-; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1\r
-; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]\r
-; AVX1-NEXT: vpmovzxwd {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero\r
-; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
+; AVX1-NEXT: vpxor %xmm1, %xmm1, %xmm1
+; AVX1-NEXT: vpshuflw {{.*#+}} xmm1 = xmm1[0,0,0,0,4,5,6,7]
+; AVX1-NEXT: vpshufd {{.*#+}} xmm2 = xmm0[2,3,0,1]
+; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3]
+; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0