[llvm] [AMDGPU][True16][Codegen] remove another build_vector pattern from true16 (PR #149861)

[Brox Chen via llvm-commits]

================
@@ -201,11 +201,6 @@ define <10 x i16> @bitcast_i160_to_v10i16(i160 %int) {
 ; GFX12-TRUE16-NEXT:    s_wait_samplecnt 0x0
 ; GFX12-TRUE16-NEXT:    s_wait_bvhcnt 0x0
 ; GFX12-TRUE16-NEXT:    s_wait_kmcnt 0x0
-; GFX12-TRUE16-NEXT:    v_mov_b16_e32 v5.l, v0.l
----------------
broxigarchen wrote:

This removed block is equivalent to:
v0  (line 207)=> v5.l | v0.h  (line 204)=> v0.1 | v0.h = v0
v2  (line 208)=> v6.l | v2.h  (line 205)=> v2.l | v2.h = v2

It's removed since it's a NOP


https://github.com/llvm/llvm-project/pull/149861