Mentions légales du service

Skip to content
Snippets Groups Projects

Feature/matvec 4all

Merged KUHN Matthieu requested to merge mkuhn/spm:feature/matvec_4all into master
All threads resolved!

Matrix vector multiply for all sparse matrix formats with the help of function pointers to reduce code duplication.

Edited by Mathieu Faverge

Merge request reports

Checking pipeline status.

Approval is optional

Merged by Mathieu FavergeMathieu Faverge 7 years ago (May 24, 2018 1:29pm UTC)

Merge details

  • Changes merged into master with 55f09455.
  • Deleted the source branch.

Activity

Filter activity
  • Approvals
  • Assignees & reviewers
  • Comments (from bots)
  • Comments (from users)
  • Commits & branches
  • Edits
  • Labels
  • Lock status
  • Mentions
  • Merge request status
  • Tracking
  • Mathieu Faverge
  • KUHN Matthieu added 2 commits

    added 2 commits

    • 0a112fbd - MR discussions
    • 78786381 - Removed unnecessary functions. Changed interfaces for matvec to spm_?_spmv and matmat to spm_?_spmm

    Compare with previous version

  • KUHN Matthieu resolved all discussions

    resolved all discussions

  • KUHN Matthieu resolved all discussions

    resolved all discussions

  • Still WIP ? :)

  • Mathieu Faverge added 40 commits

    added 40 commits

    • 78786381...e58ae509 - 35 commits from branch solverstack:master
    • 96d9c0fb - Matrix vector product for all Spm storage formats using function pointers
    • 165ca9f4 - Added tests for other matrix storage formats
    • 645b1d50 - Slightly modified for consistency with dof matvec tests
    • c10a83e1 - MR discussions
    • 5f4375dc - Removed unnecessary functions. Changed interfaces for matvec to spm_?_spmv and matmat to spm_?_spmm

    Compare with previous version

  • Mathieu Faverge added 5 commits

    added 5 commits

    • 637a8cd1 - Update cmake
    • 213ceb47 - Switch to a double for matVec/MatMat generic function to avoid troubles with…
    • d381c10a - Reformat the matrix-vector to add some inline and reduce the number of function clals in the stack
    • c490a07d - Add the new clean interface for sparse matrix-bvector and matrix-matrix products
    • b888349a - Call the new interface for products

    Compare with previous version

  • Mathieu Faverge unmarked as a Work In Progress

    unmarked as a Work In Progress

  • @all Ready for review.

    @mkuhn I changed a "bit" your code to avoid multiple function calls, and especially to change a bit the loops when I could. I didn't put the full detailed interface (!1 (closed)) as it became a total mess when I tried to do it. I'm not sure it's worth doing it. Is this interface enough ? or should we provide something more detailed ? I would say, it's fine like this.

  • RAMET Pierre
  • Mathieu Faverge resolved all discussions

    resolved all discussions

  • added 1 commit

    • 22f2ea95 - Cleanup the complex/real alpha/beta variables

    Compare with previous version

  • Mathieu Faverge mentioned in commit 55f09455

    mentioned in commit 55f09455

  • Please register or sign in to reply
    Loading