Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

Performance improvement in Normalize GPU Kernel #14139

Merged
merged 9 commits into from
Feb 14, 2019
Merged

Performance improvement in Normalize GPU Kernel #14139

merged 9 commits into from
Feb 14, 2019

Commits on Feb 14, 2019

  1. New CPU kernel for normalize

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    04fc2fd View commit details
    Browse the repository at this point in the history

  2. New GPU kernel for Normalize

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    64d2b47 View commit details
    Browse the repository at this point in the history

  3. Add launch bounds and increase threads to 32*32

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    9c175f9 View commit details
    Browse the repository at this point in the history

  4. do not hardcode number of threads

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    94787e8 View commit details
    Browse the repository at this point in the history

  5. Try fix windows build failure

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    7b8039c View commit details
    Browse the repository at this point in the history

  6. make channels as int to fix windows build issues with omp

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    a98f034 View commit details
    Browse the repository at this point in the history

  7. Simplify cuda kernels with 1 D thread block

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    eb8da43 View commit details
    Browse the repository at this point in the history

  8. Minor refactoring

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    652e1ca View commit details
    Browse the repository at this point in the history

  9. Revert thread dim for ToTensor operator

    @sandeep-krishnamurthy
    sandeep-krishnamurthy committed Feb 14, 2019
    Configuration menu
    Copy the full SHA
    f595dff View commit details
    Browse the repository at this point in the history