Fix compilation error while in a ROCm environment#1628
Fix compilation error while in a ROCm environment#1628amcadmus merged 2 commits intodeepmodeling:develfrom
Conversation
| iteratorA[ii * last_layer_size + jj] = ; | ||
| } | ||
| } | ||
| __syncthreads(); |
There was a problem hiding this comment.
Are you sure that this line should also be removed?
There was a problem hiding this comment.
Yes, there are no shared memory access operations
There was a problem hiding this comment.
I am wondering - I do see this line in CUDA implementation
deepmd-kit/source/lib/src/cuda/tabulate.cu
Line 557 in 97c99e6
There was a problem hiding this comment.
I am wondering - I do see this line in CUDA implementation
deepmd-kit/source/lib/src/cuda/tabulate.cu
Line 557 in 97c99e6
The __syncthreads function is used to synchronise threads within the entire CUDA block, but these threads do not conflict with each other under the current implementation(There is a conflict in the previous implementation, and it is a waste to keep this __syncthreads), so this function is no longer needed.
|
@wanghan-iapcm After carefully checking, I found a variable declaration error within the test_tabulate_se_r.cc file. After fixing that error, all UTs have passed in the api_cc and lib folder within the ROCm environment. |
Codecov Report
@@ Coverage Diff @@
## devel #1628 +/- ##
=======================================
Coverage 76.16% 76.16%
=======================================
Files 94 94
Lines 7849 7849
=======================================
Hits 5978 5978
Misses 1871 1871 Continue to review full report at Codecov.
|
That's great! Thanks! |
5 participants
No description provided.