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

Add fast path for ArgMin / ArgMax when axis is contiguous #411

Merged
merged 3 commits into from
Nov 16, 2024
Merged

Add fast path for ArgMin / ArgMax when axis is contiguous #411

merged 3 commits into from
Nov 16, 2024

Commits on Nov 16, 2024

  1. Add Lane::as_slice method 

    Add a method to get contiguous lanes of a tensor as a slice. This is useful for
implementing fast paths for the case where the lane is contiguous.
    @robertknight
    robertknight committed Nov 16, 2024
    Configuration menu
    Copy the full SHA
    b4e3d2b View commit details
    Browse the repository at this point in the history

  2. Add ArgMin / ArgMax fast path for contiguous (stride 1) axes

    @robertknight
    robertknight committed Nov 16, 2024
    Configuration menu
    Copy the full SHA
    4b97e9d View commit details
    Browse the repository at this point in the history

  3. Make NaN testing more efficient in reduction ops 

    Add an `IsNaN` trait which calls `f32::is_nan` to test for NaN-ness instead of
testing whether `self.partial_cmp(self)` is None.
    @robertknight
    robertknight committed Nov 16, 2024
    Configuration menu
    Copy the full SHA
    90a8604 View commit details
    Browse the repository at this point in the history