[PATCH] D18358: AMDGPU: Account for globals in AMDGPUPromoteAlloca pass

Bas Nieuwenhuizen via llvm-commits llvm-commits at lists.llvm.org
Tue Mar 22 11:07:55 PDT 2016

bnieuwenhuizen added inline comments.

Comment at: test/CodeGen/AMDGPU/promote-alloca-globals.ll:1
@@ +1,2 @@
+; RUN: opt -S -mtriple=amdgcn-unknown-unknown -amdgpu-promote-alloca < %s | FileCheck %s
arsenm wrote:
> You should also include an llc run line and check for the final LDS usage
How do you check for LDS usage? I can't find anything about LDS except the numerical representations of the RSRC registers.

Comment at: test/CodeGen/AMDGPU/promote-alloca-globals.ll:24
@@ +23,3 @@
+  store i32 %3, i32 addrspace(1)* %arrayidx13
+  %v = getelementptr inbounds [1500 x [5 x i32]], [1500 x [5 x i32]] addrspace(3)* @promote_alloca_size_1500.stack, i32 0, i32 0
+  ret void
arsenm wrote:
> This GEP isn't used? 
It isn't as we just need a user of the global array to be used by an instruction inside the function to let the AMDGPUPromoteAlloca associate it with this function. I'll probably add a user though, to prevent llc from optimizing the LDS array away.


More information about the llvm-commits mailing list