Skip to content

Fix compatibility with ROCm>=3.5.1; fix typo in hip neighbor_list#2012

Merged
wanghan-iapcm merged 18 commits intodeepmodeling:develfrom
njzjz:rocm5
Oct 27, 2022
Merged

Fix compatibility with ROCm>=3.5.1; fix typo in hip neighbor_list#2012
wanghan-iapcm merged 18 commits intodeepmodeling:develfrom
njzjz:rocm5

Conversation

@njzjz
Copy link
Member

@njzjz njzjz commented Oct 19, 2022

Fixes #1400. Fixes #2009.

  1. Uses cmake native module CMakeDetermineHIPCompiler to find the search path;
  2. for ROCm>=3.5.1, hip-hcc hiprtc is replaced by amd_comgr amdhip64, per ROCm 3.7 installation missing libraries for Tensorflow-rocm ROCm/ROCm#1200. (I am not sure about the situation of amd_comgr?)
  3. Removes -hc from the flag for ROCm>=3.5.1.
  4. Bumps from C++11 to C++14 as C++ 14 required by amd_comgr.
  5. Removes --amdgpu-target=gfx906. I don't see the reason why it is in the flag.
  6. Fixes a typo in Upload attention based model #1866.

njzjz added 11 commits October 11, 2022 19:45
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
@njzjz
Copy link
Member Author

njzjz commented Oct 19, 2022

For a CI environment, I am waiting for ROCm/ROCm#1713 to be resolved...

@codecov-commenter
Copy link

codecov-commenter commented Oct 19, 2022

Codecov Report

Base: 73.58% // Head: 73.85% // Increases project coverage by +0.26% 🎉

Coverage data is based on head (58b2d22) compared to base (c1f9e55).
Patch has no changes to coverable lines.

Additional details and impacted files
@@            Coverage Diff             @@
##            devel    #2012      +/-   ##
==========================================
+ Coverage   73.58%   73.85%   +0.26%     
==========================================
  Files         198      198              
  Lines       19908    19908              
  Branches      854      865      +11     
==========================================
+ Hits        14650    14703      +53     
+ Misses       4870     4813      -57     
- Partials      388      392       +4     
Impacted Files Coverage Δ
source/lib/include/ComputeDescriptor.h 11.49% <0.00%> (+11.49%) ⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

@njzjz
Copy link
Member Author

njzjz commented Oct 19, 2022

@galeselee please help review this PR.

@galeselee
Copy link
Contributor

In addition, the previous rocm-4 has the problem that the community user feedback calculation results are not correct, but UT can pass.

@galeselee
Copy link
Contributor

LGTM

@njzjz njzjz marked this pull request as draft October 19, 2022 07:40
@njzjz njzjz marked this pull request as ready for review October 19, 2022 08:36
@github-actions github-actions bot added the Docs label Oct 19, 2022
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
njzjz added 6 commits October 26, 2022 18:53
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
add unit tests for rocm
@wanghan-iapcm wanghan-iapcm merged commit 00c1b9e into deepmodeling:devel Oct 27, 2022
@njzjz njzjz linked an issue Oct 27, 2022 that may be closed by this pull request
mingzhong15 pushed a commit to mingzhong15/deepmd-kit that referenced this pull request Jan 15, 2023
…epmodeling#2012)

Fixes deepmodeling#1400. Fixes deepmodeling#2009.

1. Uses cmake native module `CMakeDetermineHIPCompiler` to find the
search path;
2. for ROCm>=3.5.1, `hip-hcc hiprtc` is replaced by `amd_comgr
amdhip64`, per ROCm/ROCm#1200. (I
am not sure about the situation of `amd_comgr`?)
3. Removes `-hc` from the flag for ROCm>=3.5.1.
4. Bumps from C++11 to C++14 as C++ 14 required by `amd_comgr`.
5. Removes `--amdgpu-target=gfx906`. I don't see the reason why it is in
the flag.
6. Fixes a typo in deepmodeling#1866.

Signed-off-by: Jinzhe Zeng <jinzhe.zeng@rutgers.edu>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

5 participants