Lines Matching full:tmp4
97 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[LANE]] to <8 x half>
98 // CHECK: ret <8 x half> [[TMP4]]
246 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[LANE]] to <4 x half>
247 // CHECK: ret <4 x half> [[TMP4]]
315 // CHECK: [[TMP4:%.*]] = load %struct.uint8x16x2_t, %struct.uint8x16x2_t* [[RETVAL]], align 16
316 // CHECK: ret %struct.uint8x16x2_t [[TMP4]]
330 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x2_t* [[RETVAL]] to i8*
332 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
348 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x2_t* [[RETVAL]] to i8*
350 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
366 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x2_t* [[RETVAL]] to i8*
368 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
385 // CHECK: [[TMP4:%.*]] = load %struct.int8x16x2_t, %struct.int8x16x2_t* [[RETVAL]], align 16
386 // CHECK: ret %struct.int8x16x2_t [[TMP4]]
400 // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x2_t* [[RETVAL]] to i8*
402 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
418 // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x2_t* [[RETVAL]] to i8*
420 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
436 // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x2_t* [[RETVAL]] to i8*
438 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
454 // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x2_t* [[RETVAL]] to i8*
456 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
472 // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x2_t* [[RETVAL]] to i8*
474 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
490 // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x2_t* [[RETVAL]] to i8*
492 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
509 // CHECK: [[TMP4:%.*]] = load %struct.poly8x16x2_t, %struct.poly8x16x2_t* [[RETVAL]], align 16
510 // CHECK: ret %struct.poly8x16x2_t [[TMP4]]
524 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x2_t* [[RETVAL]] to i8*
526 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
542 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x2x2_t* [[RETVAL]] to i8*
544 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 16, i1 false)
561 // CHECK: [[TMP4:%.*]] = load %struct.uint8x8x2_t, %struct.uint8x8x2_t* [[RETVAL]], align 8
562 // CHECK: ret %struct.uint8x8x2_t [[TMP4]]
576 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x2_t* [[RETVAL]] to i8*
578 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
594 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x2_t* [[RETVAL]] to i8*
596 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
612 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x2_t* [[RETVAL]] to i8*
614 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
631 // CHECK: [[TMP4:%.*]] = load %struct.int8x8x2_t, %struct.int8x8x2_t* [[RETVAL]], align 8
632 // CHECK: ret %struct.int8x8x2_t [[TMP4]]
646 // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x2_t* [[RETVAL]] to i8*
648 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
664 // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x2_t* [[RETVAL]] to i8*
666 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
682 // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x2_t* [[RETVAL]] to i8*
684 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
700 // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x2_t* [[RETVAL]] to i8*
702 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
718 // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x2_t* [[RETVAL]] to i8*
720 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
736 // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x2_t* [[RETVAL]] to i8*
738 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
755 // CHECK: [[TMP4:%.*]] = load %struct.poly8x8x2_t, %struct.poly8x8x2_t* [[RETVAL]], align 8
756 // CHECK: ret %struct.poly8x8x2_t [[TMP4]]
770 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x2_t* [[RETVAL]] to i8*
772 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
788 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x1x2_t* [[RETVAL]] to i8*
790 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 16, i32 8, i1 false)
807 // CHECK: [[TMP4:%.*]] = load %struct.uint8x16x3_t, %struct.uint8x16x3_t* [[RETVAL]], align 16
808 // CHECK: ret %struct.uint8x16x3_t [[TMP4]]
823 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x3_t* [[RETVAL]] to i8*
825 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
842 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x3_t* [[RETVAL]] to i8*
844 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
861 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x3_t* [[RETVAL]] to i8*
863 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
881 // CHECK: [[TMP4:%.*]] = load %struct.int8x16x3_t, %struct.int8x16x3_t* [[RETVAL]], align 16
882 // CHECK: ret %struct.int8x16x3_t [[TMP4]]
897 // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x3_t* [[RETVAL]] to i8*
899 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
916 // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x3_t* [[RETVAL]] to i8*
918 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
