[mlir] IntegerRangeAnalysis: add support for vector type #112292
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,106 @@ | ||
| // RUN: mlir-opt -int-range-optimizations -canonicalize %s | FileCheck %s | ||
|
|
||
|
|
||
Hardcode84 marked this conversation as resolved.
Show resolved
Hide resolved
|
||
| // CHECK-LABEL: func @constant_vec | ||
| // CHECK: test.reflect_bounds {smax = 7 : index, smin = 0 : index, umax = 7 : index, umin = 0 : index} | ||
| func.func @constant_vec() -> vector<8xindex> { | ||
| %0 = arith.constant dense<[0, 1, 2, 3, 4, 5, 6, 7]> : vector<8xindex> | ||
| %1 = test.reflect_bounds %0 : vector<8xindex> | ||
|
There was a problem hiding this comment. I don't really know what these test Ops do and I couldn't find any documentation in code. Could add some docs, pls? There was a problem hiding this comment. These test ops are from existing integer range inference tests - they have an implement of the integer range inference interface that sets attributes on There was a problem hiding this comment. Yes, I gathered that much from mlir/test/lib/Dialect/Test/TestOps.td, but it doesn’t quite clarify things for me. The lack of documentation for these operations makes it hard to understand the distinction between Given that @Hardcode84 is already using these ops for testing, it would be fantastic if some of that expertise could be shared through documentation. This would benefit everyone working with these tests! There was a problem hiding this comment. added some description to the ops |
||
| func.return %1 : vector<8xindex> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @constant_splat | ||
| // CHECK: test.reflect_bounds {smax = 3 : si32, smin = 3 : si32, umax = 3 : ui32, umin = 3 : ui32} | ||
| func.func @constant_splat() -> vector<8xi32> { | ||
| %0 = arith.constant dense<3> : vector<8xi32> | ||
| %1 = test.reflect_bounds %0 : vector<8xi32> | ||
| func.return %1 : vector<8xi32> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_splat | ||
| // CHECK: test.reflect_bounds {smax = 5 : index, smin = 4 : index, umax = 5 : index, umin = 4 : index} | ||
| func.func @vector_splat() -> vector<4xindex> { | ||
| %0 = test.with_bounds { umin = 4 : index, umax = 5 : index, smin = 4 : index, smax = 5 : index } : index | ||
| %1 = vector.splat %0 : vector<4xindex> | ||
| %2 = test.reflect_bounds %1 : vector<4xindex> | ||
| func.return %2 : vector<4xindex> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_broadcast | ||
| // CHECK: test.reflect_bounds {smax = 5 : index, smin = 4 : index, umax = 5 : index, umin = 4 : index} | ||
| func.func @vector_broadcast() -> vector<4x16xindex> { | ||
| %0 = test.with_bounds { umin = 4 : index, umax = 5 : index, smin = 4 : index, smax = 5 : index } : vector<16xindex> | ||
| %1 = vector.broadcast %0 : vector<16xindex> to vector<4x16xindex> | ||
| %2 = test.reflect_bounds %1 : vector<4x16xindex> | ||
| func.return %2 : vector<4x16xindex> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_shape_cast | ||
| // CHECK: test.reflect_bounds {smax = 5 : index, smin = 4 : index, umax = 5 : index, umin = 4 : index} | ||
| func.func @vector_shape_cast() -> vector<4x4xindex> { | ||
| %0 = test.with_bounds { umin = 4 : index, umax = 5 : index, smin = 4 : index, smax = 5 : index } : vector<16xindex> | ||
| %1 = vector.shape_cast %0 : vector<16xindex> to vector<4x4xindex> | ||
| %2 = test.reflect_bounds %1 : vector<4x4xindex> | ||
| func.return %2 : vector<4x4xindex> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_extract | ||
| // CHECK: test.reflect_bounds {smax = 6 : index, smin = 5 : index, umax = 6 : index, umin = 5 : index} | ||
| func.func @vector_extract() -> index { | ||
