[X86][SSE] Added extra (mul x, (1 << c)) -> x << c style vector tests

vXi64 will benefit more from lowering to shifts than multiplies

llvm-svn: 284461

diff --git a/llvm/test/CodeGen/X86/combine-mul.ll b/llvm/test/CodeGen/X86/combine-mul.ll
index 832054b..7d60a21 100644 (file)
--- a/llvm/test/CodeGen/X86/combine-mul.ll
+++ b/llvm/test/CodeGen/X86/combine-mul.ll
@@ -103,6 +103,38 @@ define <4 x i32> @combine_vec_mul_pow2b(<4 x i32> %x) {
   ret <4 x i32> %1
 }
 
+define <4 x i64> @combine_vec_mul_pow2c(<4 x i64> %x) {
+; SSE-LABEL: combine_vec_mul_pow2c:
+; SSE:       # BB#0:
+; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [1,2]
+; SSE-NEXT:    movdqa %xmm0, %xmm3
+; SSE-NEXT:    pmuludq %xmm2, %xmm3
+; SSE-NEXT:    psrlq $32, %xmm0
+; SSE-NEXT:    pmuludq %xmm2, %xmm0
+; SSE-NEXT:    psllq $32, %xmm0
+; SSE-NEXT:    paddq %xmm3, %xmm0
+; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [4,16]
+; SSE-NEXT:    movdqa %xmm1, %xmm3
+; SSE-NEXT:    pmuludq %xmm2, %xmm3
+; SSE-NEXT:    psrlq $32, %xmm1
+; SSE-NEXT:    pmuludq %xmm2, %xmm1
+; SSE-NEXT:    psllq $32, %xmm1
+; SSE-NEXT:    paddq %xmm3, %xmm1
+; SSE-NEXT:    retq
+;
+; AVX-LABEL: combine_vec_mul_pow2c:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovdqa {{.*#+}} ymm1 = [1,2,4,16]
+; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm2
+; AVX-NEXT:    vpsrlq $32, %ymm0, %ymm0
+; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
+; AVX-NEXT:    vpsllq $32, %ymm0, %ymm0
+; AVX-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; AVX-NEXT:    retq
+  %1 = mul <4 x i64> %x, <i64 1, i64 2, i64 4, i64 16>
+  ret <4 x i64> %1
+}
+
 ; fold (mul x, -(1 << c)) -> -(x << c) or (-x) << c
 define <4 x i32> @combine_vec_mul_negpow2a(<4 x i32> %x) {
 ; SSE-LABEL: combine_vec_mul_negpow2a:
@@ -137,6 +169,50 @@ define <4 x i32> @combine_vec_mul_negpow2b(<4 x i32> %x) {
   ret <4 x i32> %1
 }
 
+define <4 x i64> @combine_vec_mul_negpow2c(<4 x i64> %x) {
+; SSE-LABEL: combine_vec_mul_negpow2c:
+; SSE:       # BB#0:
+; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [18446744073709551615,18446744073709551614]
+; SSE-NEXT:    movdqa %xmm0, %xmm3
+; SSE-NEXT:    pmuludq %xmm2, %xmm3
+; SSE-NEXT:    movdqa {{.*#+}} xmm4 = [4294967295,4294967295]
+; SSE-NEXT:    movdqa %xmm0, %xmm5
+; SSE-NEXT:    pmuludq %xmm4, %xmm5
+; SSE-NEXT:    psllq $32, %xmm5
+; SSE-NEXT:    psrlq $32, %xmm0
+; SSE-NEXT:    pmuludq %xmm2, %xmm0
+; SSE-NEXT:    psllq $32, %xmm0
+; SSE-NEXT:    paddq %xmm5, %xmm0
+; SSE-NEXT:    paddq %xmm3, %xmm0
+; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [18446744073709551612,18446744073709551600]
+; SSE-NEXT:    movdqa %xmm1, %xmm3
+; SSE-NEXT:    pmuludq %xmm2, %xmm3
+; SSE-NEXT:    pmuludq %xmm1, %xmm4
+; SSE-NEXT:    psllq $32, %xmm4
+; SSE-NEXT:    psrlq $32, %xmm1
+; SSE-NEXT:    pmuludq %xmm2, %xmm1
+; SSE-NEXT:    psllq $32, %xmm1
+; SSE-NEXT:    paddq %xmm4, %xmm1
+; SSE-NEXT:    paddq %xmm3, %xmm1
+; SSE-NEXT:    retq
+;
+; AVX-LABEL: combine_vec_mul_negpow2c:
+; AVX:       # BB#0:
+; AVX-NEXT:    vmovdqa {{.*#+}} ymm1 = [18446744073709551615,18446744073709551614,18446744073709551612,18446744073709551600]
+; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm2
+; AVX-NEXT:    vpbroadcastq {{.*}}(%rip), %ymm3
+; AVX-NEXT:    vpmuludq %ymm3, %ymm0, %ymm3
+; AVX-NEXT:    vpsllq $32, %ymm3, %ymm3
+; AVX-NEXT:    vpsrlq $32, %ymm0, %ymm0
+; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
+; AVX-NEXT:    vpsllq $32, %ymm0, %ymm0
+; AVX-NEXT:    vpaddq %ymm0, %ymm3, %ymm0
+; AVX-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
+; AVX-NEXT:    retq
+  %1 = mul <4 x i64> %x, <i64 -1, i64 -2, i64 -4, i64 -16>
+  ret <4 x i64> %1
+}
+
 ; (mul (shl X, c1), c2) -> (mul X, c2 << c1)
 define <4 x i32> @combine_vec_mul_shl_const(<4 x i32> %x) {
 ; SSE-LABEL: combine_vec_mul_shl_const: