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

Adding f16 as Dtype #696

Merged
merged 52 commits into from
Apr 27, 2023
Merged

Adding f16 as Dtype #696

merged 52 commits into from
Apr 27, 2023

Commits on Apr 11, 2023

  1. Adding f16 as Dtype

    @coreylowman
    coreylowman committed Apr 11, 2023
    Configuration menu
    Copy the full SHA
    43edb29 View commit details
    Browse the repository at this point in the history

  2. Adding matmul kernels

    @coreylowman
    coreylowman committed Apr 11, 2023
    Configuration menu
    Copy the full SHA
    c8494fc View commit details
    Browse the repository at this point in the history

  3. Adding AssertClose impl for f16/bf16

    @coreylowman
    coreylowman committed Apr 11, 2023
    Configuration menu
    Copy the full SHA
    d0af7e9 View commit details
    Browse the repository at this point in the history

Commits on Apr 12, 2023

  1. Adding f16 to mnist

    @coreylowman
    coreylowman committed Apr 12, 2023
    Configuration menu
    Copy the full SHA
    185f92f View commit details
    Browse the repository at this point in the history

  2. Fixing cudarc dependency, adding f16 feature

    @coreylowman
    coreylowman committed Apr 12, 2023
    Configuration menu
    Copy the full SHA
    c582d1e View commit details
    Browse the repository at this point in the history

  3. Adding half/use-intrinsics on nightly

    @coreylowman
    coreylowman committed Apr 12, 2023
    Configuration menu
    Copy the full SHA
    d622e6d View commit details
    Browse the repository at this point in the history

  4. Merge branch 'main' into f16

    @coreylowman
    coreylowman committed Apr 12, 2023
    Configuration menu
    Copy the full SHA
    e000a6e View commit details
    Browse the repository at this point in the history

  5. Adding test_dtype!() macro

    @coreylowman
    coreylowman committed Apr 12, 2023
    Configuration menu
    Copy the full SHA
    f237018 View commit details
    Browse the repository at this point in the history

  6. Adds compile error if both f16 and f64 are tested

    @coreylowman
    coreylowman committed Apr 12, 2023
    Configuration menu
    Copy the full SHA
    880364b View commit details
    Browse the repository at this point in the history

Commits on Apr 19, 2023

  1. Draft: Add f16 CUDA kernels (#705) 

    * Begin working on f16 cuda kernels

* Prevent compilation for incompatible test features
    @ViliamVadocz
    ViliamVadocz authored Apr 19, 2023
    Configuration menu
    Copy the full SHA
    f126af9 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'main' into f16

    @coreylowman
    coreylowman committed Apr 19, 2023
    Configuration menu
    Copy the full SHA
    e9e36b1 View commit details
    Browse the repository at this point in the history

  3. Fixing merge conflict

    @coreylowman
    coreylowman committed Apr 19, 2023
    Configuration menu
    Copy the full SHA
    7883460 View commit details
    Browse the repository at this point in the history

Commits on Apr 20, 2023

  1. Adding to_dtype for tests

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    6a44060 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'main' into f16

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    a9e54bb View commit details
    Browse the repository at this point in the history

  3. Merge branch 'main' into f16

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    3121bda View commit details
    Browse the repository at this point in the history

  4. Adding scalar arithmetic for f16/f32. Fixing more tests

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    fafad01 View commit details
    Browse the repository at this point in the history

  5. Tests compiling

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    588be14 View commit details
    Browse the repository at this point in the history

  6. f32 tests passing

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    d6eccef View commit details
    Browse the repository at this point in the history

  7. f64 tests passing for cpu

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    aee414a View commit details
    Browse the repository at this point in the history

  8. Using NumCast instead of TestDtype

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    36d4c88 View commit details
    Browse the repository at this point in the history

  9. Adding bf16 feature flag

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    ded9941 View commit details
    Browse the repository at this point in the history

  10. Fixing f16/bf16 stuff for Device impl for cuda

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    9868f57 View commit details
    Browse the repository at this point in the history

  11. Adding cuda kernel sketches to all ops

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    c174ba1 View commit details
    Browse the repository at this point in the history

  12. Formatting

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    b571d66 View commit details
    Browse the repository at this point in the history

  13. Updating f16 kernels

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    6277aa3 View commit details
    Browse the repository at this point in the history

  14. Fixing cuda compilation errors

    @coreylowman
    coreylowman committed Apr 20, 2023
    Configuration menu
    Copy the full SHA
    2791cf2 View commit details
    Browse the repository at this point in the history

Commits on Apr 24, 2023

  1. Untested atomicMin and atomicMax implementations (#732)

    @ViliamVadocz
    ViliamVadocz authored Apr 24, 2023
    Configuration menu
    Copy the full SHA
    58ce2fa View commit details
    Browse the repository at this point in the history

