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

CPU AVX implementation for Softmax, Norm #357

Merged
merged 76 commits into from
Jan 9, 2024
Merged

CPU AVX implementation for Softmax, Norm #357

merged 76 commits into from
Jan 9, 2024

Commits on Aug 4, 2023

  1. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash

initial commit

works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash

change imports

fix for diff size, compiledmodule error fix
    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    14cbb3b View commit details
    Browse the repository at this point in the history

  2. works on multidimensional, axis=-1

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    7896c45 View commit details
    Browse the repository at this point in the history

  3. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash
    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    ff90ed5 View commit details
    Browse the repository at this point in the history

  4. change imports

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    fc61204 View commit details
    Browse the repository at this point in the history

  5. fix for diff size, compiledmodule error fix

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    f84201f View commit details
    Browse the repository at this point in the history

  6. works on multidimensional, axis=-1

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    6f2e43c View commit details
    Browse the repository at this point in the history

  7. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash
    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    25f22cf View commit details
    Browse the repository at this point in the history

  8. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash
    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    aafbb0f View commit details
    Browse the repository at this point in the history

  9. change imports

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    44993e2 View commit details
    Browse the repository at this point in the history

  10. fix for diff size, compiledmodule error fix

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    a86d866 View commit details
    Browse the repository at this point in the history

  11. works on multidimensional, axis=-1

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    b59ffa2 View commit details
    Browse the repository at this point in the history

  12. wrap up softmax, starting layernorm

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    7edf0eb View commit details
    Browse the repository at this point in the history

  13. layernorm kinda works but not rly

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    44c04b3 View commit details
    Browse the repository at this point in the history

  14. better code for softmax

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    2ccc4b6 View commit details
    Browse the repository at this point in the history

  15. layernorm works for last layer

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    13ea5dc View commit details
    Browse the repository at this point in the history

  16. move find sum and find max to registered function

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    d89036d View commit details
    Browse the repository at this point in the history

  17. find max in registered func

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    b0659f6 View commit details
    Browse the repository at this point in the history

  18. not working softmax on not last dim, minor changes

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    904760b View commit details
    Browse the repository at this point in the history

  19. layernorm works for any dims

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    29b7ba7 View commit details
    Browse the repository at this point in the history

  20. comments

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    0c8dc3a View commit details
    Browse the repository at this point in the history

  21. tuning, fix for flowgraph operator resolve

    @fishingguy456
    fishingguy456 committed Aug 4, 2023
    Configuration menu
    Copy the full SHA
    77fe8d9 View commit details
    Browse the repository at this point in the history

Commits on Aug 5, 2023

  1. softmax works

    @fishingguy456
    fishingguy456 committed Aug 5, 2023
    Configuration menu
    Copy the full SHA
    ac40695 View commit details
    Browse the repository at this point in the history

  2. commented tensors dont work, i.e. axis is not last 2 AND not multiple… 

    … of 8
    @fishingguy456
    fishingguy456 committed Aug 5, 2023
    Configuration menu
    Copy the full SHA
    4938a1f View commit details
    Browse the repository at this point in the history

Commits on Aug 8, 2023

  1. actually works rn frfr so fast 💯

    @fishingguy456
    fishingguy456 committed Aug 8, 2023
    Configuration menu
    Copy the full SHA
    1d447cf View commit details
    Browse the repository at this point in the history

  2. cleanup

    @fishingguy456
    fishingguy456 committed Aug 8, 2023
    Configuration menu
    Copy the full SHA
    30224ce View commit details
    Browse the repository at this point in the history

Commits on Aug 9, 2023

  1. more cleanup

    @fishingguy456
    fishingguy456 committed Aug 9, 2023
    Configuration menu
    Copy the full SHA
    67d4d56 View commit details
    Browse the repository at this point in the history

Commits on Aug 11, 2023

  1. random testing stuff

    @fishingguy456
    fishingguy456 committed Aug 11, 2023
    Configuration menu
    Copy the full SHA
    09ca2f8 View commit details
    Browse the repository at this point in the history

Commits on Aug 18, 2023

  1. allow epilogue

    @fishingguy456
    fishingguy456 committed Aug 18, 2023
    Configuration menu
    Copy the full SHA
    8352dd8 View commit details
    Browse the repository at this point in the history

  2. better epiloguing

    @fishingguy456
    fishingguy456 committed Aug 18, 2023
    Configuration menu
    Copy the full SHA
    27f6cbb View commit details
    Browse the repository at this point in the history

Commits on Aug 25, 2023

  1. janky matmul resolve

    @fishingguy456
    fishingguy456 committed Aug 25, 2023
    Configuration menu
    Copy the full SHA
    cce1d42 View commit details
    Browse the repository at this point in the history

  2. still epilogue problem?

    @fishingguy456
    fishingguy456 committed Aug 25, 2023
    Configuration menu
    Copy the full SHA
    f92de53 View commit details
    Browse the repository at this point in the history

Commits on Sep 14, 2023

  1. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash

initial commit

works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash

change imports

fix for diff size, compiledmodule error fix
    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    63dfed4 View commit details
    Browse the repository at this point in the history

  2. works on multidimensional, axis=-1

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    73a063a View commit details
    Browse the repository at this point in the history

  3. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash
    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    1c129c0 View commit details
    Browse the repository at this point in the history

  4. change imports

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    bf8a5b5 View commit details
    Browse the repository at this point in the history

  5. fix for diff size, compiledmodule error fix

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    3aa5cb6 View commit details
    Browse the repository at this point in the history

  6. works on multidimensional, axis=-1

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    b849ebf View commit details
    Browse the repository at this point in the history

  7. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash
    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    12fdbd1 View commit details
    Browse the repository at this point in the history

