eb6c4197d5
Do not run `cf-to-llvm` as part of `func-to-llvm`. This commit fixes https://github.com/llvm/llvm-project/issues/70982. This commit changes the way how `func.func` ops are lowered to LLVM. Previously, the signature of the entire region (i.e., entry block and all other blocks in the `func.func` op) was converted as part of the `func.func` lowering pattern. Now, only the entry block is converted. The remaining block signatures are converted together with `cf.br` and `cf.cond_br` as part of `cf-to-llvm`. All unstructured control flow is not converted as part of a single pass (`cf-to-llvm`). `func-to-llvm` no longer deals with unstructured control flow. Also add more test cases for control flow dialect ops. Note: This PR is in preparation of #120431, which adds an additional GPU-specific lowering for `cf.assert`. This was a problem because `cf.assert` used to be converted as part of `func-to-llvm`. Note for LLVM integration: If you see failures, add `-convert-cf-to-llvm` to your pass pipeline.
55 lines
2.0 KiB
MLIR
55 lines
2.0 KiB
MLIR
// RUN: mlir-opt %s -test-vector-scan-lowering -test-lower-to-llvm | \
|
|
// RUN: mlir-cpu-runner -e entry -entry-point-result=void \
|
|
// RUN: -shared-libs=%mlir_c_runner_utils | \
|
|
// RUN: FileCheck %s
|
|
|
|
func.func @entry() {
|
|
%f1 = arith.constant 1.0: f32
|
|
%f2 = arith.constant 2.0: f32
|
|
%f3 = arith.constant 3.0: f32
|
|
%f4 = arith.constant 4.0: f32
|
|
%f5 = arith.constant 5.0: f32
|
|
%f6 = arith.constant 6.0: f32
|
|
|
|
// Construct test vector.
|
|
%0 = vector.broadcast %f1 : f32 to vector<3x2xf32>
|
|
%1 = vector.insert %f2, %0[0, 1] : f32 into vector<3x2xf32>
|
|
%2 = vector.insert %f3, %1[1, 0] : f32 into vector<3x2xf32>
|
|
%3 = vector.insert %f4, %2[1, 1] : f32 into vector<3x2xf32>
|
|
%4 = vector.insert %f5, %3[2, 0] : f32 into vector<3x2xf32>
|
|
%x = vector.insert %f6, %4[2, 1] : f32 into vector<3x2xf32>
|
|
vector.print %x : vector<3x2xf32>
|
|
// CHECK: ( ( 1, 2 ), ( 3, 4 ), ( 5, 6 ) )
|
|
|
|
%y = vector.broadcast %f6 : f32 to vector<2xf32>
|
|
%z = vector.broadcast %f6 : f32 to vector<3xf32>
|
|
// Scan
|
|
%a:2 = vector.scan <add>, %x, %y {inclusive = true, reduction_dim = 0} :
|
|
vector<3x2xf32>, vector<2xf32>
|
|
%b:2 = vector.scan <add>, %x, %z {inclusive = true, reduction_dim = 1} :
|
|
vector<3x2xf32>, vector<3xf32>
|
|
%c:2 = vector.scan <add>, %x, %y {inclusive = false, reduction_dim = 0} :
|
|
vector<3x2xf32>, vector<2xf32>
|
|
%d:2 = vector.scan <add>, %x, %z {inclusive = false, reduction_dim = 1} :
|
|
vector<3x2xf32>, vector<3xf32>
|
|
|
|
// CHECK: ( ( 1, 2 ), ( 4, 6 ), ( 9, 12 ) )
|
|
// CHECK: ( 9, 12 )
|
|
// CHECK: ( ( 1, 3 ), ( 3, 7 ), ( 5, 11 ) )
|
|
// CHECK: ( 3, 7, 11 )
|
|
// CHECK: ( ( 6, 6 ), ( 7, 8 ), ( 10, 12 ) )
|
|
// CHECK: ( 10, 12 )
|
|
// CHECK: ( ( 6, 7 ), ( 6, 9 ), ( 6, 11 ) )
|
|
// CHECK: ( 7, 9, 11 )
|
|
vector.print %a#0 : vector<3x2xf32>
|
|
vector.print %a#1 : vector<2xf32>
|
|
vector.print %b#0 : vector<3x2xf32>
|
|
vector.print %b#1 : vector<3xf32>
|
|
vector.print %c#0 : vector<3x2xf32>
|
|
vector.print %c#1 : vector<2xf32>
|
|
vector.print %d#0 : vector<3x2xf32>
|
|
vector.print %d#1 : vector<3xf32>
|
|
|
|
return
|
|
}
|