Date: Wed, 6 Jan 1999 15:41:04 -0600 (CST) From: Steve Huss-Lederman To: snir@XXXXXXXXXX Subject: MPI question Cc: lusk@XXXXXXXXXXX,gropp@XXXXXXXXXXX X-UIDL: e4a5f8a539d8e7b809ceb0f223a7e42c I got the attached message to the comment address. p. 120 talks about MPI_REDUCE in the presence of MPI_ACCUMULATE. Here it should be limited to where the basic underlying primative type is the same. However, for other collective ops from MPI-1 the restriction is only predefined for predefined ops. Is this correct? Steve P.S. - I am leaving for a week of R&R today so I will be offline. ------- Start of forwarded message ------- Reply-To: From: "Fernando Elson Mourao" To: "MPI-Forum" Subject: MPI_REPLACE operation Date: Wed, 6 Jan 1999 16:31:59 -0000 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0 Importance: Normal X-MimeOLE: Produced By Microsoft MimeOLE V4.72.2106.4 Sender: owner-mpi-comments@XXXXXXXXXXX Precedence: bulk Hello, My question about this operation is just to be sure about one thing. Is the MPI_REPLACE operation apliable to all MPI datatypes or only to predefined ones? Regards, Elson Mourao ======================================================================= Fernando Elson Mourao elson@XXXXXXXXXXXXX - ----------------------------------------------------------------------- Departamento de Engenharia Informatica | Tels: +351 (0)39 790.000 Universidade de Coimbra | (direct)+351 (0)39 790.082 Vila Franca- Polo II | Fax: +351 (0)39 701.266 3030-397 COIMBRA | PORTUGAL | ======================================================================= ------- End of forwarded message -------