935 // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x3_t* [[RETVAL]] to i8*
937 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
954 // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x3_t* [[RETVAL]] to i8*
956 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
973 // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x3_t* [[RETVAL]] to i8*
975 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
992 // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x3_t* [[RETVAL]] to i8*
994 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
1012 // CHECK: [[TMP4:%.*]] = load %struct.poly8x16x3_t, %struct.poly8x16x3_t* [[RETVAL]], align 16
1013 // CHECK: ret %struct.poly8x16x3_t [[TMP4]]
1028 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x3_t* [[RETVAL]] to i8*
1030 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
1047 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x2x3_t* [[RETVAL]] to i8*
1049 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 48, i32 16, i1 false)
1067 // CHECK: [[TMP4:%.*]] = load %struct.uint8x8x3_t, %struct.uint8x8x3_t* [[RETVAL]], align 8
1068 // CHECK: ret %struct.uint8x8x3_t [[TMP4]]
1083 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x3_t* [[RETVAL]] to i8*
1085 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1102 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x3_t* [[RETVAL]] to i8*
1104 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1121 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x3_t* [[RETVAL]] to i8*
1123 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1141 // CHECK: [[TMP4:%.*]] = load %struct.int8x8x3_t, %struct.int8x8x3_t* [[RETVAL]], align 8
1142 // CHECK: ret %struct.int8x8x3_t [[TMP4]]
1157 // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x3_t* [[RETVAL]] to i8*
1159 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1176 // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x3_t* [[RETVAL]] to i8*
1178 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1195 // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x3_t* [[RETVAL]] to i8*
1197 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1214 // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x3_t* [[RETVAL]] to i8*
1216 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1233 // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x3_t* [[RETVAL]] to i8*
1235 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1252 // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x3_t* [[RETVAL]] to i8*
1254 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1272 // CHECK: [[TMP4:%.*]] = load %struct.poly8x8x3_t, %struct.poly8x8x3_t* [[RETVAL]], align 8
1273 // CHECK: ret %struct.poly8x8x3_t [[TMP4]]
1288 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x3_t* [[RETVAL]] to i8*
1290 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1307 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x1x3_t* [[RETVAL]] to i8*
1309 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 24, i32 8, i1 false)
1327 // CHECK: [[TMP4:%.*]] = load %struct.uint8x16x4_t, %struct.uint8x16x4_t* [[RETVAL]], align 16
1328 // CHECK: ret %struct.uint8x16x4_t [[TMP4]]
1342 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x8x4_t* [[RETVAL]] to i8*
1344 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1360 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x4x4_t* [[RETVAL]] to i8*
1362 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1378 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x2x4_t* [[RETVAL]] to i8*
1380 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1397 // CHECK: [[TMP4:%.*]] = load %struct.int8x16x4_t, %struct.int8x16x4_t* [[RETVAL]], align 16
1398 // CHECK: ret %struct.int8x16x4_t [[TMP4]]
1412 // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x8x4_t* [[RETVAL]] to i8*
1414 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1430 // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x4x4_t* [[RETVAL]] to i8*
1432 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1448 // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x2x4_t* [[RETVAL]] to i8*
1450 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1466 // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x8x4_t* [[RETVAL]] to i8*
1468 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1484 // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x4x4_t* [[RETVAL]] to i8*
1486 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1502 // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x2x4_t* [[RETVAL]] to i8*
1504 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1521 // CHECK: [[TMP4:%.*]] = load %struct.poly8x16x4_t, %struct.poly8x16x4_t* [[RETVAL]], align 16
1522 // CHECK: ret %struct.poly8x16x4_t [[TMP4]]
1536 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x8x4_t* [[RETVAL]] to i8*
1538 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1554 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x2x4_t* [[RETVAL]] to i8*
1556 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 64, i32 16, i1 false)
1573 // CHECK: [[TMP4:%.*]] = load %struct.uint8x8x4_t, %struct.uint8x8x4_t* [[RETVAL]], align 8
1574 // CHECK: ret %struct.uint8x8x4_t [[TMP4]]
1588 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint16x4x4_t* [[RETVAL]] to i8*
