Files
llvm-project/llvm/lib/CodeGen/SplitKit.cpp
Romanov Vlad a704195054 [AMDGPU] Shrink S_MOV_B64 to S_MOV_B32 during rematerialization (#184333)
When rematerializing S_MOV_B64 or S_MOV_B64_IMM_PSEUDO and only a single
32-bit lane of the result is used at the remat point, emit S_MOV_B32
with the appropriate half of the 64-bit immediate instead.

This reduces register pressure by defining a 32-bit register instead of
a 64-bit pair when the other half is unused.
2026-03-20 15:09:13 +01:00

70 KiB