Yaxun (Sam) Liu

A recent clang change (https://github.com/llvm/llvm-project/pull/90152) revealed an issue in develop branch of composable_kernel: https://github.com/ROCm/composable_kernel/blob/08d51d9bc4ec275fce3ad0a01a08ab1fd45636bc/include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops.hpp#L799 ``` composable_kernel/include/ck/tensor_operation/gpu/block/blockwise_gemm_xdlops.hpp:799:32: error: no member named 'a_origin' in 'BlockwiseGemmXdlops_v2' 799 | : a_thread_copy_(other.a_origin), b_thread_copy_(other.b_origin) | ~~~~~...

[Issue]: calling function min with (int, unsigned int) causing ambiguity

3

### Problem Description when we try to fix a compiler bug about function min by https://github.com/llvm/llvm-project/pull/82956 we encountered build failure in hipCUB. Then we found the line that causes the...

### Problem Description when we try to fix a compiler bug about function min by https://github.com/llvm/llvm-project/pull/82956 we encountered build failure in Tensile/rocBLAS. Then we found the line that causes the...

https://github.com/ROCm/hipBLASLt/blob/04afd1fdb9ebbd99464e4ed0e79db0314242beec/tensilelite/Tensile/Source/ReductionTemplate.h#L45 There is a `constexpr max(size_t,size_t)` function defined in global namespace which conflicts with a clang PR: https://github.com/llvm/llvm-project/pull/82956 Due to this conflict, we cannot fix a clang issue. HIP apps...

[CUDA][HIP] Fix CTAD for host/device constructors (#168711)

4

Clang currently does not allow using CTAD in CUDA/HIP device functions since deduction guides are treated as host-only. This patch fixes that by treating deduction guides as host+device. The rationale...