Skip to content

[Tutorial] Use OpenCL on ARM Board #525

Closed
Closed
[Tutorial] Use OpenCL on ARM Board#525
Labels
status: help wanted

Description

@tqchen
tqchen
opened on Oct 9, 2017

While http://docs.tvmlang.org/tutorials/deployment/cross_compilation_and_rpc.html#sphx-glr-tutorials-deployment-cross-compilation-and-rpc-py covers how to cross compile and deploy CPU code on ARM board, it did not touch how to do so for the OpenCL code.

We have unittest examples to demonstrate that https://github.com/dmlc/tvm/blob/master/tests/python/unittest/test_runtime_rpc.py#L92 already. It would be nice to translate these lessons into the cross compilation tutorial

Activity

  1. tqchen
    added
    status: help wanted
    on Oct 9, 2017
  2. tqchen
    changed the title [-][Tutorial] Tutorial on using OpenCL on ARM Board[/-] [+][Tutorial] Use OpenCL on ARM Board[/+] on Oct 9, 2017

  3. ma-xiaodong commented on Oct 26, 2017

    @ma-xiaodong
    ma-xiaodong
    on Oct 26, 2017

    Here is an example of using OpenCL on ARM Board.
    https://github.com/ma-xiaodong/tvm_tuning/blob/master/02_conv_order.py
    which corresponding the program
    http://docs.tvmlang.org/tutorials/optimize/opt_conv_cuda.html#sphx-glr-tutorials-optimize-opt-conv-cuda-py
    of optimizing conv on cuda.

  4. kaishijeng commented on Oct 27, 2017

    @kaishijeng
    kaishijeng
    on Oct 27, 2017

    ma-xiaodong:
    Are you to tune Mali-T860 opencl to pass test_topi_conv2d_nchw.py unittest in topi/tests/python directory?

    I always encounter an error because of insufficient resources on a firefly3399 platform running Ubuntu 16.04
    #530

    Thanks,

  5. ma-xiaodong commented on Oct 28, 2017

    @ma-xiaodong
    ma-xiaodong
    on Oct 28, 2017

    I also encountered the error CL_OUT_OF_RESOURCES.
    In order to use opencl, you must split a dimension of an array and bind the results to threadIdx and blockIdx.
    I guess this error occurs because three reasons:

    1. You split a inner dimension instead of an outer dimension.
    2. You split a dimension whose length is short.
    3. Maybe the split factor is too big.

    I suggest:

    1. Split the out dimension first
    2. Split dimension with big length (you can reorder the array if needet)
    3. Use a small factor(4, or 8, no big than 16)

  6. ma-xiaodong commented on Oct 28, 2017

    @ma-xiaodong
    ma-xiaodong
    on Oct 28, 2017

    #530

  7. kaishijeng commented on Oct 28, 2017

    @kaishijeng
    kaishijeng
    on Oct 28, 2017

    In order to use opencl, you must split a dimension of an array and bind the results to threadIdx and >>blockIdx.

    Which file do I need to change threadidx and blockidx? Is there a program which can tune to get optimal values of threadidx and blockidx?

  8. ma-xiaodong commented on Oct 28, 2017

    @ma-xiaodong
    ma-xiaodong
    on Oct 28, 2017

    Which file do I need to change threadidx and blockidx? Is there a program which can tune to get optimal values of threadidx and blockidx?

    Here is an example of using OpenCL on ARM Board.
    https://github.com/ma-xiaodong/tvm_tuning/blob/master/02_conv_order.py

  9. kaishijeng commented on Nov 3, 2017

    @kaishijeng
    kaishijeng
    on Nov 3, 2017

    @ma-xiaodong

    Why are you not using the following conv2d modules

    https://github.com/dmlc/tvm/blob/master/topi/python/topi/cuda/conv2d_nchw.py
    https://github.com/dmlc/tvm/blob/master/topi/python/topi/cuda/conv2d_hwcn.py

    conv2d_nchw.py looks more complex than your 02_conv_order.py.
    My firefly 3399 has a problem to pass conv2d_nchw.py due to insufficient resources.

    Thanks,

  10. tqchen
    closed this as completedon Nov 12, 2017

  11. ndcuong91 commented on Jul 4, 2018

    @ndcuong91
    ndcuong91
    on Jul 4, 2018

    @ma-xiaodong i couldn't access to your file: https://github.com/ma-xiaodong/tvm_tuning/blob/master/02_conv_order.py, can you update the link?

  12. ma-xiaodong commented on Jul 5, 2018

    @ma-xiaodong
    ma-xiaodong
    on Jul 5, 2018

    @titikid My example is too old, so I remove it. TVM already has good example.
    https://github.com/dmlc/tvm/blob/master/topi/python/topi/cuda/conv2d_hwcn.py

  13. ndcuong91 commented on Jul 5, 2018

    @ndcuong91
    ndcuong91
    on Jul 5, 2018

    @ma-xiaodong i see. Anyway, after use smaller split factor in conv2d_nchw.py as your hint i could solve my issue related to openCL. Thank you

  14. zxybazh
    added 4 commits that reference this issue on Jan 7, 2022

    Squashed commit

    f98a8e9

    Squashed commit

    d1fce89

    User-Interface: Tune-TIR (apache#525)

    28a8537

    Squashed commit

    05f3aba
  15. junrushao
    added a commit that references this issue on Jan 27, 2022

    Squashed commit

    3c6882a
  16. Hzfengsy
    added a commit that references this issue on Feb 2, 2022

    Squashed commit

    fc54ebb
  17. Hzfengsy
    added a commit that references this issue on Feb 12, 2022

    Squashed commit

    d440d79
  18. Hzfengsy
    added 2 commits that reference this issue on Feb 19, 2022

    Squashed commit

    17018cc

    Squashed commit

    623c0a6
  19. junrushao
    added 3 commits that reference this issue on Feb 20, 2022

    Squashed commit

    1d8cc54

    Squashed commit

    893555f

    Squashed commit

    46a98f1
  20. zxybazh
    added 5 commits that reference this issue on Feb 21, 2022

    Squashed commit

    eb8def7

    Squashed commit

    6f04005

    Squashed commit

    ed92495

    Squashed commit

    fc2f258

    Squashed commit

    0af0f71
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

    status: help wanted

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions