvc.mcc.ac.uk ([] helo=man.ac.uk)
		by serenity.mcc.ac.uk with esmtp (Exim 2.05 #4)
		id 13veJE-000Mi8-00
		for atlas-comm@cs.utk.edu; Tue, 14 Nov 2000 11:31:28 +0000
	Message-ID: <3A112286.2E134FB4@man.ac.uk>
	Date: Tue, 14 Nov 2000 11:31:18 +0000
	From: Dan Kidger <d.kidger@man.ac.uk>
	Organization: Manchester Computing
	X-Mailer: Mozilla 4.6 [en] (Win98; I)
	X-Accept-Language: en
	MIME-Version: 1.0
	To: atlas-comm@cs.utk.edu
	Subject: Re: subscribe
	References: <200011131918.OAA27438@enterprise.cs.utk.edu> <3A104FAF.D1266CEC@cs.utexas.edu>
	Content-Type: text/plain; charset=us-ascii
	Content-Transfer-Encoding: 7bit
	Robert van de Geijn wrote:
	> What is GPL?
	> gemm numbers are on the webpage now:
	> http://www.cs.utexas.edu/users/flame/ITXGEMM/performance.html
	> The best performance I have seen with our stuff on my laptop has been better
	> than 80% of peak.
	Can I just check what you mean by peak?
	I thought a P III could issue 2 flops every 3 cycles, such that a 650Mhz chip
	maxed out at 433 Mflops/s ?
	Dr. Daniel Kidger                               | E: d.kidger@man.ac.uk
	High Performance Computing Group                | W: www.csar.cfs.ac.uk
	Manchester Computing, University of Manchester, | T:   +44 161 275 7038
	Manchester, M13 9PL, United Kingdom.            | F:   +44 161 275 6800
	--------------------Q: what's up ?   A: Z cross X ---------------------

BTW, notice that peak performance is not particularly
important.  What is important is how the matrix-matrix
multiply benefits an application, which is why we
have LU with pivoting numbers on the webpage, and
how well it does for all matrix sizes, which is why
we show data for m,n,k=32:32:512 and m,n,k=40:40:520
on our webpage.