Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.
And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.
[CUDA] Support popcount intrinsic
See original GitHub issueUsing tvm.popcount in a function seems to cause problem when a gpu schedule is applied. For example, the function
matmul = tvm.compute((batch, out_dim), lambda i, j: \ tvm.sum(tvm.popcount(data[i, k] ^ weight[j, k]), axis=k), \ tag='binary_dense')
runs fine on cpu but throws the error
TVMError: [04:35:49] src/codegen/llvm/codegen_llvm.cc:630: unknown intrinsic popcount
when scheduled for the GPU despite a popcount intrinsic existing for cuda. This is quite an important feature for supporting binarization which is an increasingly popular method of network optimization.
Issue Analytics
- State:
- Created 6 years ago
- Comments:9 (9 by maintainers)
Top Results From Across the Web
Top Related Medium Post
Top Related StackOverflow Question
Troubleshoot Live Code
Top Related Reddit Thread
Top Related Hackernoon Post
Top Related Tweet
Top Related Dev.to Post
Top Related Hashnode Post
#774
any update on this?