One-Shot Bufferize currently does not support loops where a yielded
value bufferizes to a buffer that is different from the buffer of the
region iter_arg. In such a case, the bufferization fails with an error
such as:
```
Yield operand #0 is not equivalent to the corresponding iter bbArg
scf.yield %0 : tensor<5xf32>
```
One common reason for non-equivalent buffers is that an op on the path
from the region iter_arg to the terminator bufferizes out-of-place. Ops
that are analyzed earlier are more likely to bufferize in-place.
This commit adds a new heuristic that gives preference to ops that are
reachable on the reverse SSA use-def chain from a region terminator and
are within the parent region of the terminator. This is expected to work
better than the existing heuristics for loops where an iter_arg is
written to multiple times within a loop, but only one write is fed into
the terminator.
Current users of One-Shot Bufferize are not affected by this change.
"Bottom-up" is still the default heuristic. Users can switch to the new
heuristic manually.
This commit also turns the "fuzzer" pass option into a heuristic,
cleaning up the code a bit.
This commit removes the deallocation capabilities of
one-shot-bufferization. One-shot-bufferization should never deallocate
any memrefs as this should be entirely handled by the
ownership-based-buffer-deallocation pass going forward. This means the
`allow-return-allocs` pass option will default to true now,
`create-deallocs` defaults to false and they, as well as the escape
attribute indicating whether a memref escapes the current region, will
be removed. A new `allow-return-allocs-from-loops` option is added as a
temporary workaround for some bufferization limitations.
This reverts commit 6a91dfedeb956dfa092a6a3f411e8b02f0d5d289.
This caused problems in downstream projects. We are reverting to give
them more time for integration.
This is the first commit in a series with the goal to rework the
BufferDeallocation pass. Currently, this pass heavily relies on copies
to perform correct deallocations, which leads to very slow code and
potentially high memory usage. Additionally, there are unsupported cases
such as returning memrefs which this series of commits aims to add
support for as well.
This first commit removes the deallocation capabilities of
one-shot-bufferization.One-shot-bufferization should never deallocate any
memrefs as this should be entirely handled by the buffer-deallocation pass
going forward. This means the allow-return-allocs pass option will
default to true now, create-deallocs defaults to false and they, as well
as the escape attribute indicating whether a memref escapes the current region,
will be removed.
The documentation should w.r.t. these pass option changes should also be
updated in this commit.
Reviewed By: springerm
Differential Revision: https://reviews.llvm.org/D156662
To keep the pass simple, users should apply cleanup passes manually when necessary. In particular, `-cse -canonicalize` are often desireable to fold away self-copies that are created by the bufferization.
This addresses a comment in D120191.
Differential Revision: https://reviews.llvm.org/D155923
Bufferization already makes the assumption that buffers pass function
boundaries in the strided form and uses the corresponding affine map layouts.
Switch it to use the recently introduced strided layout instead to avoid
unnecessary casts when bufferizing further operations to the memref dialect
counterparts that now largely rely on the strided layout attribute.
Depends On D133947
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D133951
This allows for better type inference during bufferization and is in preparation of supporting memory spaces.
Differential Revision: https://reviews.llvm.org/D128581
This change changes the bufferization so that it utilizes the new TensorCopyInsertion pass. One-Shot Bufferize no longer calls the One-Shot Analysis. Instead, it relies on the TensorCopyInsertion pass to make the entire IR fully inplacable. The `bufferize` implementations of all ops are simplified; they no longer have to account for out-of-place bufferization decisions. These were already materialized in the IR in the form of `bufferization.alloc_tensor` ops during the TensorCopyInsertion pass.
Differential Revision: https://reviews.llvm.org/D127652
No longer pass static dim sizes as an attribute. This was redundant and required extra checks in the verifier. This change also makes the op symmetrical to memref::AllocOp.
Differential Revision: https://reviews.llvm.org/D126178
This change adds a new op `alloc_tensor` to the bufferization dialect. During bufferization, this op is always lowered to a buffer allocation (unless it is "eliminated" by a pre-processing pass). It is useful to have such an op in tensor land, because it allows users to model tensor SSA use-def chains (which drive bufferization decisions) and because tensor SSA use-def chains can be analyzed by One-Shot Bufferize, while memref values cannot.
