21 Commits

Author SHA1 Message Date
Matthias Springer
c1029b6a9b
[mlir][Transform] apply_conversion_patterns: Update handles (#83950)
Until now, `transform.apply_conversion_patterns` consumed the target
handle and potentially invalidated handles. This commit adds tracking
functionality similar to `transform.apply_patterns`, such that handles
are no longer invalidated, but updated based on op replacements
performed by the dialect conversion.

This new functionality is hidden behind a `preserve_handles` attribute
for now.
2024-03-10 12:10:53 +09:00
Oleksandr "Alex" Zinenko
5468f88413
[mlir] update remaining transform tests to main pass (#81279)
Use the main transform interpreter pass instead of the test pass. The
only tests that are not updated are specific to the operation of the
test pass.
2024-02-28 11:06:53 +01:00
Oleksandr "Alex" Zinenko
2798b72ae7
[mlir] introduce debug transform dialect extension (#77595)
Introduce a new extension for simple print-debugging of the transform
dialect scripts. The initial version of this extension consists of two
ops that are printing the payload objects associated with transform
dialect values. Similar ops were already available in the test extenion
and several downstream projects, and were extensively used for testing.
2024-01-12 13:24:02 +01:00
Billy Zhu
34a65980d7
[MLIR] Erase location of folded constants (#75415)
Follow up to the discussion from #75258, and serves as an alternate
solution for #74670.

Set the location to Unknown for deduplicated / moved / materialized
constants by OperationFolder. This makes sure that the folded constants
don't end up with an arbitrary location of one of the original ops that
became it, and that hoisted ops don't confuse the stepping order.
2023-12-21 09:54:48 -08:00
Matthias Springer
e8ae0e72b7
[mlir][transform] TrackingListener: Improve dead handles detection (#74290)
The tracking listener should not report op replacement errors for
payload ops that are not mapped to any live handles. The handle liveless
analysis did not work properly with transform IR that has named
sequences.

A handle is live if it has a user after the transform op that is
currently being applied. With named sequences, we need to maintain a
stack of currently applied transform ops. That stack already exists
(`regionStack`), the only thing that's missing is the current transform
op for each stack frame.

This commit fixes #72931.
2023-12-06 16:32:22 +09:00
Ingo Müller
f40e620956 Reapply "[mlir][transform] Improve error message of tracking listener. (#66987)"
This commit reapplies #66987, which got original contained a memory leak
and got reverted by 78c8ab5844e618162c4cf3982d05102d4da10d23. The leak
is now fixed.

Original description:

This PR extends the error message of the tracking listener when
replacement ops cannot be found. That may happen if the applied patterns
replace an op by an op of a different kind or by block arguments.
However, this only matters if there are alive handles to the replaced
op. The new error message mentions that explicitly and reports the alive
handles.
2023-09-26 12:56:38 +00:00
Vitaly Buka
78c8ab5844 Revert "[mlir][transform] Improve error message of tracking listener. (#66987)"
Breaks https://lab.llvm.org/buildbot/#/builders/5/builds/36953

This reverts commit a7530452fd163c84e83e662b549ade7b0fae9edf.
2023-09-25 09:07:17 -07:00
Ingo Müller
a7530452fd
[mlir][transform] Improve error message of tracking listener. (#66987)
This PR extends the error message of the tracking listener when
replacement ops cannot be found. That may happen if the applied patterns
replace an op by an op of a different kind or by block arguments.
However, this only matters if there are alive handles to the replaced
op. The new error message mentions that explicitly and reports the alive
handles.
2023-09-25 13:56:59 +02:00
Matthias Springer
6ebeecf452 [mlir][transform] ApplyConversionPatternsOp: Make conversion target optional
A conversion target is not needed. In a partial dialect conversion, ops are rewritten when possible. The dialect conversion succeeds if there are no illegal ops in the resulting IR.

Differential Revision: https://reviews.llvm.org/D157595
2023-08-10 13:53:25 +02:00
Matthias Springer
1102489891 [mlir][transform] Fix typo in ApplyConversionPatternsOp
Also add extra test cases for partial/full conversion.

Differential Revision: https://reviews.llvm.org/D157594
2023-08-10 13:52:56 +02:00
Matthias Springer
0bb4d4d32f [mlir][transform] Add ApplyToLLVMConversionPatternsOp
This op populates conversion patterns by querying the
ConvertToLLVMPatternInterface. Only dialects that support this interface
are supported.

Differential Revision: https://reviews.llvm.org/D157487
2023-08-09 13:44:47 +02:00
Matthias Springer
0cd5e02847 [mlir][transform] Improve assembly format of apply_conversion_patterns
This change was extracted from D157278.

Differential Revision: https://reviews.llvm.org/D157482
2023-08-09 11:07:22 +02:00
Matthias Springer
bcfdb3e4bc [mlir][transform] Add apply_conversion_patterns op
This transform op applies a dialect conversion to the targeted ops. Its design is similar to `apply_patterns`.

Patterns are specified in the first region of `apply_conversion_patterns`. They must implement the `ConversionPatternDescriptorOpInterface`. Regular rewrite patterns and dialect conversion patterns should not be mixed, so the interface is separate from the `PatternDescriptorOpInterface`.

The type converter is specified as the single op of the second region. It is optional; if no type converter is specified, it is expected that pattern descriptors provide their own type converters. If both the pattern descriptors and the `apply_conversion_patterns` op specify a type converter, the type converter of the pattern descriptor is used.

Differential Revision: https://reviews.llvm.org/D157109
2023-08-07 08:49:55 +02:00
Matthias Springer
20245ed4de [mlir][transform] Add apply_cse option to transform.apply_patterns op
Applying the canonicalizer and CSE in an interleaved fashion is useful after bufferization (and maybe other transforms) to fold away self copies.

Differential Revision: https://reviews.llvm.org/D155933
2023-07-21 15:13:56 +02:00
Matthias Springer
29e1fd9bdf [mlir][transform] Fix TrackingListener in regions that are isolated from above
When an operation is removed/replaced, the TrackingListener updates the internal transform state mapping between handles and payload IR. All handles must be updated, even the ones that are defined in a region that is beyond the most recent region that is isolated from above.

This fixes a bug, where a payload op was erased in a named sequence. Not only handles defined inside of the named region must be updated, but also all other handles such as the ones where the sequence is included.

Differential Revision: https://reviews.llvm.org/D153767
2023-06-26 18:05:24 +02:00
Matthias Springer
c63d2b2c71 [mlir][transform] Add TransformRewriter
All `apply` functions now have a `TransformRewriter &` parameter. This rewriter should be used to modify the IR. It has a `TrackingListener` attached and updates the internal handle-payload mappings based on rewrites.

Implementations no longer need to create their own `TrackingListener` and `IRRewriter`. Error checking is integrated into `applyTransform`. Tracking listener errors are reported only for ops with the `ReportTrackingListenerFailuresOpTrait` trait attached, allowing for a gradual migration. Furthermore, errors can be silenced with an op attribute.

Additional API will be added to `TransformRewriter` in subsequent revisions. This revision just adds an "empty" `TransformRewriter` class and updates all `apply` implementations.

Differential Revision: https://reviews.llvm.org/D152427
2023-06-20 10:49:59 +02:00
Matthias Springer
726d076784 [mlir][transform] ApplyPatternsOp: Add check to prevent modifying the transform IR
Add an extra check to make sure that transform IR is not getting modified by this op while it is being interpreted. This generally dangerous and we may want to enforce this for all transform ops that modify the payload in the future.

Users should generally try to apply patterns only to the piece of IR where it is needed (e.g., a matched function) and not the entire module (which may contain the transform IR).

This revision is in response to a crash in a downstream compiler that was caused by a dead `transform.structured.match` op that was removed by the GreedyPatternRewriteDriver's DCE while the enclosing sequence was being interpreted.

Differential Revision: https://reviews.llvm.org/D153113
2023-06-19 08:58:40 +02:00
Matthias Springer
cc7f52432b [mlir][transform] Use separate ops instead of PatternRegistry
* Remove `transform::PatternRegistry`.
* Add a new op for each currently registered pattern set.
* Change names of vector dialect pattern selector ops, so that they are consistent with the remaining code base.
* Remove redundant `transform.vector.extract_address_computations` op.

Differential Revision: https://reviews.llvm.org/D152249
2023-06-06 11:53:03 +02:00
Matthias Springer
5a10f207cc [mlir][transform] Add region to ApplyPatternsOp
Patterns should be selected by adding ops that implement `PatternDescriptorOpInterface` to the region of `apply_pattern` ops. Such ops can have operands, allowing for pattern parameterization. The existing way of selecting patterns from the PatternRegistry is deprecated.

Differential Revision: https://reviews.llvm.org/D152167
2023-06-06 09:09:41 +02:00
Matthias Springer
60f06bc5bb [mlir][transform] ApplyPatternsOp: Register canonicalization patterns
Also support replacing payload ops with ConstantLike ops in the TrackingListener, even if the replacement op does not have the same name. (Not supported for ops with multiple results, as this would require splitting the handle.)

Differential Revision: https://reviews.llvm.org/D152127
2023-06-05 11:37:43 +02:00
Matthias Springer
0b52fa900a [mlir][transform] Add ApplyPatternsOp and PatternRegistry
Add a new transform op that applies patterns to a targeted payload op. Patterns can be registered by transform dialect extensions in a pattern registry.

Differential Revision: https://reviews.llvm.org/D151983
2023-06-02 14:58:20 +02:00