|
| 1 | +; NOTE: Assertions have been autogenerated by utils/update_test_checks.py |
| 2 | +; RUN: opt < %s -vector-combine -S -mtriple=x86_64-- -mattr=SSE2 | FileCheck %s --check-prefixes=CHECK,SSE |
| 3 | +; RUN: opt < %s -vector-combine -S -mtriple=x86_64-- -mattr=AVX2 | FileCheck %s --check-prefixes=CHECK,AVX |
| 4 | + |
| 5 | +define <16 x i8> @bitcast_shuf_narrow_element(<4 x i32> %v) { |
| 6 | +; CHECK-LABEL: @bitcast_shuf_narrow_element( |
| 7 | +; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <4 x i32> [[V:%.*]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 8 | +; CHECK-NEXT: [[R:%.*]] = bitcast <4 x i32> [[SHUF]] to <16 x i8> |
| 9 | +; CHECK-NEXT: ret <16 x i8> [[R]] |
| 10 | +; |
| 11 | + %shuf = shufflevector <4 x i32> %v, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 12 | + %r = bitcast <4 x i32> %shuf to <16 x i8> |
| 13 | + ret <16 x i8> %r |
| 14 | +} |
| 15 | + |
| 16 | +define <4 x float> @bitcast_shuf_same_size(<4 x i32> %v) { |
| 17 | +; CHECK-LABEL: @bitcast_shuf_same_size( |
| 18 | +; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <4 x i32> [[V:%.*]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 19 | +; CHECK-NEXT: [[R:%.*]] = bitcast <4 x i32> [[SHUF]] to <4 x float> |
| 20 | +; CHECK-NEXT: ret <4 x float> [[R]] |
| 21 | +; |
| 22 | + %shuf = shufflevector <4 x i32> %v, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 23 | + %r = bitcast <4 x i32> %shuf to <4 x float> |
| 24 | + ret <4 x float> %r |
| 25 | +} |
| 26 | + |
| 27 | +define <4 x i32> @bitcast_shuf_wide_element(<8 x i16> %v) { |
| 28 | +; CHECK-LABEL: @bitcast_shuf_wide_element( |
| 29 | +; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <8 x i16> [[V:%.*]], <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 2, i32 3, i32 2, i32 3> |
| 30 | +; CHECK-NEXT: [[R:%.*]] = bitcast <8 x i16> [[SHUF]] to <4 x i32> |
| 31 | +; CHECK-NEXT: ret <4 x i32> [[R]] |
| 32 | +; |
| 33 | + %shuf = shufflevector <8 x i16> %v, <8 x i16> undef, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 2, i32 3, i32 2, i32 3> |
| 34 | + %r = bitcast <8 x i16> %shuf to <4 x i32> |
| 35 | + ret <4 x i32> %r |
| 36 | +} |
| 37 | + |
| 38 | +declare void @use(<4 x i32>) |
| 39 | + |
| 40 | +define <16 x i8> @bitcast_shuf_uses(<4 x i32> %v) { |
| 41 | +; CHECK-LABEL: @bitcast_shuf_uses( |
| 42 | +; CHECK-NEXT: [[SHUF:%.*]] = shufflevector <4 x i32> [[V:%.*]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 43 | +; CHECK-NEXT: call void @use(<4 x i32> [[SHUF]]) |
| 44 | +; CHECK-NEXT: [[R:%.*]] = bitcast <4 x i32> [[SHUF]] to <16 x i8> |
| 45 | +; CHECK-NEXT: ret <16 x i8> [[R]] |
| 46 | +; |
| 47 | + %shuf = shufflevector <4 x i32> %v, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 48 | + call void @use(<4 x i32> %shuf) |
| 49 | + %r = bitcast <4 x i32> %shuf to <16 x i8> |
| 50 | + ret <16 x i8> %r |
| 51 | +} |
| 52 | + |
| 53 | +define <2 x i64> @PR35454_1(<2 x i64> %v) { |
| 54 | +; CHECK-LABEL: @PR35454_1( |
| 55 | +; CHECK-NEXT: [[BC:%.*]] = bitcast <2 x i64> [[V:%.*]] to <4 x i32> |
| 56 | +; CHECK-NEXT: [[PERMIL:%.*]] = shufflevector <4 x i32> [[BC]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 57 | +; CHECK-NEXT: [[BC1:%.*]] = bitcast <4 x i32> [[PERMIL]] to <16 x i8> |
| 58 | +; CHECK-NEXT: [[ADD:%.*]] = shl <16 x i8> [[BC1]], <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> |
| 59 | +; CHECK-NEXT: [[BC2:%.*]] = bitcast <16 x i8> [[ADD]] to <4 x i32> |
| 60 | +; CHECK-NEXT: [[PERMIL1:%.*]] = shufflevector <4 x i32> [[BC2]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 61 | +; CHECK-NEXT: [[BC3:%.*]] = bitcast <4 x i32> [[PERMIL1]] to <2 x i64> |
| 62 | +; CHECK-NEXT: ret <2 x i64> [[BC3]] |
| 63 | +; |
| 64 | + %bc = bitcast <2 x i64> %v to <4 x i32> |
| 65 | + %permil = shufflevector <4 x i32> %bc, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 66 | + %bc1 = bitcast <4 x i32> %permil to <16 x i8> |
| 67 | + %add = shl <16 x i8> %bc1, <i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1, i8 1> |
| 68 | + %bc2 = bitcast <16 x i8> %add to <4 x i32> |
| 69 | + %permil1 = shufflevector <4 x i32> %bc2, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 70 | + %bc3 = bitcast <4 x i32> %permil1 to <2 x i64> |
| 71 | + ret <2 x i64> %bc3 |
| 72 | +} |
| 73 | + |
| 74 | +define <2 x i64> @PR35454_2(<2 x i64> %v) { |
| 75 | +; CHECK-LABEL: @PR35454_2( |
| 76 | +; CHECK-NEXT: [[BC:%.*]] = bitcast <2 x i64> [[V:%.*]] to <4 x i32> |
| 77 | +; CHECK-NEXT: [[PERMIL:%.*]] = shufflevector <4 x i32> [[BC]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 78 | +; CHECK-NEXT: [[BC1:%.*]] = bitcast <4 x i32> [[PERMIL]] to <8 x i16> |
| 79 | +; CHECK-NEXT: [[ADD:%.*]] = shl <8 x i16> [[BC1]], <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> |
| 80 | +; CHECK-NEXT: [[BC2:%.*]] = bitcast <8 x i16> [[ADD]] to <4 x i32> |
| 81 | +; CHECK-NEXT: [[PERMIL1:%.*]] = shufflevector <4 x i32> [[BC2]], <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 82 | +; CHECK-NEXT: [[BC3:%.*]] = bitcast <4 x i32> [[PERMIL1]] to <2 x i64> |
| 83 | +; CHECK-NEXT: ret <2 x i64> [[BC3]] |
| 84 | +; |
| 85 | + %bc = bitcast <2 x i64> %v to <4 x i32> |
| 86 | + %permil = shufflevector <4 x i32> %bc, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 87 | + %bc1 = bitcast <4 x i32> %permil to <8 x i16> |
| 88 | + %add = shl <8 x i16> %bc1, <i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1, i16 1> |
| 89 | + %bc2 = bitcast <8 x i16> %add to <4 x i32> |
| 90 | + %permil1 = shufflevector <4 x i32> %bc2, <4 x i32> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0> |
| 91 | + %bc3 = bitcast <4 x i32> %permil1 to <2 x i64> |
| 92 | + ret <2 x i64> %bc3 |
| 93 | +} |
0 commit comments