[llvm] [WIP] Fold (bitwiseop X, (add (not Y), Z)) -> (bitwiseop X, (not (sub Y, Z))). (PR #141476)

[Xu Zhang via llvm-commits]

================
@@ -162,17 +163,19 @@ define void @program_1(ptr %dest, ptr %t0, <4 x float> %p0, <4 x float> %p1, <4
 ; X64-NEXT:    movaps (%rsp), %xmm0 ## 16-byte Reload
 ; X64-NEXT:    minps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; X64-NEXT:    movaps %xmm0, (%rsp) ## 16-byte Spill
+; X64-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 ## 16-byte Reload
+; X64-NEXT:    psubd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm1
 ; X64-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
-; X64-NEXT:    psubd {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
-; X64-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
-; X64-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
-; X64-NEXT:    psubd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
+; X64-NEXT:    paddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; X64-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
 ; X64-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
-; X64-NEXT:    orps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Folded Reload
+; X64-NEXT:    xorps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0
 ; X64-NEXT:    movaps %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
-; X64-NEXT:    xorps %xmm3, %xmm3
-; X64-NEXT:    xorps %xmm4, %xmm4
+; X64-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 ## 16-byte Reload
+; X64-NEXT:	   por %xmm1, %xmm0
+; X64-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) ## 16-byte Spill
+; X64-NEXT:	   xorps %xmm3, %xmm3
+; X64-NEXT:	   xorps %xmm4, %xmm4
----------------
simonzgx wrote:

Oh! I didn’t realize this could be auto-generated! I’ll update it with your suggestion, thanks a lot.

https://github.com/llvm/llvm-project/pull/141476