; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_03_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,3,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 3, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_04_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,4,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 4, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_05_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,5,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 5, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_06_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,6,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 6, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,9,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,9,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,9,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,9,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_09_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,9,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 9, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,10,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,10,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,10,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,10,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_10_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,10,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 10, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,11,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,11,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,11,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,11,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_11_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,11,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 11, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,12,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,12,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,12,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,12,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_12_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,12,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 12, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,13,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,13,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,13,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,13,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_13_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,13,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 13, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,14,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,14,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,14,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,14,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 14, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,15,0,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,15,0,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,15,0,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,15,0,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_01_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_02_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_07_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,7,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 7, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_08_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,8,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 8, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: vpshufb {{.*#+}} xmm0 = xmm0[0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_14_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,14,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 0, i32 14, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 14, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: movl $15, %eax
-; AVX2OR512VL-NEXT: vmovd %eax, %xmm1
-; AVX2OR512VL-NEXT: vpshufb %xmm1, %xmm0, %xmm0
-; AVX2OR512VL-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-LABEL: shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX2: # %bb.0:
+; AVX2-NEXT: movl $15, %eax
+; AVX2-NEXT: vmovd %eax, %xmm1
+; AVX2-NEXT: vpshufb %xmm1, %xmm0, %xmm0
+; AVX2-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-NEXT: retq
+;
+; AVX512VLBW-LABEL: shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLBW: # %bb.0:
+; AVX512VLBW-NEXT: movl $15, %eax
+; AVX512VLBW-NEXT: vmovd %eax, %xmm1
+; AVX512VLBW-NEXT: vpshufb %xmm1, %xmm0, %xmm0
+; AVX512VLBW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLBW-NEXT: retq
+;
+; AVX512VLVBMI-SLOW-LABEL: shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-SLOW: # %bb.0:
+; AVX512VLVBMI-SLOW-NEXT: movl $15, %eax
+; AVX512VLVBMI-SLOW-NEXT: vmovd %eax, %xmm1
+; AVX512VLVBMI-SLOW-NEXT: vpshufb %xmm1, %xmm0, %xmm0
+; AVX512VLVBMI-SLOW-NEXT: vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VLVBMI-SLOW-NEXT: retq
+;
+; AVX512VLVBMI-FAST-LABEL: shuffle_v32i8_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00_15_00_00_00_00_00_00_00_00_00_00_00_00_00_00_00:
+; AVX512VLVBMI-FAST: # %bb.0:
+; AVX512VLVBMI-FAST-NEXT: vmovdqa {{.*#+}} ymm1 = [15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,15,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0]
+; AVX512VLVBMI-FAST-NEXT: vpermb %ymm0, %ymm1, %ymm0
+; AVX512VLVBMI-FAST-NEXT: retq
%shuffle = shufflevector <32 x i8> %a, <32 x i8> %b, <32 x i32> <i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 15, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 0>
ret <32 x i8> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8f32_00000010:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_00000010:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8f32_00000010:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,0,1,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8f32_00000010:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8f32_00000010:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,0,1,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
ret <8 x float> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8f32_00000200:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,2]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_00000200:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,2]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8f32_00000200:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,2,0,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8f32_00000200:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,2]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8f32_00000200:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,2,0,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
ret <8 x float> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8f32_00003000:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,3,0]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_00003000:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,3,0]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8f32_00003000:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,3,0,0,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8f32_00003000:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,3,0]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8f32_00003000:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,3,0,0,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 3, i32 0, i32 0, i32 0>
ret <8 x float> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8f32_00001111:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_00001111:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,1]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8f32_00001111:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,1,1,1,1]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8f32_00001111:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,1]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8f32_00001111:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,1,1,1,1]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1>
ret <8 x float> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm2, %ymm0
; AVX1-NEXT: retq
