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This repository has been archived by the owner on Nov 17, 2023.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

CudnnFind() usage improvements #12804

Merged
merged 14 commits into from
Oct 26, 2018
Merged

CudnnFind() usage improvements #12804

merged 14 commits into from
Oct 26, 2018

Commits on Oct 10, 2018

  1. Add mx.context.gpu_memory_info() to python api for flexible tests.

    @DickJC123
    DickJC123 committed Oct 10, 2018
    Configuration menu
    Copy the full SHA
    7ad40a2 View commit details
    Browse the repository at this point in the history

Commits on Oct 11, 2018

  1. Add test_gluon_gpu.py:test_large_models to show cudnnFind headroom is… 

    …sue.
    @DickJC123
    DickJC123 committed Oct 11, 2018
    2 Configuration menu
    Copy the full SHA
    94614a5 View commit details
    Browse the repository at this point in the history

Commits on Oct 12, 2018

  1. Output model sizes tried by test_gluon_gpu.py:test_large_models.

    @DickJC123
    DickJC123 committed Oct 12, 2018
    Configuration menu
    Copy the full SHA
    8369e1e View commit details
    Browse the repository at this point in the history

  2. Fix perl interface to MXGetGPUMemoryInformation.

    @DickJC123
    DickJC123 committed Oct 12, 2018
    Configuration menu
    Copy the full SHA
    aff8df4 View commit details
    Browse the repository at this point in the history

Commits on Oct 13, 2018

  1. Increase difficulty of test_gluon_gpu.py:test_large_models.

    @DickJC123
    DickJC123 committed Oct 13, 2018
    Configuration menu
    Copy the full SHA
    e95eb25 View commit details
    Browse the repository at this point in the history

  2. Forgot a file in fix for perl.

    @DickJC123
    DickJC123 committed Oct 13, 2018
    Configuration menu
    Copy the full SHA
    7434d4b View commit details
    Browse the repository at this point in the history

Commits on Oct 14, 2018

  1. Modify test to pass on no-cudnn CI runner.

    @DickJC123
    DickJC123 committed Oct 14, 2018
    Configuration menu
    Copy the full SHA
    7bf8d51 View commit details
    Browse the repository at this point in the history

  2. Mutex algo reg updates, serialize cudnnFind calls.

    @DickJC123
    DickJC123 committed Oct 14, 2018
    Configuration menu
    Copy the full SHA
    7f724c9 View commit details
    Browse the repository at this point in the history

  3. Fix for cudnnFind memory headroom issue.

    @DickJC123
    DickJC123 committed Oct 14, 2018
    Configuration menu
    Copy the full SHA
    7596ad6 View commit details
    Browse the repository at this point in the history

  4. Fix cpplint.

    @DickJC123
    DickJC123 committed Oct 14, 2018
    Configuration menu
    Copy the full SHA
    aa60af6 View commit details
    Browse the repository at this point in the history

Commits on Oct 25, 2018

  1. Respond to reviewers comments.

    @DickJC123
    DickJC123 committed Oct 25, 2018
    Configuration menu
    Copy the full SHA
    08b9f55 View commit details
    Browse the repository at this point in the history

  2. Guard against improper MXNET_GPU_MEM_LARGE_ALLOC_ROUND_SIZE values.

    @DickJC123
    DickJC123 committed Oct 25, 2018
    Configuration menu
    Copy the full SHA
    83df6d5 View commit details
    Browse the repository at this point in the history

  3. Merge branch 'master' into cudnnfind_usage_improvements

    @DickJC123
    DickJC123 authored Oct 25, 2018
    Configuration menu
    Copy the full SHA
    e043791 View commit details
    Browse the repository at this point in the history

  4. Fix potentially unassigned var.

    @DickJC123
    DickJC123 committed Oct 25, 2018
    Configuration menu
    Copy the full SHA
    810af44 View commit details
    Browse the repository at this point in the history