Skip to content

Provide a matrix transpose operation #50

Closed

Description

@ViralBShah

It would be greatly beneficial if openblas could provide a user callable fast matrix transpose operation as part of the library. I am sure this capability exists internally.

Activity

  1. xianyi commented on Aug 21, 2011

    @xianyi
    Collaborator

    Hi Viral,

    Do you mean some BLAS-like extensions, e.g. mkl_?imatcopy and
    mkl_?omatcopy in Intel MKL?

    We will add this into feature list.

    Xianyi

    ViralBShah write:Sent: 2011/8/21 21:48

    It would be greatly beneficial if openblas could provide a user callable fast matrix transpose operation as part of the library. I am sure this capability exists internally.

  2. ViralBShah commented on Aug 21, 2011

    @ViralBShah
    ContributorAuthor

    I am not sure what those functions do, but they sound like copy and transposed copy. If so, those would be nice to have.

    -viral

    On 21-Aug-2011, at 9:26 PM, xianyi wrote:

    Hi Viral,

    Do you mean some BLAS-like extensions, e.g. mkl_?imatcopy and
    mkl_?omatcopy in Intel MKL?

    We will add this into feature list.

    Xianyi

    ViralBShah write:Sent: 2011/8/21 21:48

    It would be greatly beneficial if openblas could provide a user callable fast matrix transpose operation as part of the library. I am sure this capability exists internally.

    Reply to this email directly or view it on GitHub:
    #50 (comment)

  3. xianyi commented on Aug 22, 2011

    @xianyi
    Collaborator

    I added this feature into next milestone. I think it's easy to implement. However, we need time to tune the performance as same as Intel MKL.

  4. ViralBShah commented on Aug 22, 2011

    @ViralBShah
    ContributorAuthor

    That would be fantastic! Even exposing the API immediately would be useful, and performance can be improved over time.

    Thanks,

    -viral

    On 22-Aug-2011, at 8:30 AM, xianyi wrote:

    I added this feature into next milestone. I think it's easy to implement. However, we need time to tune the performance as same as Intel MKL.

    Reply to this email directly or view it on GitHub:
    #50 (comment)

  5. certik commented on Sep 26, 2013

    @certik

    For the record, the new issue for this feature is #51.

  6. added a commit that references this issue on Apr 27, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions