diff options
| -rw-r--r-- | llvm/test/CodeGen/X86/combine-shl.ll | 33 |
1 files changed, 33 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/X86/combine-shl.ll b/llvm/test/CodeGen/X86/combine-shl.ll index 55ad952c359..40971dc3eaf 100644 --- a/llvm/test/CodeGen/X86/combine-shl.ll +++ b/llvm/test/CodeGen/X86/combine-shl.ll @@ -832,3 +832,36 @@ define <4 x i32> @combine_vec_shl_mul1(<4 x i32> %x) { %2 = shl <4 x i32> %1, <i32 1, i32 2, i32 3, i32 4> ret <4 x i32> %2 } + +; fold (add (shl x, c1), c2) -> (or (shl x, c1), c2) +; TODO: Handle minimum shift value case +define <4 x i32> @combine_vec_add_shl_nonsplat(<4 x i32> %a0) { +; SSE2-LABEL: combine_vec_add_shl_nonsplat: +; SSE2: # %bb.0: +; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [4,8,16,32] +; SSE2-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3] +; SSE2-NEXT: pmuludq %xmm1, %xmm0 +; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3] +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3] +; SSE2-NEXT: pmuludq %xmm2, %xmm1 +; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3] +; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1] +; SSE2-NEXT: paddd {{.*}}(%rip), %xmm0 +; SSE2-NEXT: retq +; +; SSE41-LABEL: combine_vec_add_shl_nonsplat: +; SSE41: # %bb.0: +; SSE41-NEXT: pmulld {{.*}}(%rip), %xmm0 +; SSE41-NEXT: por {{.*}}(%rip), %xmm0 +; SSE41-NEXT: retq +; +; AVX-LABEL: combine_vec_add_shl_nonsplat: +; AVX: # %bb.0: +; AVX-NEXT: vpsllvd {{.*}}(%rip), %xmm0, %xmm0 +; AVX-NEXT: vpbroadcastd {{.*#+}} xmm1 = [3,3,3,3] +; AVX-NEXT: vpaddd %xmm1, %xmm0, %xmm0 +; AVX-NEXT: retq + %1 = shl <4 x i32> %a0, <i32 2, i32 3, i32 4, i32 5> + %2 = add <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3> + ret <4 x i32> %2 +} |