| %0 = test.with_bounds { umin = 5 : index, umax = 6 : index, smin = 5 : index, smax = 6 : index } : vector<4xindex> | ||
| %1 = vector.extract %0[0] : index from vector<4xindex> | ||
| %2 = test.reflect_bounds %1 : index | ||
| func.return %2 : index | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_extractelement | ||
| // CHECK: test.reflect_bounds {smax = 7 : index, smin = 6 : index, umax = 7 : index, umin = 6 : index} | ||
| func.func @vector_extractelement() -> index { | ||
| %c0 = arith.constant 0 : index | ||
| %0 = test.with_bounds { umin = 6 : index, umax = 7 : index, smin = 6 : index, smax = 7 : index } : vector<4xindex> | ||
| %1 = vector.extractelement %0[%c0 : index] : vector<4xindex> | ||
| %2 = test.reflect_bounds %1 : index | ||
| func.return %2 : index | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_add | ||
| // CHECK: test.reflect_bounds {smax = 12 : index, smin = 10 : index, umax = 12 : index, umin = 10 : index} | ||
| func.func @vector_add() -> vector<4xindex> { | ||
| %0 = test.with_bounds { umin = 4 : index, umax = 5 : index, smin = 4 : index, smax = 5 : index } : vector<4xindex> | ||
| %1 = test.with_bounds { umin = 6 : index, umax = 7 : index, smin = 6 : index, smax = 7 : index } : vector<4xindex> | ||
| %2 = arith.addi %0, %1 : vector<4xindex> | ||
| %3 = test.reflect_bounds %2 : vector<4xindex> | ||
| func.return %3 : vector<4xindex> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_insert | ||
| // CHECK: test.reflect_bounds {smax = 8 : index, smin = 5 : index, umax = 8 : index, umin = 5 : index} | ||
| func.func @vector_insert() -> vector<4xindex> { | ||
| %0 = test.with_bounds { umin = 5 : index, umax = 7 : index, smin = 5 : index, smax = 7 : index } : vector<4xindex> | ||
| %1 = test.with_bounds { umin = 6 : index, umax = 8 : index, smin = 6 : index, smax = 8 : index } : index | ||
| %2 = vector.insert %1, %0[0] : index into vector<4xindex> | ||
| %3 = test.reflect_bounds %2 : vector<4xindex> | ||
| func.return %3 : vector<4xindex> | ||
| } | ||
|
|
||
| // CHECK-LABEL: func @vector_insertelement | ||
| // CHECK: test.reflect_bounds {smax = 8 : index, smin = 5 : index, umax = 8 : index, umin = 5 : index} | ||
| func.func @vector_insertelement() -> vector<4xindex> { | ||
| %c0 = arith.constant 0 : index | ||
| %0 = test.with_bounds { umin = 5 : index, umax = 7 : index, smin = 5 : index, smax = 7 : index } : vector<4xindex> | ||
| %1 = test.with_bounds { umin = 6 : index, umax = 8 : index, smin = 6 : index, smax = 8 : index } : index | ||
| %2 = vector.insertelement %1, %0[%c0 : index] : vector<4xindex> | ||
| %3 = test.reflect_bounds %2 : vector<4xindex> | ||
| func.return %3 : vector<4xindex> | ||
| } | ||
|
There was a problem hiding this comment. So, just because I can't remember exactly func.func @test_loaded_vector_extract(%memref : memref<16xi32>) -> i32 {
%c0 = arith.constant 0 : index
%v = vector.load %memref[%c0] : vector<4xi32>
%e = vector.extract %v[0]
%bounds = test.reflect_bounds %e : i32
func.return %bounds : i32
} |
||
|
|
||
| // CHECK-LABEL: func @test_loaded_vector_extract | ||
| // No bounds | ||
| // CHECK: test.reflect_bounds %{{.*}} : i32 | ||
| func.func @test_loaded_vector_extract(%memref : memref<16xi32>) -> i32 { | ||
| %c0 = arith.constant 0 : index | ||
| %v = vector.load %memref[%c0] : memref<16xi32>, vector<4xi32> | ||
| %e = vector.extract %v[0] : i32 from vector<4xi32> | ||
| %bounds = test.reflect_bounds %e : i32 | ||
| func.return %bounds : i32 | ||
| } | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
These look like unrelated changes?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I've added
mlir/Interfaces/InferIntRangeInterface.tdand sorted rests of the includes.