[llvm] [X86] Narrow vXi32/vXi64 usubsat to vpsubusb/vpsubusw when LHS is known to fit in fewer bits (PR #206592)

[Phoebe Wang via llvm-commits]

================
@@ -100,15 +100,18 @@ define i8 @concat_icmp_v8i32_v4i32(<4 x i32> %a0, <4 x i32> %a1) {
 define i16 @concat_icmp_v16i16_v8i16(<8 x i16> %a0, <8 x i16> %a1) {
 ; SSE2-LABEL: concat_icmp_v16i16_v8i16:
 ; SSE2:       # %bb.0:
-; SSE2-NEXT:    movdqa {{.*#+}} xmm2 = [2,2,2,2,2,2,2,2]
-; SSE2-NEXT:    movdqa %xmm2, %xmm3
-; SSE2-NEXT:    psubusw %xmm0, %xmm3
----------------
phoebewang wrote:

The date type is v8i16 already. Narrowing to i8 increase instructions.

https://github.com/llvm/llvm-project/pull/206592