xref: /llvm-project/llvm/test/Analysis/CostModel/RISCV/shuffle-interleave.ll (revision 55304d0d907fb26c298b84447a85e3a987d0adbc) 
1; NOTE: Assertions have been autogenerated by utils/update_analyze_test_checks.py UTC_ARGS: --version 2
2; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv32 -mattr=+v | FileCheck --check-prefixes=CHECK,RV32 %s
3; RUN: opt < %s -passes="print<cost-model>" 2>&1 -disable-output -mtriple=riscv64 -mattr=+v | FileCheck --check-prefixes=CHECK,RV64 %s
4
5; The mask here interleaves (%v1, %v0), not (%v0, %v1): it should still be cheap.
6define <4 x i8> @interleave2_v2i8(<2 x i8> %v0, <2 x i8> %v1) {
7; CHECK-LABEL: 'interleave2_v2i8'
8; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %concat = shufflevector <2 x i8> %v0, <2 x i8> %v1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
9; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res = shufflevector <4 x i8> %concat, <4 x i8> poison, <4 x i32> <i32 2, i32 0, i32 3, i32 1>
10; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <4 x i8> %res
11;
12  %concat = shufflevector <2 x i8> %v0, <2 x i8> %v1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
13  %res = shufflevector <4 x i8> %concat, <4 x i8> poison, <4 x i32> <i32 2, i32 0, i32 3, i32 1>
14  ret <4 x i8> %res
15}
16
17define <8 x i8> @interleave2_v8i8(<4 x i8> %v0, <4 x i8> %v1) {
18; CHECK-LABEL: 'interleave2_v8i8'
19; CHECK-NEXT:  Cost Model: Found an estimated cost of 1 for instruction: %concat = shufflevector <4 x i8> %v0, <4 x i8> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
20; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %res = shufflevector <8 x i8> %concat, <8 x i8> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
21; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i8> %res
22;
23  %concat = shufflevector <4 x i8> %v0, <4 x i8> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
24  %res = shufflevector <8 x i8> %concat, <8 x i8> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
25  ret <8 x i8> %res
26}
27
28define <8 x i32> @interleave2_v8i32(<4 x i32> %v0, <4 x i32> %v1) {
29; CHECK-LABEL: 'interleave2_v8i32'
30; CHECK-NEXT:  Cost Model: Found an estimated cost of 2 for instruction: %concat = shufflevector <4 x i32> %v0, <4 x i32> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
31; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %res = shufflevector <8 x i32> %concat, <8 x i32> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
32; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i32> %res
33;
34  %concat = shufflevector <4 x i32> %v0, <4 x i32> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
35  %res = shufflevector <8 x i32> %concat, <8 x i32> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
36  ret <8 x i32> %res
37}
38
39; Should be expensive on RV32 because it can't widen
40define <8 x i64> @interleave2_v8i64(<4 x i64> %v0, <4 x i64> %v1) {
41; RV32-LABEL: 'interleave2_v8i64'
42; RV32-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %concat = shufflevector <4 x i64> %v0, <4 x i64> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
43; RV32-NEXT:  Cost Model: Found an estimated cost of 19 for instruction: %res = shufflevector <8 x i64> %concat, <8 x i64> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
44; RV32-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %res
45;
46; RV64-LABEL: 'interleave2_v8i64'
47; RV64-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %concat = shufflevector <4 x i64> %v0, <4 x i64> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
48; RV64-NEXT:  Cost Model: Found an estimated cost of 22 for instruction: %res = shufflevector <8 x i64> %concat, <8 x i64> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
49; RV64-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret <8 x i64> %res
50;
51  %concat = shufflevector <4 x i64> %v0, <4 x i64> %v1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
52  %res = shufflevector <8 x i64> %concat, <8 x i64> poison, <8 x i32> <i32 0, i32 4, i32 1, i32 5, i32 2, i32 6, i32 3, i32 7>
53  ret <8 x i64> %res
54}
55
56; TODO: getInstructionCost doesn't call getShuffleCost here because the shuffle changes length
57define {<4 x i8>, <4 x i8>} @deinterleave_2(<8 x i8> %v) {
58; CHECK-LABEL: 'deinterleave_2'
59; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %v0 = shufflevector <8 x i8> %v, <8 x i8> poison, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
60; CHECK-NEXT:  Cost Model: Found an estimated cost of 4 for instruction: %v1 = shufflevector <8 x i8> %v, <8 x i8> poison, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
61; CHECK-NEXT:  Cost Model: Found an estimated cost of -1 for instruction: %res0 = insertvalue { <4 x i8>, <4 x i8> } poison, <4 x i8> %v0, 0
62; CHECK-NEXT:  Cost Model: Found an estimated cost of -1 for instruction: %res1 = insertvalue { <4 x i8>, <4 x i8> } %res0, <4 x i8> %v1, 1
63; CHECK-NEXT:  Cost Model: Found an estimated cost of 0 for instruction: ret { <4 x i8>, <4 x i8> } %res1
64;
65  %v0 = shufflevector <8 x i8> %v, <8 x i8> poison, <4 x i32> <i32 0, i32 2, i32 4, i32 6>
66  %v1 = shufflevector <8 x i8> %v, <8 x i8> poison, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
67  %res0 = insertvalue {<4 x i8>, <4 x i8>} poison, <4 x i8> %v0, 0
68  %res1 = insertvalue {<4 x i8>, <4 x i8>} %res0, <4 x i8> %v1, 1
69  ret {<4 x i8>, <4 x i8>} %res1
70}
71