[PATCH] D24125: [AMDGPU] Promote uniform i16 ops to i32 ops

[Tom Stellard via llvm-commits]

tstellarAMD requested changes to this revision.
This revision now requires changes to proceed.

================
Comment at: lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp:361
@@ +360,3 @@
+  // TODO: Should we promote smaller types that will be legalized to i16?
+  if (!ST->has16BitInsts() && isI16Ty(I.getType()) && DA->isUniform(&I))
+    Changed |= promoteUniformI16OpToI32Op(I);
----------------
I just noticed that this condition is wrong.  We should only be doing the promotion when the target supports 16-bit operations not when it does not support them.


https://reviews.llvm.org/D24125