[llvm] [NVPTX] Add TMA bulk tensor reduction intrinsics (PR #116854)

[Durgadoss R via llvm-commits]

================
@@ -4157,9 +4157,9 @@ bool NVPTXScopes::empty() const { return Scopes.size() == 0; }
        ? NVPTX::CP_ASYNC_BULK_TENSOR_##dir##_##dim##_SHARED32_##mode##suffix   \
        : NVPTX::CP_ASYNC_BULK_TENSOR_##dir##_##dim##_##mode##suffix)
 
-#define GET_CP_ASYNC_BULK_TENSOR_OPCODE_S2G(dim, mode)                         \
-  (IsCacheHint ? (CP_ASYNC_BULK_TENSOR_OPCODE(S2G, dim, mode, _CH))            \
-               : (CP_ASYNC_BULK_TENSOR_OPCODE(S2G, dim, mode, )))
+#define GET_CP_ASYNC_BULK_TENSOR_OPCODE_CH(op, dim, mode)                      \
+  (IsCacheHint ? (CP_ASYNC_BULK_TENSOR_OPCODE(op, dim, mode, _CH))             \
----------------
durga4github wrote:

Yes, I made all the macros self-contained in the latest revision.

Resolving.

https://github.com/llvm/llvm-project/pull/116854