[llvm] [LLVM][NVPTX] Add NVPTX codegen support for fence.proxy.tensormap (PR #100748)

[Pradeep Kumar via llvm-commits]

================
@@ -251,6 +251,34 @@ Overview:
 The '``@llvm.nvvm.barrier0()``' intrinsic emits a PTX ``bar.sync 0``
 instruction, equivalent to the ``__syncthreads()`` call in CUDA.
 
+Membar/Fences
+-------------
+
+
+'``llvm.nvvm.fence.proxy.tensormap.*``'
----------------
schwarzschild-radius wrote:

Added `generic` explicitly in the intrinsic to mark the direction of the fencing

https://github.com/llvm/llvm-project/pull/100748