  8. initial commit 

    works on 8x8 at least but bad exp

save for omp changes

working and faster than pytorch

works and is fast but exp is WIP

remove useless files

minor changes for rebase

delete trash

fix trash

fix trash
    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    9c7ecd0 View commit details
    Browse the repository at this point in the history

  9. change imports

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    b155bbd View commit details
    Browse the repository at this point in the history

  10. fix for diff size, compiledmodule error fix

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    de72bc6 View commit details
    Browse the repository at this point in the history

  11. works on multidimensional, axis=-1

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    17b8d76 View commit details
    Browse the repository at this point in the history

  12. wrap up softmax, starting layernorm

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    1b52167 View commit details
    Browse the repository at this point in the history

  13. layernorm kinda works but not rly

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    e479db7 View commit details
    Browse the repository at this point in the history

  14. better code for softmax

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    c623630 View commit details
    Browse the repository at this point in the history

  15. layernorm works for last layer

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    b44b69e View commit details
    Browse the repository at this point in the history

  16. move find sum and find max to registered function

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    29ea558 View commit details
    Browse the repository at this point in the history

  17. find max in registered func

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    339e549 View commit details
    Browse the repository at this point in the history

  18. not working softmax on not last dim, minor changes

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    88c423c View commit details
    Browse the repository at this point in the history

  19. layernorm works for any dims

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    9c91875 View commit details
    Browse the repository at this point in the history

  20. comments

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    6e0d8e5 View commit details
    Browse the repository at this point in the history

  21. tuning, fix for flowgraph operator resolve

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    552aebb View commit details
    Browse the repository at this point in the history

  22. softmax works

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    dc258e3 View commit details
    Browse the repository at this point in the history

  23. commented tensors dont work, i.e. axis is not last 2 AND not multiple… 

    … of 8
    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    95f6be7 View commit details
    Browse the repository at this point in the history

  24. actually works rn frfr so fast 💯

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    d0b99a4 View commit details
    Browse the repository at this point in the history

  25. cleanup

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    67a43a5 View commit details
    Browse the repository at this point in the history

  26. more cleanup

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    4443780 View commit details
    Browse the repository at this point in the history

  27. random testing stuff

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    4088fc6 View commit details
    Browse the repository at this point in the history

  28. allow epilogue

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    7430696 View commit details
    Browse the repository at this point in the history

  29. better epiloguing

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    8a1167e View commit details
    Browse the repository at this point in the history

  30. janky matmul resolve

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    0f4876f View commit details
    Browse the repository at this point in the history

  31. still epilogue problem?

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    49c072f View commit details
    Browse the repository at this point in the history

  32. Merge remote-tracking branch 'origin/main'

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    0bd13d8 View commit details
    Browse the repository at this point in the history

  33. clean up for pr

    @fishingguy456
    fishingguy456 committed Sep 14, 2023
    Configuration menu
    Copy the full SHA
    de74231 View commit details
    Browse the repository at this point in the history

Commits on Sep 18, 2023

  1. fix test

    @fishingguy456
    fishingguy456 committed Sep 18, 2023
    Configuration menu
    Copy the full SHA
    9ab0bac View commit details
    Browse the repository at this point in the history

  2. lint

    @fishingguy456
    fishingguy456 committed Sep 18, 2023
    Configuration menu
    Copy the full SHA
    f779a1d View commit details
    Browse the repository at this point in the history

Commits on Sep 19, 2023

  1. minor pr edits

    @fishingguy456
    fishingguy456 committed Sep 19, 2023
    Configuration menu
    Copy the full SHA
    124fb09 View commit details
    Browse the repository at this point in the history

  2. pytests, cpu child class

    @fishingguy456
    fishingguy456 committed Sep 19, 2023
    Configuration menu
    Copy the full SHA
    6c4efd9 View commit details
    Browse the repository at this point in the history

  3. potential fix for failing tests? but prob not will have to investigat… 

    …e more
    @fishingguy456
    fishingguy456 committed Sep 19, 2023
    Configuration menu
    Copy the full SHA
    40fd71f View commit details
    Browse the repository at this point in the history

Commits on Jan 4, 2024

  1. weird diff

    @fishingguy456
    fishingguy456 committed Jan 4, 2024
    Configuration menu
    Copy the full SHA
    90c4ffb View commit details
    Browse the repository at this point in the history

  2. merge conflict resolve build.py

    @fishingguy456
    fishingguy456 committed Jan 4, 2024
    Configuration menu
    Copy the full SHA
    587ba64 View commit details
    Browse the repository at this point in the history

  3. remove shady batch mat mul

    @fishingguy456
    fishingguy456 committed Jan 4, 2024
    Configuration menu
    Copy the full SHA
    89d5646 View commit details
    Browse the repository at this point in the history

Commits on Jan 5, 2024

  1. lint thing

    @fishingguy456
    fishingguy456 committed Jan 5, 2024
    Configuration menu
    Copy the full SHA
    a3a4b03 View commit details
    Browse the repository at this point in the history

Commits on Jan 8, 2024

  1. move helpers to new file

    @fishingguy456
    fishingguy456 committed Jan 8, 2024
    Configuration menu
    Copy the full SHA
    aec95d2 View commit details
    Browse the repository at this point in the history

  2. lint

    @fishingguy456
    fishingguy456 committed Jan 8, 2024
    Configuration menu
    Copy the full SHA
    7a41b5c View commit details
    Browse the repository at this point in the history

  3. change tolerance for flaky test for test_dynamic_shape

    @fishingguy456
    fishingguy456 committed Jan 8, 2024
    Configuration menu
    Copy the full SHA
    dcc6a45 View commit details
    Browse the repository at this point in the history