Skip to content

Commit 85c79d6

Browse files
committed
[X86] Add test coverage for #103564
1 parent 8aacbfc commit 85c79d6

File tree

1 file changed

+69
-0
lines changed

1 file changed

+69
-0
lines changed

llvm/test/CodeGen/X86/vector-shuffle-256-v32.ll

Lines changed: 69 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -5187,6 +5187,75 @@ define <4 x i64> @PR66150(ptr %b) {
51875187
ret <4 x i64> %tmp6
51885188
}
51895189

5190+
define <64 x i8> @PR103564(<32 x i8> %a0, <32 x i8> %a1) {
5191+
; AVX1-LABEL: PR103564:
5192+
; AVX1: # %bb.0:
5193+
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
5194+
; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm3 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
5195+
; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm3, %ymm2
5196+
; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
5197+
; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
5198+
; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
5199+
; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
5200+
; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm0, %ymm1
5201+
; AVX1-NEXT: vmovaps %ymm2, %ymm0
5202+
; AVX1-NEXT: retq
5203+
;
5204+
; AVX2-LABEL: PR103564:
5205+
; AVX2: # %bb.0:
5206+
; AVX2-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
5207+
; AVX2-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
5208+
; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm1[0,1],ymm2[0,1]
5209+
; AVX2-NEXT: vperm2i128 {{.*#+}} ymm1 = ymm1[2,3],ymm2[2,3]
5210+
; AVX2-NEXT: retq
5211+
;
5212+
; AVX512VLBW-LABEL: PR103564:
5213+
; AVX512VLBW: # %bb.0:
5214+
; AVX512VLBW-NEXT: vextracti128 $1, %ymm1, %xmm2
5215+
; AVX512VLBW-NEXT: vextracti128 $1, %ymm0, %xmm3
5216+
; AVX512VLBW-NEXT: vpunpckhbw {{.*#+}} xmm4 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
5217+
; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} xmm2 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
5218+
; AVX512VLBW-NEXT: vinserti128 $1, %xmm4, %ymm2, %ymm2
5219+
; AVX512VLBW-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
5220+
; AVX512VLBW-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
5221+
; AVX512VLBW-NEXT: vinserti128 $1, %xmm3, %ymm0, %ymm0
5222+
; AVX512VLBW-NEXT: vinserti64x4 $1, %ymm2, %zmm0, %zmm0
5223+
; AVX512VLBW-NEXT: retq
5224+
;
5225+
; AVX512VLVBMI-LABEL: PR103564:
5226+
; AVX512VLVBMI: # %bb.0:
5227+
; AVX512VLVBMI-NEXT: # kill: def $ymm1 killed $ymm1 def $zmm1
5228+
; AVX512VLVBMI-NEXT: # kill: def $ymm0 killed $ymm0 def $zmm0
5229+
; AVX512VLVBMI-NEXT: vinserti64x4 $1, %ymm1, %zmm1, %zmm1
5230+
; AVX512VLVBMI-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm2
5231+
; AVX512VLVBMI-NEXT: vmovdqa64 {{.*#+}} zmm0 = [0,64,1,65,2,66,3,67,4,68,5,69,6,70,7,71,8,72,9,73,10,74,11,75,12,76,13,77,14,78,15,79,48,112,49,113,50,114,51,115,52,116,53,117,54,118,55,119,56,120,57,121,58,122,59,123,60,124,61,125,62,126,63,127]
5232+
; AVX512VLVBMI-NEXT: vpermi2b %zmm1, %zmm2, %zmm0
5233+
; AVX512VLVBMI-NEXT: retq
5234+
;
5235+
; XOPAVX1-LABEL: PR103564:
5236+
; XOPAVX1: # %bb.0:
5237+
; XOPAVX1-NEXT: vpunpckhbw {{.*#+}} xmm2 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
5238+
; XOPAVX1-NEXT: vpunpcklbw {{.*#+}} xmm3 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
5239+
; XOPAVX1-NEXT: vinsertf128 $1, %xmm2, %ymm3, %ymm2
5240+
; XOPAVX1-NEXT: vextractf128 $1, %ymm1, %xmm1
5241+
; XOPAVX1-NEXT: vextractf128 $1, %ymm0, %xmm0
5242+
; XOPAVX1-NEXT: vpunpckhbw {{.*#+}} xmm3 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
5243+
; XOPAVX1-NEXT: vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
5244+
; XOPAVX1-NEXT: vinsertf128 $1, %xmm3, %ymm0, %ymm1
5245+
; XOPAVX1-NEXT: vmovaps %ymm2, %ymm0
5246+
; XOPAVX1-NEXT: retq
5247+
;
5248+
; XOPAVX2-LABEL: PR103564:
5249+
; XOPAVX2: # %bb.0:
5250+
; XOPAVX2-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
5251+
; XOPAVX2-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
5252+
; XOPAVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm1[0,1],ymm2[0,1]
5253+
; XOPAVX2-NEXT: vperm2i128 {{.*#+}} ymm1 = ymm1[2,3],ymm2[2,3]
5254+
; XOPAVX2-NEXT: retq
5255+
%r = shufflevector <32 x i8> %a0, <32 x i8> %a1, <64 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
5256+
ret <64 x i8> %r
5257+
}
5258+
51905259
define <32 x i8> @insert_dup_mem_v32i8_i32(ptr %ptr) {
51915260
; AVX1-LABEL: insert_dup_mem_v32i8_i32:
51925261
; AVX1: # %bb.0:

0 commit comments

Comments
 (0)