[llvm] [DAG] SimplifyMultipleUseDemandedBits - bypass ADD nodes if either operand is zero (PR #112588)
Matt Arsenault via llvm-commits
llvm-commits at lists.llvm.org
Mon Oct 28 09:01:37 PDT 2024
================
@@ -39,8 +39,8 @@ define amdgpu_kernel void @srem_i16_7(ptr addrspace(1) %out, ptr addrspace(1) %i
; TAHITI-NEXT: s_waitcnt vmcnt(0)
; TAHITI-NEXT: v_readfirstlane_b32 s0, v0
; TAHITI-NEXT: s_mulk_i32 s0, 0x4925
-; TAHITI-NEXT: s_lshr_b32 s1, s0, 31
; TAHITI-NEXT: s_ashr_i32 s0, s0, 17
+; TAHITI-NEXT: s_bfe_u32 s1, s0, 0x1000f
----------------
arsenm wrote:
This is worse. The BFE has a bigger encoding because it needs to embed the weird immediate
https://github.com/llvm/llvm-project/pull/112588
More information about the llvm-commits
mailing list