bcahoon 17c8c1c509 [AMDGPU] Do not fold into v_accvpr_mov/write/read (#120475) ...

In SIFoldOperands, leave copies for moving between agpr and vgpr
registers. The register coalescer is able to handle the copies
more efficiently than v_accvgpr_mov, v_accvgpr_write, and
v_accvgpr_read. Otherwise, the compiler generates unneccesary
instructions such as v_accvgpr_mov a0, a0.

2025-01-07 09:25:01 -06:00

80 KiB

Raw Blame History

View Raw