Subject: C518 review From: "David W Walker" Date: Tue, 18 Sep 2001 15:40:58 +0100 To: "\"Geoffrey Fox(CandC:PandE Referee Report)\"" X-UIDL: 5c15477b61280000 X-Mozilla-Status: 0001 X-Mozilla-Status2: 00000000 Received: by mailer.csit.fsu.edu (mbox gcfpc) (with Cubic Circle's cucipop (v1.31 1998/05/13) Sun Oct 14 21:55:48 2001) X-From_: fox@mailer.csit.fsu.edu Sun Oct 14 21:45:29 2001 Return-Path: Delivered-To: gcfpc@csit.fsu.edu Received: from dirac.csit.fsu.edu (dirac.csit.fsu.edu [144.174.128.44]) by mailer.csit.fsu.edu (Postfix) with ESMTP id A310423A05 for ; Sun, 14 Oct 2001 21:45:29 -0400 (EDT) Received: from localhost by dirac.csit.fsu.edu (AIX4.2/UCB 8.7) id VAA61204; Sun, 14 Oct 2001 21:45:28 -0400 (EDT) Resent-Message-Id: <200110150145.VAA61204@dirac.csit.fsu.edu> Delivered-To: fox@csit.fsu.edu Received: from fins.uits.indiana.edu (fins.uits.indiana.edu [129.79.6.185]) by mailer.csit.fsu.edu (Postfix) with ESMTP id E46C423A45 for ; Tue, 18 Sep 2001 10:41:44 -0400 (EDT) Received: from sentinel.cs.cf.ac.uk (sentinel.cs.cf.ac.uk [131.251.42.18]) by fins.uits.indiana.edu (8.10.1/8.10.1/IUPO) with ESMTP id f8IEfh024943 for ; Tue, 18 Sep 2001 09:41:43 -0500 (EST) Received: from zardoz.cs.cf.ac.uk ([131.251.42.89] helo=DWWCOMSC) by sentinel.cs.cf.ac.uk with smtp (Exim 3.22 #4) id 15jM4D-0006Oz-00 for gcf@indiana.edu; Tue, 18 Sep 2001 15:41:41 +0100 Message-ID: <0d8601c1404f$ee71e440$592afb83@DWWCOMSC> MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_NextPart_000_0D82_01C14058.502B9DE0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 Resent-To: Geoffrey Fox Resent-Date: Sun, 14 Oct 2001 21:45:28 -0400 Resent-From: Geoffrey Fox Geoffrey I have reviewed the paper by Valsalam and Skjellum. It is basically about getting the best possible performance for sequential matrix multiplication so I am not sure if it is suitable for the journal. In other respects it is pretty much OK, but rather long. I hope this helps David Prof David W Walker Department of Computer Science Cardiff University PO Box 916 Cardiff CF24 3XF United Kingdom +44 (0)29 2087-4205 (phone) +44 (0)29 2087-4598 (fax) http://www.cs.cf.ac.uk/User/David.W.Walker/ CandC:PandE Referee Report Form -- Electronic Transimission to "Geoffrey Fox(CandC:PandE Referee Report)" strongly preferred Referees Home Page: http://aspen.csit.fsu.edu/CandCPandE/ Email "Geoffrey Fox(CandC:PandE Referee Report)" for URL of full paper to be reviewed WILEY Journal Home Page John Wiley and Sons, Ltd. Baffins Lane, Chichester West Sussex, PO19 1UD, England Telephone: (01243) 779777 Fax: (01243) 770379 REFEREE'S REPORT Concurrency and Computation:Practice and Experience A: General Information Please return to: Geoffrey C. Fox Electronically Preferred "Geoffrey Fox(CandC:PandE Referee Report)" Concurrency and Computation: Practice and Experience Computational Science and Information Technology Florida State University 400 Dirac Science Library Tallahassee Florida 32306-4130 Office FAX 850-644-0098 Office Phone 850-644-4587 but best is cell phone 3152546387 Please fill in Summary Conclusions (Sec. C) and details as appropriate in Secs. D, E and F. B: Refereeing Philosophy We encourage a broad range of readers and contributors. Please judge papers on their technical merit and separate comments on this from those on style and approach. Keep in mind the strong practical orientation that we are trying to give the journal. Note that the forms attached provide separate paper for comments that you wish only the editor to see and those that both the editor and author receive. Your identity will of course not be revealed to the author. C: Paper and Referee Metadata * Paper Number: C518 * Date: 18 September 2001 * Paper Title: A framework for high-performance matrix multiplication based on hierarchical abstractions, algorithms and optimized low-level kernels * Author(s): Vinod Valsalam and Anthony Skjellum * Referee: David Walker * Address: Cardiff University Referee Recommendations. Please indicate overall recommendations here, and details in following sections. 1. publish as is 2. accepted provided changes suggested are made X 3. reject X if it is decided not appropriate for the journal D: Referee Comments (For Editor Only) I am not sure if CandC:PandE is the most suitable journal for this paper as it is mostly about ways to improve the performance of matrix multiplication on sequential computers through the use of storage formats and algorithmic techniques that result in efficient use of cache. Extending these ideas to hierarchical memory in general it would be expected that the same approach could lead to a better parallel algorithm but this is not discussed in the paper. The paper is also rather long. E: Referee Comments (For Author and Editor) This is a careful piece of work that presents new results on sequential matrix multiplication algorithms. Have the authors tried to come up with an analytical model of the performance of the various algorithms discussed? It would be useful to develop such a model that reflects the costs of data movement and computation, and to compare it with the results given in the paper. Similarly it should be possible to model the cost of converting between different storage schemes (e.g., between hierarchical and row-major or column-major). This would confirm the relatively small cost of conversion. It would also be interesting to extend the approach to parallel algorithms for matrix multiplication. F: Presentation Changes The authors might like to consider splitting the paper into 2 separate papers. One giving an overview of matrix multiplication algorithms, including the new approaches presented in the original paper, and a second presenting the results, analysis and interpretation.