1590 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1606 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint32x2x4_t* [[RETVAL]] to i8*
1608 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1624 // CHECK: [[TMP4:%.*]] = bitcast %struct.uint64x1x4_t* [[RETVAL]] to i8*
1626 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1643 // CHECK: [[TMP4:%.*]] = load %struct.int8x8x4_t, %struct.int8x8x4_t* [[RETVAL]], align 8
1644 // CHECK: ret %struct.int8x8x4_t [[TMP4]]
1658 // CHECK: [[TMP4:%.*]] = bitcast %struct.int16x4x4_t* [[RETVAL]] to i8*
1660 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1676 // CHECK: [[TMP4:%.*]] = bitcast %struct.int32x2x4_t* [[RETVAL]] to i8*
1678 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1694 // CHECK: [[TMP4:%.*]] = bitcast %struct.int64x1x4_t* [[RETVAL]] to i8*
1696 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1712 // CHECK: [[TMP4:%.*]] = bitcast %struct.float16x4x4_t* [[RETVAL]] to i8*
1714 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1730 // CHECK: [[TMP4:%.*]] = bitcast %struct.float32x2x4_t* [[RETVAL]] to i8*
1732 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1748 // CHECK: [[TMP4:%.*]] = bitcast %struct.float64x1x4_t* [[RETVAL]] to i8*
1750 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1767 // CHECK: [[TMP4:%.*]] = load %struct.poly8x8x4_t, %struct.poly8x8x4_t* [[RETVAL]], align 8
1768 // CHECK: ret %struct.poly8x8x4_t [[TMP4]]
1782 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly16x4x4_t* [[RETVAL]] to i8*
1784 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1800 // CHECK: [[TMP4:%.*]] = bitcast %struct.poly64x1x4_t* [[RETVAL]] to i8*
1802 // CHECK: call void @llvm.memcpy.p0i8.p0i8.i64(i8* [[TMP4]], i8* [[TMP5]], i64 32, i32 8, i1 false)
1822 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
1823 // CHECK: [[VLD1_LANE:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[TMP4]], i32 7
1834 // CHECK: [[TMP4:%.*]] = load i32, i32* [[TMP3]]
1835 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[TMP4]], i32 3
1846 // CHECK: [[TMP4:%.*]] = load i64, i64* [[TMP3]]
1847 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP4]], i32 1
1866 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
1867 TMP4]], i32 7
1878 // CHECK: [[TMP4:%.*]] = load i32, i32* [[TMP3]]
1879 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x i32> [[TMP2]], i32 [[TMP4]], i32 3
1890 // CHECK: [[TMP4:%.*]] = load i64, i64* [[TMP3]]
1891 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP4]], i32 1
1902 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
1903 // CHECK: [[VLD1_LANE:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[TMP4]], i32 7
1915 // CHECK: [[TMP4:%.*]] = load float, float* [[TMP3]]
1916 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x float> [[TMP2]], float [[TMP4]], i32 3
1927 // CHECK: [[TMP4:%.*]] = load double, double* [[TMP3]]
1928 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x double> [[TMP2]], double [[TMP4]], i32 1
1947 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
1948 // CHECK: [[VLD1_LANE:%.*]] = insertelement <8 x i16> [[TMP2]], i16 [[TMP4]], i32 7
1959 // CHECK: [[TMP4:%.*]] = load i64, i64* [[TMP3]]
1960 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x i64> [[TMP2]], i64 [[TMP4]], i32 1
1979 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
1980 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x i16> [[TMP2]], i16 [[TMP4]], i32 3
1991 // CHECK: [[TMP4:%.*]] = load i32, i32* [[TMP3]]
1992 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x i32> [[TMP2]], i32 [[TMP4]], i32 1
2003 // CHECK: [[TMP4:%.*]] = load i64, i64* [[TMP3]]
2004 // CHECK: [[VLD1_LANE:%.*]] = insertelement <1 x i64> [[TMP2]], i64 [[TMP4]], i32 0
2023 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
2024 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x i16> [[TMP2]], i16 [[TMP4]], i32 3
2035 // CHECK: [[TMP4:%.*]] = load i32, i32* [[TMP3]]
2036 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x i32> [[TMP2]], i32 [[TMP4]], i32 1
2047 // CHECK: [[TMP4:%.*]] = load i64, i64* [[TMP3]]
2048 // CHECK: [[VLD1_LANE:%.*]] = insertelement <1 x i64> [[TMP2]], i64 [[TMP4]], i32 0
2059 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
2060 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x i16> [[TMP2]], i16 [[TMP4]], i32 3
2072 // CHECK: [[TMP4:%.*]] = load float, float* [[TMP3]]
2073 // CHECK: [[VLD1_LANE:%.*]] = insertelement <2 x float> [[TMP2]], float [[TMP4]], i32 1
2084 // CHECK: [[TMP4:%.*]] = load double, double* [[TMP3]]
2085 // CHECK: [[VLD1_LANE:%.*]] = insertelement <1 x double> [[TMP2]], double [[TMP4]], i32 0
2104 // CHECK: [[TMP4:%.*]] = load i16, i16* [[TMP3]]
2105 // CHECK: [[VLD1_LANE:%.*]] = insertelement <4 x i16> [[TMP2]], i16 [[TMP4]], i32 3
2116 // CHECK: [[TMP4:%.*]] = load i64, i64* [[TMP3]]
2117 // CHECK: [[VLD1_LANE:%.*]] = insertelement <1 x i64> [[TMP2]], i64 [[TMP4]], i32 0
2139 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
2140 // CHECK: [[VLD2_LANE:%.*]] = call { <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld2lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], i64 15, i8* %ptr)
2168 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