This change also replaces all uses of linalg.init_tensor in bufferization-related code with bufferization.alloc_tensor.
linalg.init_tensor and bufferization.alloc_tensor are similar, but the purpose of the former one is just to carry a shape. It does not indicate a memory allocation.
linalg.init_tensor is not suitable for modelling SSA use-def chains for bufferization purposes, because linalg.init_tensor is marked as not having side effects (in contrast to alloc_tensor). As such, it is legal to move linalg.init_tensor ops around/CSE them/etc. This is not desirable for alloc_tensor; it represents an explicit buffer allocation while still in tensor land and such allocations should not suddenly disappear or get moved around when running the canonicalizer/CSE/etc.
BEGIN_PUBLIC
No public commit message needed for presubmit.
END_PUBLIC
Differential Revision: https://reviews.llvm.org/D126003
This changes replaces the `fully-dynamic-layout-maps` options (which was badly named) with two new options:
* `unknown-type-conversion` controls the layout maps on buffer types for which no layout map can be inferred.
* `function-boundary-type-conversion` controls the layout maps on buffer types inside of function signatures.
Differential Revision: https://reviews.llvm.org/D125615
* Move Module Bufferization to the bufferization dialect. The implementation is split into `OneShotModuleBufferize.cpp` and `FuncBufferizableOpInterfaceImpl.cpp`, so that the external model implementation can be easily moved to the func dialect in the future.
* Split and clean up test cases. A few test cases are still remaining in Linalg and will be updated separately.
* `linalg.inplaceable` is renamed to `bufferization.writable` to accurately reflect its current usage.
* Attributes and their verifiers are moved from the Linalg dialect to the Bufferization dialect.
* Expand documentation.
* Add a new flag to One-Shot Bufferize to allow for function boundary bufferization.
Differential Revision: https://reviews.llvm.org/D122229
This improves the modularity of the bufferization.
From now on, all ops that do not implement BufferizableOpInterface are considered hoisting barriers. Previously, all ops that do not implement the interface were not considered barriers and such ops had to be marked as barriers explicitly. This was unsafe because we could've hoisted across unknown ops where it was not safe to hoist.
As a side effect, this allows for cleaning up AffineBufferizableOpInterfaceImpl. This build unit no longer needed and can be deleted.
Differential Revision: https://reviews.llvm.org/D121519
Also add a TODO to switch to a custom walk instead of the GreedyPatternRewriter, which should be more efficient. (The bufferization pattern is guaranteed to apply only a single time for every op, so a simple walk should suffice.)
We currently specify a top-to-bottom walk order. This is important because other walk orders could introduce additional casts and/or buffer copies. These canonicalize away again, but it is more efficient to never generate them in the first place.
Note: A few of these canonicalizations are not yet implemented.
Differential Revision: https://reviews.llvm.org/D121518
The revision removes the linalg.fill operation and renames the OpDSL generated linalg.fill_tensor operation to replace it. After the change, all named structured operations are defined via OpDSL and there are no handwritten operations left.
A side-effect of the change is that the pretty printed form changes from:
```
%1 = linalg.fill(%cst, %0) : f32, tensor<?x?xf32> -> tensor<?x?xf32>
```
changes to
```
%1 = linalg.fill ins(%cst : f32) outs(%0 : tensor<?x?xf32>) -> tensor<?x?xf32>
```
Additionally, the builder signature now takes input and output value ranges as it is the case for all other OpDSL operations:
```
rewriter.create<linalg::FillOp>(loc, val, output)
```
changes to
```
rewriter.create<linalg::FillOp>(loc, ValueRange{val}, ValueRange{output})
```
All other changes remain minimal. In particular, the canonicalization patterns are the same and the `value()`, `output()`, and `result()` methods are now implemented by the FillOpInterface.
Depends On D120726
Reviewed By: nicolasvasilache
Differential Revision: https://reviews.llvm.org/D120728
The related functionality is moved over to the bufferization dialect. Test cases are cleaned up a bit.
Differential Revision: https://reviews.llvm.org/D120191