  2. Use atomicCAS for atomicMin/Max (#733)

    @ViliamVadocz
    ViliamVadocz authored Apr 24, 2023
    Configuration menu
    Copy the full SHA
    9aa4a5c View commit details
    Browse the repository at this point in the history

  3. Fixing accuracy issues & build.rs errors

    @coreylowman
    coreylowman committed Apr 24, 2023
    Configuration menu
    Copy the full SHA
    eee6ae3 View commit details
    Browse the repository at this point in the history

  4. Improving error messages in build.rs

    @coreylowman
    coreylowman committed Apr 24, 2023
    Configuration menu
    Copy the full SHA
    c513841 View commit details
    Browse the repository at this point in the history

  5. Fixing rmsprop for f16

    @coreylowman
    coreylowman committed Apr 24, 2023
    Configuration menu
    Copy the full SHA
    6a308d9 View commit details
    Browse the repository at this point in the history

  6. Improving some f16 code

    @coreylowman
    coreylowman committed Apr 24, 2023
    Configuration menu
    Copy the full SHA
    5c9be5c View commit details
    Browse the repository at this point in the history

Commits on Apr 25, 2023

  1. Work on compatibility (#735) 

    * Fix struct initialization for optimizers

* Add some compatibility functions, mainly atomicCAS

* Temporarily add an example Windows include path

* Swap order of shorts in atomicCAS
    @ViliamVadocz
    ViliamVadocz authored Apr 25, 2023
    Configuration menu
    Copy the full SHA
    3fe083d View commit details
    Browse the repository at this point in the history

  2. Adding env!(CUDA_INCLUDE_DIR)

    @coreylowman
    coreylowman committed Apr 25, 2023
    Configuration menu
    Copy the full SHA
    3e48a84 View commit details
    Browse the repository at this point in the history

  3. Adding pool2d f16

    @coreylowman
    coreylowman committed Apr 25, 2023
    Configuration menu
    Copy the full SHA
    8684914 View commit details
    Browse the repository at this point in the history

  4. Fixing compatibility functions for half & double

    @coreylowman
    coreylowman committed Apr 25, 2023
    Configuration menu
    Copy the full SHA
    1705e48 View commit details
    Browse the repository at this point in the history

Commits on Apr 26, 2023

  1. Moving from impl Into<E> to impl Into<f64>

    @coreylowman
    coreylowman committed Apr 26, 2023
    Configuration menu
    Copy the full SHA
    f739880 View commit details
    Browse the repository at this point in the history

  2. Merge branch 'main' into f16

    @coreylowman
    coreylowman committed Apr 26, 2023
    Configuration menu
    Copy the full SHA
    c9752fb View commit details
    Browse the repository at this point in the history

  3. Adding f32 rhs for scalar cmp operators

    @coreylowman
    coreylowman committed Apr 26, 2023
    Configuration menu
    Copy the full SHA
    c052024 View commit details
    Browse the repository at this point in the history

  4. Fixing inclukdiung compat headers in JIT kernels

    @coreylowman
    coreylowman committed Apr 26, 2023
    Configuration menu
    Copy the full SHA
    fb6a15d View commit details
    Browse the repository at this point in the history

Commits on Apr 27, 2023

  1. Remove broken atomicCAS and redo atomicAdd,Max,Min (#742)

    @ViliamVadocz
    ViliamVadocz authored Apr 27, 2023
    Configuration menu
    Copy the full SHA
    a5b129a View commit details
    Browse the repository at this point in the history

  2. [Breaking] Optimizer config values are now all f64 (#744)

    @coreylowman
    coreylowman authored Apr 27, 2023
    Configuration menu
    Copy the full SHA
    7fa7932 View commit details
    Browse the repository at this point in the history

  3. Fixing optimizer cuda kernels

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    e23dfa9 View commit details
    Browse the repository at this point in the history

  4. Removing bf16

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    84f95ac View commit details
    Browse the repository at this point in the history

  5. Revert changes to 06-mnist

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    433b49f View commit details
    Browse the repository at this point in the history

  6. Updating test data for div

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    7dad9d3 View commit details
    Browse the repository at this point in the history

  7. Merge branch 'main' into f16

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    7d2bd9f View commit details
    Browse the repository at this point in the history

  8. Fixing CUDA_INCLUDE_DIR when ci-check is on

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    1bdc025 View commit details
    Browse the repository at this point in the history

  9. Fixing serde test failures

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    61c1261 View commit details
    Browse the repository at this point in the history

  10. Fixing ci checks

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    b3e8e35 View commit details
    Browse the repository at this point in the history

  11. Removing NumCast from optimizer tests

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    cb18b92 View commit details
    Browse the repository at this point in the history

  12. Addressing PR comments

    @coreylowman
    coreylowman committed Apr 27, 2023
    Configuration menu
    Copy the full SHA
    aa23ef1 View commit details
    Browse the repository at this point in the history