Skip to content

<algorithm>: vectorize min_element / max_element using SSE4.1/AVX for floats #2439

Description

Similar to #2438 .

I assume it is safe for /fp:fast. Not sure about other FP modes and how to handle FP corner cases such as denormals or singalings.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    fixedSomething works now, yay!performanceMust go faster

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions