[Trilinos-Users] Transposing a Tpetra::MultiVector

Hoemmen, Mark mhoemme at sandia.gov
Wed Dec 13 00:09:16 EST 2017


Could you please post feature requests as issues on our GitHub site?


Thanks!  The feature request here is for an update operation 

Y := alpha * X^T + beta * Y

where X and Y are Tpetra::MultiVector instances, and alpha and beta are scalars.  Could you please explain your use case?  This is not an operation that commonly arises in iterative linear solvers or eigensolvers.

Thanks and best regards,
Mark Hoemmen

On 12/6/17, 9:26 AM, "Trilinos-Users on behalf of trilinos-users-request at trilinos.org" <trilinos-users-bounces at trilinos.org on behalf of trilinos-users-request at trilinos.org> wrote:
>Message: 9
>Date: Tue, 21 Nov 2017 15:52:31 +0000
>From: <Nigel.Overmars at dlr.de>
>To: <trilinos-users at trilinos.org>
>Subject: [Trilinos-Users] Transposing a Tpetra::MultiVector
>	<DD07DFB16C113A47B9FAE7783DA77F9584386C at dlrexmbx02.intra.dlr.de>
>Content-Type: text/plain; charset="us-ascii"
>Dear all,
>For two suitable MultiVector's mvec1 and mvec2 and two scalars alpha and beta I want to implement
>mvec2 = alpha * mvec1 ^T + beta * mvec2
>Is there a possible way to do this directly in Tpetra? I can (ab)use the multiply member function, but there must be a more canonical way. Even more so because this would mean that I have to make an allocation. The dimensions of the objects I am considering are rather low, but it is still something I wish to prevent.

>Best regards,
>Nigel Overmars

More information about the Trilinos-Users mailing list