2169 // CHECK: [[VLD2_LANE:%.*]] = call { <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld2lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4
2197 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
2198 // CHECK: [[VLD2_LANE:%.*]] = call { <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld2lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], i64 15, i8* %ptr)
2226 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
2230 // CHECK: [[VLD3_LANE:%.*]] = call { <16 x i8>, <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld3lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i64 15, i8* %ptr)
2258 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
2262 // CHECK: [[VLD3_LANE:%.*]] = call { <16 x i8>, <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld3lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i64 15, i8* %ptr)
2288 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
2289 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
2322 // CHECK: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[ARRAYIDX]], align 16
2323 // CHECK: [[TMP5:%.*]] = bitcast <4 x i32> [[TMP4]] to <16 x i8>
2356 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
2357 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
2390 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
2391 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
2424 // CHECK: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[ARRAYIDX]], align 16
2425 // CHECK: [[TMP5:%.*]] = bitcast <4 x i32> [[TMP4]] to <16 x i8>
2458 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
2459 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
2492 // CHECK: [[TMP4:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX]], align 16
2493 // CHECK: [[TMP5:%.*]] = bitcast <8 x half> [[TMP4]] to <16 x i8>
2526 // CHECK: [[TMP4:%.*]] = load <4 x float>, <4 x float>* [[ARRAYIDX]], align 16
2527 // CHECK: [[TMP5:%.*]] = bitcast <4 x float> [[TMP4]] to <16 x i8>
2560 // CHECK: [[TMP4:%.*]] = load <2 x double>, <2 x double>* [[ARRAYIDX]], align 16
2561 // CHECK: [[TMP5:%.*]] = bitcast <2 x double> [[TMP4]] to <16 x i8>
2594 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
2595 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
2628 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
2629 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
2664 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
2665 // CHECK: [[VLD2_LANE:%.*]] = call { <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld2lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], i64 7, i8* %a)
2691 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
2692 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
2725 // CHECK: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[ARRAYIDX]], align 8
2726 // CHECK: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <8 x i8>
2759 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
2760 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
2795 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
2796 // CHECK: [[VLD2_LANE:%.*]] = call { <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld2lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], i64 7, i8* %a)
2822 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
2823 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
2856 // CHECK: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[ARRAYIDX]], align 8
2857 // CHECK: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <8 x i8>
2890 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
2891 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
2924 // CHECK: [[TMP4:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX]], align 8
2925 // CHECK: [[TMP5:%.*]] = bitcast <4 x half> [[TMP4]] to <8 x i8>
2958 // CHECK: [[TMP4:%.*]] = load <2 x float>, <2 x float>* [[ARRAYIDX]], align 8
2959 // CHECK: [[TMP5:%.*]] = bitcast <2 x float> [[TMP4]] to <8 x i8>
2992 // CHECK: [[TMP4:%.*]] = load <1 x double>, <1 x double>* [[ARRAYIDX]], align 8
2993 // CHECK: [[TMP5:%.*]] = bitcast <1 x double> [[TMP4]] to <8 x i8>
3028 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
3029 // CHECK: [[VLD2_LANE:%.*]] = call { <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld2lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], i64 7, i8* %a)
3055 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
3056 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
3089 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
3090 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
3123 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
3124 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
3162 // CHECK: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[ARRAYIDX]], align 16
3163 // CHECK: [[TMP5:%.*]] = bitcast <4 x i32> [[TMP4]] to <16 x i8>
3201 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
3202 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
3240 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
3241 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
3279 // CHECK: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[ARRAYIDX]], align 16
3280 // CHECK: [[TMP5:%.*]] = bitcast <4 x i32> [[TMP4]] to <16 x i8>
3318 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
