Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[AMD] Introduce an OptimizeLDSUsage pass #3730

Merged
merged 22 commits into from
Jul 20, 2024
Merged

[AMD] Introduce an OptimizeLDSUsage pass #3730

merged 22 commits into from
Jul 20, 2024

Commits on Jul 18, 2024

  1. [AMD] OptimizeLDSUsage pass 

    This PR inroduces OptimizeLDSUsage pass which generalizes LDS optimization,
which was part of DecomposeUnsupportedLayouts pass.
    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    324240f View commit details
    Browse the repository at this point in the history

  2. remove redundant change

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    d946d78 View commit details
    Browse the repository at this point in the history

  3. add license

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    bf459a1 View commit details
    Browse the repository at this point in the history

  4. add pass argument

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    4e32f8d View commit details
    Browse the repository at this point in the history

  5. bring back old DecomposeUnsupportedConversions

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    124132d View commit details
    Browse the repository at this point in the history

  6. unify decompose unsupported convert and optimize lds

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    46b2c32 View commit details
    Browse the repository at this point in the history

  7. add precise scratch buffer size target

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    487d507 View commit details
    Browse the repository at this point in the history

  8. add additional test for small layout_convert

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    48e634e View commit details
    Browse the repository at this point in the history

  9. change comment for LDS reduction algorithm

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    5d3c61c View commit details
    Browse the repository at this point in the history

  10. move buffer analysis to Allocation class

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    80e25cb View commit details
    Browse the repository at this point in the history

  11. support arbitrary rank of tensors

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    7828db0 View commit details
    Browse the repository at this point in the history

  12. Address review comments: 

    - use arch name to infer lds size
- remove unused code, simplify code, rename entities, etc.
    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    2a4f018 View commit details
    Browse the repository at this point in the history

  13. move inveriants out of tmp layout creation loop

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    4a1704b View commit details
    Browse the repository at this point in the history

  14. move includes in proper order

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    4772207 View commit details
    Browse the repository at this point in the history

  15. post rebase fix

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    2e0c156 View commit details
    Browse the repository at this point in the history

  16. Fix macOS build issues

    @antiagainst @binarman
    antiagainst authored and binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    cc51df9 View commit details
    Browse the repository at this point in the history

  17. revert constant LDS size

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    55a7075 View commit details
    Browse the repository at this point in the history

  18. revert DecomposeUnsupportedConversions.cpp

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    be3be24 View commit details
    Browse the repository at this point in the history

  19. after rebase fixes

    @binarman
    binarman committed Jul 18, 2024
    Configuration menu
    Copy the full SHA
    28620cd View commit details
    Browse the repository at this point in the history

Commits on Jul 19, 2024

  1. Add more details on pass lds limit argument

    @binarman
    binarman committed Jul 19, 2024
    Configuration menu
    Copy the full SHA
    b08d270 View commit details
    Browse the repository at this point in the history

Commits on Jul 20, 2024

  1. Improve various pieces

    @antiagainst
    antiagainst committed Jul 20, 2024
    Configuration menu
    Copy the full SHA
    01ab78f View commit details
    Browse the repository at this point in the history

  2. Merge branch 'main' into reduce_lds_usage

    @antiagainst
    antiagainst authored Jul 20, 2024
    Configuration menu
    Copy the full SHA
    23419b0 View commit details
    Browse the repository at this point in the history