%2 = add <4 x i32> %1, <i32 3, i32 3, i32 3, i32 3>
ret <4 x i32> %2
}
+
+define <4 x i32> @combine_vec_add_shuffle_shl(<4 x i32> %a0) {
+; SSE2-LABEL: combine_vec_add_shuffle_shl:
+; SSE2: # %bb.0:
+; SSE2-NEXT: movdqa %xmm0, %xmm1
+; SSE2-NEXT: pslld $3, %xmm1
+; SSE2-NEXT: pslld $2, %xmm0
+; SSE2-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
+; SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,3,3,0]
+; SSE2-NEXT: por {{.*}}(%rip), %xmm0
+; SSE2-NEXT: retq
+;
+; SSE41-LABEL: combine_vec_add_shuffle_shl:
+; SSE41: # %bb.0:
+; SSE41-NEXT: movdqa %xmm0, %xmm1
+; SSE41-NEXT: pslld $3, %xmm1
+; SSE41-NEXT: pslld $2, %xmm0
+; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7]
+; SSE41-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,1,0]
+; SSE41-NEXT: por {{.*}}(%rip), %xmm0
+; SSE41-NEXT: retq
+;
+; AVX-LABEL: combine_vec_add_shuffle_shl:
+; AVX: # %bb.0:
+; AVX-NEXT: vpsllvd {{.*}}(%rip), %xmm0, %xmm0
+; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,1,0]
+; 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 0, i32 1>
+ %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <4 x i32> <i32 0, i32 1, i32 1, i32 0>
+ %3 = add <4 x i32> %2, <i32 3, i32 3, i32 3, i32 3>
+ ret <4 x i32> %3
+}
%6 = uitofp <4 x i32> %5 to <4 x float>
ret <4 x float> %6
}
+
+define <2 x double> @knownbits_lshr_subvector_uitofp(<4 x i32> %x) {
+; X32-LABEL: knownbits_lshr_subvector_uitofp:
+; X32: # %bb.0:
+; X32-NEXT: vpsrld $2, %xmm0, %xmm1
+; X32-NEXT: vpsrld $1, %xmm0, %xmm0
+; X32-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7]
+; X32-NEXT: vpmovzxdq {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero
+; X32-NEXT: vmovdqa {{.*#+}} xmm1 = [4.503599627370496E+15,4.503599627370496E+15]
+; X32-NEXT: vpor %xmm1, %xmm0, %xmm0
+; X32-NEXT: vsubpd %xmm1, %xmm0, %xmm0
+; X32-NEXT: retl
+;
+; X64-LABEL: knownbits_lshr_subvector_uitofp:
+; X64: # %bb.0:
+; X64-NEXT: vpsrld $2, %xmm0, %xmm1
+; X64-NEXT: vpsrld $1, %xmm0, %xmm0
+; X64-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7]
+; X64-NEXT: vpmovzxdq {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero
+; X64-NEXT: vmovdqa {{.*#+}} xmm1 = [4.503599627370496E+15,4.503599627370496E+15]
+; X64-NEXT: vpor %xmm1, %xmm0, %xmm0
+; X64-NEXT: vsubpd %xmm1, %xmm0, %xmm0
+; X64-NEXT: retq
+ %1 = lshr <4 x i32> %x, <i32 1, i32 2, i32 0, i32 0>
+ %2 = shufflevector <4 x i32> %1, <4 x i32> undef, <2 x i32> <i32 0, i32 1>
+ %3 = uitofp <2 x i32> %2 to <2 x double>
+ ret <2 x double> %3
+}