3319 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
3357 // CHECK: [[TMP4:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX]], align 16
3358 // CHECK: [[TMP5:%.*]] = bitcast <8 x half> [[TMP4]] to <16 x i8>
3396 // CHECK: [[TMP4:%.*]] = load <4 x float>, <4 x float>* [[ARRAYIDX]], align 16
3397 // CHECK: [[TMP5:%.*]] = bitcast <4 x float> [[TMP4]] to <16 x i8>
3435 // CHECK: [[TMP4:%.*]] = load <2 x double>, <2 x double>* [[ARRAYIDX]], align 16
3436 // CHECK: [[TMP5:%.*]] = bitcast <2 x double> [[TMP4]] to <16 x i8>
3476 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
3480 // CHECK: [[VLD3_LANE:%.*]] = call { <16 x i8>, <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld3lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i64 15, i8* %a)
3506 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
3507 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
3545 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
3546 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
3586 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
3590 // CHECK: [[VLD3_LANE:%.*]] = call { <8 x i8>, <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld3lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], i64 7, i8* %a)
3616 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
3617 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
3655 // CHECK: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[ARRAYIDX]], align 8
3656 // CHECK: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <8 x i8>
3694 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
3695 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
3735 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
3739 // CHECK: [[VLD3_LANE:%.*]] = call { <8 x i8>, <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld3lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], i64 7, i8* %a)
3765 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
3766 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
3804 // CHECK: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[ARRAYIDX]], align 8
3805 // CHECK: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <8 x i8>
3843 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
3844 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
3882 // CHECK: [[TMP4:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX]], align 8
3883 // CHECK: [[TMP5:%.*]] = bitcast <4 x half> [[TMP4]] to <8 x i8>
3921 // CHECK: [[TMP4:%.*]] = load <2 x float>, <2 x float>* [[ARRAYIDX]], align 8
3922 // CHECK: [[TMP5:%.*]] = bitcast <2 x float> [[TMP4]] to <8 x i8>
3960 // CHECK: [[TMP4:%.*]] = load <1 x double>, <1 x double>* [[ARRAYIDX]], align 8
3961 // CHECK: [[TMP5:%.*]] = bitcast <1 x double> [[TMP4]] to <8 x i8>
4001 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
4005 TMP4]], <8 x i8> [[TMP5]], i64 7, i8* %a)
4031 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
4032 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
4070 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
4071 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
4111 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
4118 // CHECK: [[VLD4_LANE:%.*]] = call { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld4lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], <16 x i8> [[TMP6]], i64 15, i8* %a)
4144 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
4145 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
4188 // CHECK: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[ARRAYIDX]], align 16
4189 // CHECK: [[TMP5:%.*]] = bitcast <4 x i32> [[TMP4]] to <16 x i8>
4232 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
4233 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
4278 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
4285 // CHECK: [[VLD4_LANE:%.*]] = call { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld4lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], <16 x i8> [[TMP6]], i64 15, i8* %a)
4311 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
4312 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
4355 // CHECK: [[TMP4:%.*]] = load <4 x i32>, <4 x i32>* [[ARRAYIDX]], align 16
4356 // CHECK: [[TMP5:%.*]] = bitcast <4 x i32> [[TMP4]] to <16 x i8>
4399 // CHECK: [[TMP4
4400 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
4443 // CHECK: [[TMP4:%.*]] = load <8 x half>, <8 x half>* [[ARRAYIDX]], align 16
4444 // CHECK: [[TMP5:%.*]] = bitcast <8 x half> [[TMP4]] to <16 x i8>
4487 // CHECK: [[TMP4:%.*]] = load <4 x float>, <4 x float>* [[ARRAYIDX]], align 16
4488 // CHECK: [[TMP5:%.*]] = bitcast <4 x float> [[TMP4]] to <16 x i8>
4531 // CHECK: [[TMP4:%.*]] = load <2 x double>, <2 x double>* [[ARRAYIDX]], align 16
4532 // CHECK: [[TMP5:%.*]] = bitcast <2 x double> [[TMP4]] to <16 x i8>
4577 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX2]], align 16
4584 // CHECK: [[VLD4_LANE:%.*]] = call { <16 x i8>, <16 x i8>, <16 x i8>, <16 x i8> } @llvm.aarch64.neon.ld4lane.v16i8.p0i8(<16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], <16 x i8> [[TMP6]], i64 15, i8* %a)
4610 // CHECK: [[TMP4:%.*]] = load <8 x i16>, <8 x i16>* [[ARRAYIDX]], align 16
4611 // CHECK: [[TMP5:%.*]] = bitcast <8 x i16> [[TMP4]] to <16 x i8>
