[llvm] [NVPTX] Add Intrinsics for discard.* (PR #128404)

[Durgadoss R via llvm-commits]

================
@@ -0,0 +1,35 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
+; RUN: llc < %s -mtriple=nvptx64 -mcpu=sm_80 -mattr=+ptx74| FileCheck --check-prefixes=CHECK-PTX64 %s
+; RUN: %if ptxas-11.4 %{ llc < %s -mtriple=nvptx64 -mcpu=sm_80 -mattr=+ptx74| %ptxas-verify -arch=sm_80 %}
+
+target triple = "nvptx64-nvidia-cuda"
+
+declare void  @llvm.nvvm.discard.global.L2(ptr addrspace(1) %global_ptr, i64 immarg)
----------------
durga4github wrote:

nit: Do we want to use an arg-name (to be consistent with the the ptr-arg)? 
i64 immarg %size?

https://github.com/llvm/llvm-project/pull/128404