;
-; AVX2-LABEL: shuffle_v8f32_08991abb:
-; AVX2: # %bb.0:
-; AVX2-NEXT: vmovaps {{.*#+}} ymm2 = <u,0,1,1,u,2,3,3>
-; AVX2-NEXT: vpermps %ymm1, %ymm2, %ymm1
-; AVX2-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,1,1,3]
-; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
-; AVX2-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
-; AVX2-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_08991abb:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vmovaps {{.*#+}} ymm2 = <u,0,1,1,u,2,3,3>
+; AVX2-SLOW-NEXT: vpermps %ymm1, %ymm2, %ymm1
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,1,1,3]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
+; AVX2-SLOW-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8f32_08991abb:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm2 = <0,1,1,3,1,3,u,u>
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm2, %ymm0
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm2 = <u,0,1,1,u,2,3,3>
+; AVX2-FAST-NEXT: vpermps %ymm1, %ymm2, %ymm1
+; AVX2-FAST-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
+; AVX2-FAST-NEXT: retq
;
; AVX512VL-LABEL: shuffle_v8f32_08991abb:
; AVX512VL: # %bb.0:
-; AVX512VL-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[0,1,1,3]
-; AVX512VL-NEXT: vmovaps {{.*#+}} ymm0 = [8,0,1,1,10,2,3,3]
-; AVX512VL-NEXT: vpermi2ps %ymm2, %ymm1, %ymm0
+; AVX512VL-NEXT: vmovaps {{.*#+}} ymm2 = [8,0,1,1,9,2,3,3]
+; AVX512VL-NEXT: vpermi2ps %ymm0, %ymm1, %ymm2
+; AVX512VL-NEXT: vmovaps %ymm2, %ymm0
; AVX512VL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 8, i32 9, i32 9, i32 1, i32 10, i32 11, i32 11>
ret <8 x float> %shuffle
; AVX1-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
; AVX1-NEXT: retq
;
-; AVX2-LABEL: shuffle_v8f32_09ab1def:
-; AVX2: # %bb.0:
-; AVX2-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,1,1,3]
-; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
-; AVX2-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
-; AVX2-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_09ab1def:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,1,1,3]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
+; AVX2-SLOW-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
+; AVX2-SLOW-NEXT: retq
;
-; AVX512VL-SLOW-LABEL: shuffle_v8f32_09ab1def:
-; AVX512VL-SLOW: # %bb.0:
-; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,1,1,3]
-; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,1,3]
-; AVX512VL-SLOW-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
-; AVX512VL-SLOW-NEXT: retq
+; AVX2-FAST-LABEL: shuffle_v8f32_09ab1def:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm2 = <0,1,1,3,1,3,u,u>
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm2, %ymm0
+; AVX2-FAST-NEXT: vblendps {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7]
+; AVX2-FAST-NEXT: retq
;
-; AVX512VL-FAST-LABEL: shuffle_v8f32_09ab1def:
-; AVX512VL-FAST: # %bb.0:
-; AVX512VL-FAST-NEXT: vpermilps {{.*#+}} xmm2 = xmm0[0,1,1,3]
-; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm0 = [8,1,2,3,10,5,6,7]
-; AVX512VL-FAST-NEXT: vpermi2ps %ymm2, %ymm1, %ymm0
-; AVX512VL-FAST-NEXT: retq
+; AVX512VL-LABEL: shuffle_v8f32_09ab1def:
+; AVX512VL: # %bb.0:
+; AVX512VL-NEXT: vmovaps {{.*#+}} ymm2 = [8,1,2,3,9,5,6,7]
+; AVX512VL-NEXT: vpermi2ps %ymm0, %ymm1, %ymm2
+; AVX512VL-NEXT: vmovaps %ymm2, %ymm0
+; AVX512VL-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 0, i32 9, i32 10, i32 11, i32 1, i32 13, i32 14, i32 15>
ret <8 x float> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8f32_32103210:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8f32_32103210:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8f32_32103210:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [3,2,1,0,3,2,1,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8f32_32103210:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8f32_32103210:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [3,2,1,0,3,2,1,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x float> %a, <8 x float> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>
ret <8 x float> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8i32_00000010:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8i32_00000010:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8i32_00000010:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,0,1,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8i32_00000010:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,0]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,0,1]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8i32_00000010:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,0,1,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 0, i32 1, i32 0>
ret <8 x i32> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8i32_00000200:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,2]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8i32_00000200:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,2]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8i32_00000200:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,2,0,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8i32_00000200:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,2]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8i32_00000200:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,0,2,0,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 0, i32 2, i32 0, i32 0>
ret <8 x i32> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8i32_00003000:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,3,0]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8i32_00003000:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,3,0]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8i32_00003000:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,3,0,0,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8i32_00003000:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,3,0]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,0]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8i32_00003000:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,3,0,0,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 3, i32 0, i32 0, i32 0>
ret <8 x i32> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8i32_00001111:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8i32_00001111:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,1]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8i32_00001111:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,1,1,1,1]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8i32_00001111:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,1,1]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,0,1,1]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8i32_00001111:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [0,0,0,0,1,1,1,1]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 1, i32 1, i32 1, i32 1>
ret <8 x i32> %shuffle
}
; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0
; AVX1-NEXT: retq
;
-; AVX2OR512VL-LABEL: shuffle_v8i32_32103210:
-; AVX2OR512VL: # %bb.0:
-; AVX2OR512VL-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
-; AVX2OR512VL-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
-; AVX2OR512VL-NEXT: retq
+; AVX2-SLOW-LABEL: shuffle_v8i32_32103210:
+; AVX2-SLOW: # %bb.0:
+; AVX2-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; AVX2-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX2-SLOW-NEXT: retq
+;
+; AVX2-FAST-LABEL: shuffle_v8i32_32103210:
+; AVX2-FAST: # %bb.0:
+; AVX2-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [3,2,1,0,3,2,1,0]
+; AVX2-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX2-FAST-NEXT: retq
+;
+; AVX512VL-SLOW-LABEL: shuffle_v8i32_32103210:
+; AVX512VL-SLOW: # %bb.0:
+; AVX512VL-SLOW-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[3,2,1,0]
+; AVX512VL-SLOW-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,1,0,1]
+; AVX512VL-SLOW-NEXT: retq
+;
+; AVX512VL-FAST-LABEL: shuffle_v8i32_32103210:
+; AVX512VL-FAST: # %bb.0:
+; AVX512VL-FAST-NEXT: vmovaps {{.*#+}} ymm1 = [3,2,1,0,3,2,1,0]
+; AVX512VL-FAST-NEXT: vpermps %ymm0, %ymm1, %ymm0
+; AVX512VL-FAST-NEXT: retq
%shuffle = shufflevector <8 x i32> %a, <8 x i32> %b, <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0>
ret <8 x i32> %shuffle
}