4654 // CHECK: [[TMP4:%.*]] = load <2 x i64>, <2 x i64>* [[ARRAYIDX]], align 16
4655 // CHECK: [[TMP5:%.*]] = bitcast <2 x i64> [[TMP4]] to <16 x i8>
4700 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
4707 // CHECK: [[VLD4_LANE:%.*]] = call { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld4lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], <8 x i8> [[TMP6]], i64 7, i8* %a)
4733 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
4734 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
4777 // CHECK: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[ARRAYIDX]], align 8
4778 // CHECK: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <8 x i8>
4821 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
4822 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
4867 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
4874 // CHECK: [[VLD4_LANE:%.*]] = call { <8 x i8>, <8 x i8>, <8 x i8>, <8 x i8> } @llvm.aarch64.neon.ld4lane.v8i8.p0i8(<8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], <8 x i8> [[TMP6]], i64 7, i8* %a)
4900 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
4901 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
4944 // CHECK: [[TMP4:%.*]] = load <2 x i32>, <2 x i32>* [[ARRAYIDX]], align 8
4945 // CHECK: [[TMP5:%.*]] = bitcast <2 x i32> [[TMP4]] to <8 x i8>
4988 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
4989 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
5032 // CHECK: [[TMP4:%.*]] = load <4 x half>, <4 x half>* [[ARRAYIDX]], align 8
5033 // CHECK: [[TMP5:%.*]] = bitcast <4 x half> [[TMP4]] to <8 x i8>
5076 // CHECK: [[TMP4:%.*]] = load <2 x float>, <2 x float>* [[ARRAYIDX]], align 8
5077 // CHECK: [[TMP5:%.*]] = bitcast <2 x float> [[TMP4]] to <8 x i8>
5120 // CHECK: [[TMP4:%.*]] = load <1 x double>, <1 x double>* [[ARRAYIDX]], align 8
5121 // CHECK: [[TMP5:%.*]] = bitcast <1 x double> [[TMP4]] to <8 x i8>
5166 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX2]], align 8
5173 TMP4]], <8 x i8> [[TMP5]], <8 x i8> [[TMP6]], i64 7, i8* %a)
5199 // CHECK: [[TMP4:%.*]] = load <4 x i16>, <4 x i16>* [[ARRAYIDX]], align 8
5200 // CHECK: [[TMP5:%.*]] = bitcast <4 x i16> [[TMP4]] to <8 x i8>
5243 // CHECK: [[TMP4:%.*]] = load <1 x i64>, <1 x i64>* [[ARRAYIDX]], align 8
5244 // CHECK: [[TMP5:%.*]] = bitcast <1 x i64> [[TMP4]] to <8 x i8>
5286 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5287 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5298 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i32*
5299 // CHECK: store i32 [[TMP3]], i32* [[TMP4]]
5310 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i64*
5311 // CHECK: store i64 [[TMP3]], i64* [[TMP4]]
5330 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5331 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5342 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i32*
5343 // CHECK: store i32 [[TMP3]], i32* [[TMP4]]
5354 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i64*
5355 // CHECK: store i64 [[TMP3]], i64* [[TMP4]]
5366 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5367 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5378 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to float*
5379 // CHECK: store float [[TMP3]], float* [[TMP4]]
5390 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to double*
5391 // CHECK: store double [[TMP3]], double* [[TMP4]]
5410 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5411 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5422 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i64*
5423 // CHECK: store i64 [[TMP3]], i64* [[TMP4]]
5442 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5443 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5454 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i32*
5455 // CHECK: store i32 [[TMP3]], i32* [[TMP4]]
5466 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i64*
5467 // CHECK: store i64 [[TMP3]], i64* [[TMP4]]
5486 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5487 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5498 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i32*
5499 // CHECK: store i32 [[TMP3]], i32* [[TMP4]]
5510 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i64*
5511 // CHECK: store i64 [[TMP3]], i64* [[TMP4]]
5522 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5523 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5534 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to float*
5535 // CHECK: store float [[TMP3]], float* [[TMP4]]
5546 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to double*
5547 // CHECK: store double [[TMP3]], double* [[TMP4]]
5566 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i16*
5567 // CHECK: store i16 [[TMP3]], i16* [[TMP4]]
5578 // CHECK: [[TMP4:%.*]] = bitcast i8* [[TMP0]] to i64*
5579 // CHECK: store i64 [[TMP3]], i64* [[TMP4]]
5617 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
5622 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
5642 // CHECK: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
5647 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x i32>
5667 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
5672 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
5712 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
5717 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
5737 // CHECK: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
5742 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x i32>
5762 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
5767 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
5787 // CHECK: [[TMP4:%.*]] = bitcast <8 x half> [[TMP3]] to <16 x i8>
5792 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
5812 // CHECK: [[TMP4:%.*]] = bitcast <4 x float> [[TMP3]] to <16 x i8>
5817 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x float>
5837 // CHECK: [[TMP4:%.*]] = bitcast <2 x double> [[TMP3]] to <16 x i8>
5842 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x double>
5882 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
5887 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
5907 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
5912 // CHECK: [[TMP7:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
5952 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
5957 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
5977 // CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <8 x i8>
5982 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x i32>
6002 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
6007 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
6047 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
6052 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
6072 // CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <8 x i8>
6077 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x i32>
6097 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
6102 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
6122 // CHECK: [[TMP4:%.*]] = bitcast <4 x half> [[TMP3]] to <8 x i8>
6127 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
6147 // CHECK: [[TMP4:%.*]] = bitcast <2 x float> [[TMP3]] to <8 x i8>
6152 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x float>
6172 // CHECK: [[TMP4:%.*]] = bitcast <1 x double> [[TMP3]] to <8 x i8>
6177 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x double>
6217 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
6222 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
6242 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
6247 // CHECK: [[TMP7:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
6271 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX4]], align 16
6272 // CHECK: call void @llvm.aarch64.neon.st3lane.v16i8.p0i8(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], i64 15, i8* %a)
6290 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
6299 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
6320 // CHECK: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
6329 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x i32>
6350 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
6359 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
6384 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX4]], align 16
6385 // CHECK: call void @llvm.aarch64.neon.st3lane.v16i8.p0i8(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], i64 15, i8* %a)
6403 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
6412 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
6433 // CHECK: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
6442 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x i32>
6463 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
6472 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
6493 // CHECK: [[TMP4:%.*]] = bitcast <8 x half> [[TMP3]] to <16 x i8>
6502 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
6523 // CHECK: [[TMP4:%.*]] = bitcast <4 x float> [[TMP3]] to <16 x i8>
6532 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x float>
6553 // CHECK: [[TMP4:%.*]] = bitcast <2 x double> [[TMP3]] to <16 x i8>
6562 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x double>
6587 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX4]], align 16
6588 // CHECK: call void @llvm.aarch64.neon.st3lane.v16i8.p0i8(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], i64 15, i8* %a)
6606 // CHECK: [[TMP4
6615 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
6636 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
6645 // CHECK: [[TMP9:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
6670 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX4]], align 8
6671 // CHECK: call void @llvm.aarch64.neon.st3lane.v8i8.p0i8(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], i64 7, i8* %a)
6689 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
6698 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
6719 // CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <8 x i8>
6728 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x i32>
6749 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
6758 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
6783 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX4]], align 8
6784 // CHECK: call void @llvm.aarch64.neon.st3lane.v8i8.p0i8(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], i64 7, i8* %a)
6802 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
6811 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
6832 // CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <8 x i8>
6841 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x i32>
6862 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
6871 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
6892 // CHECK: [[TMP4:%.*]] = bitcast <4 x half> [[TMP3]] to <8 x i8>
6901 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
6922 // CHECK: [[TMP4:%.*]] = bitcast <2 x float> [[TMP3]] to <8 x i8>
6931 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x float>
6952 // CHECK: [[TMP4:%.*]] = bitcast <1 x double> [[TMP3]] to <8 x i8>
6961 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x double>
6986 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX4]], align 8
6987 // CHECK: call void @llvm.aarch64.neon.st3lane.v8i8.p0i8(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], i64 7, i8* %a)
7005 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
7014 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
7035 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
7044 // CHECK: [[TMP9:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
7069 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX4]], align 16
7073 // CHECK: call void @llvm.aarch64.neon.st4lane.v16i8.p0i8(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i64 15, i8* %a)
7091 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
7104 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
7126 // CHECK: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
7139 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x i32>
7161 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
7174 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
7200 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX4]], align 16
7204 // CHECK: call void @llvm.aarch64.neon.st4lane.v16i8.p0i8(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i64 15, i8* %a)
7222 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
7235 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
7257 // CHECK: [[TMP4:%.*]] = bitcast <4 x i32> [[TMP3]] to <16 x i8>
7270 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x i32>
7292 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
7305 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
7327 // CHECK: [[TMP4:%.*]] = bitcast <8 x half> [[TMP3]] to <16 x i8>
7340 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
7362 // CHECK: [[TMP4:%.*]] = bitcast <4 x float> [[TMP3]] to <16 x i8>
7375 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <4 x float>
7397 // CHECK: [[TMP4:%.*]] = bitcast <2 x double> [[TMP3]] to <16 x i8>
7410 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x double>
7436 // CHECK: [[TMP4:%.*]] = load <16 x i8>, <16 x i8>* [[ARRAYIDX4]], align 16
7440 // CHECK: call void @llvm.aarch64.neon.st4lane.v16i8.p0i8(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], i64 15, i8* %a)
7458 // CHECK: [[TMP4:%.*]] = bitcast <8 x i16> [[TMP3]] to <16 x i8>
7471 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <8 x i16>
7493 // CHECK: [[TMP4:%.*]] = bitcast <2 x i64> [[TMP3]] to <16 x i8>
7506 // CHECK: [[TMP11:%.*]] = bitcast <16 x i8> [[TMP4]] to <2 x i64>
7532 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX4]], align 8
7536 // CHECK: call void @llvm.aarch64.neon.st4lane.v8i8.p0i8(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], i64 7, i8* %a)
7554 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
7567 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
7589 // CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <8 x i8>
7602 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x i32>
7624 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
7637 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
7663 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX4]], align 8
7667 // CHECK: call void @llvm.aarch64.neon.st4lane.v8i8.p0i8(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], i64 7, i8* %a)
7685 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
7698 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
7720 // CHECK: [[TMP4:%.*]] = bitcast <2 x i32> [[TMP3]] to <8 x i8>
7733 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x i32>
7755 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
7768 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
7790 // CHECK: [[TMP4:%.*]] = bitcast <4 x half> [[TMP3]] to <8 x i8>
7803 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
7825 // CHECK: [[TMP4:%.*]] = bitcast <2 x float> [[TMP3]] to <8 x i8>
7838 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <2 x float>
7860 // CHECK: [[TMP4:%.*]] = bitcast <1 x double> [[TMP3]] to <8 x i8>
7873 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x double>
7899 // CHECK: [[TMP4:%.*]] = load <8 x i8>, <8 x i8>* [[ARRAYIDX4]], align 8
7903 // CHECK: call void @llvm.aarch64.neon.st4lane.v8i8.p0i8(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], i64 7, i8* %a)
7921 // CHECK: [[TMP4:%.*]] = bitcast <4 x i16> [[TMP3]] to <8 x i8>
7934 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <4 x i16>
7956 // CHECK: [[TMP4:%.*]] = bitcast <1 x i64> [[TMP3]] to <8 x i8>
7969 // CHECK: [[TMP11:%.*]] = bitcast <8 x i8> [[TMP4]] to <1 x i64>
