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.
Add a new interface method to `BufferizableOpInterface`:
`hasTensorSemantics`. This method returns "true" if the op has tensor
semantics and should be bufferized.
Until now, we assumed that an op has tensor semantics if it has tensor
operands and/or tensor op results. However, there are ops like
`ml_program.global` that do not have any results/operands but must still
be bufferized (#75103). The new interface method can return "true" for
such ops.
This change also decouples `bufferization::bufferizeOp` a bit from the
func dialect.
There is currently an op interface for subset insertion ops
(`SubsetInsertionOpInterface`), but not for subset extraction ops. This
commit adds `SubsetExtractionOpInterface` to `mlir/Interfaces`, as well
as a common dependent op interface: `SubsetOpInterface`.
- `SubsetOpInterface` is for ops that operate on tensor subsets. It
provides interface methods to check if two subset ops operate on
equivalent or disjoint subsets. Ops that implement this interface must
implement either `SubsetExtractionOpInterface` or
`SubsetInsertionOpInterface`.
- `SubsetExtractionOpInterface` is for ops that extract from a tensor at
a subset. E.g., `tensor.extract_slice`, `tensor.gather`,
`vector.transfer_read`. Current implemented only on
`tensor.extract_slice`.
- `SubsetInsertionOpInterface` is for ops that insert into a destination
tensor at a subset. E.g., `tensor.insert_slice`,
`tensor.parallel_insert_slice`, `tensor.scatter`,
`vector.transfer_write`. Currently only implemented on
`tensor.insert_slice`, `tensor.parallel_insert_slice`.
Other changes:
- Rename `SubsetInsertionOpInterface.td` to `SubsetOpInterface.td`.
- Add helper functions to `ValueBoundsOpInterface.cpp` for checking
whether two slices are disjoint.
The new interfaces will be utilized by a new "loop-invariant subset
hoisting"
transformation. (This new transform is roughly
what `Linalg/Transforms/SubsetHoisting.cpp` is doing, but in a generic
and interface-driven way.)
`SubsetInsertionOpInterface` is an interface for ops that insert into a
destination tensor at a subset. It is currently used by the
bufferization framework to support efficient
`tensor.extract_slice/insert_slice` bufferization and to drive "empty
tensor elimination".
This commit moves the interface to `mlir/Interfaces`. This is in
preparation of adding a new "loop-invariant subset hoisting"
transformation to
`mlir/Transforms/Utils/LoopInvariantCodeMotionUtils.cpp`, which will
utilize `SubsetInsertionOpInterface`. (This new transform is roughly
what `Linalg/Transforms/SubsetHoisting.cpp` is doing, but in a generic
and interface-driven way.)
Remove the yielded tensor analysis. This analysis was used to detect
cases where One-Shot Bufferize cannot deallocate buffers. Deallocation
has recently been removed from One-Shot Bufferize. Buffers are now
deallocated by the buffer deallocation pass. This analysis is no longer
needed.
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 commit generalizes the special
tensor.extract_slice/tensor.insert_slice bufferization rules to tensor
subset ops.
Ops that insert a tensor into a tensor at a specified subset (e.g.,
tensor.insert_slice, tensor.scatter) can implement the
`SubsetInsertionOpInterface`.
Apart from adding a new op interface (extending the API), this change is
NFC. The only ops that currently implement the new interface are
tensor.insert_slice and tensor.parallel_insert_slice, and those ops were
are supported by One-Shot Bufferize.
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
One-Shot Bufferize correctly handles RaW conflicts around repetitive regions (loops). Specical handling is needed for parallel regions. These are a special kind of repetitive regions that can have additional RaW conflicts that would not be present if the regions would be executed sequentially.
Example:
```
%0 = bufferization.alloc_tensor()
scf.forall ... {
%1 = linalg.fill ins(...) outs(%0)
...
scf.forall.in_parallel {
tensor.parallel_insert_slice %1 into ...
}
}
```
A separate (private) buffer must be allocated for each iteration of the `scf.forall` loop.
This change adds a new interface method to `BufferizableOpInterface` to detect parallel regions. By default, regions are assumed to be sequential.
A buffer is privatized if an OpOperand bufferizes to a memory read inside a parallel region that is different from the parallel region where operand's value is defined.
Differential Revision: https://reviews.llvm.org/D159286
This revision adds support for unstructured control flow to the bufferization infrastructure. In particular: regions with multiple blocks, `cf.br`, `cf.cond_br`.
Two helper templates are added to `BufferizableOpInterface.h`, which can be implemented by ops that supported unstructured control flow in their regions (e.g., `func.func`) and ops that branch to another block (e.g., `cf.br`).
A block signature is always bufferized together with the op that owns the block.
Differential Revision: https://reviews.llvm.org/D158094
The operands for which elementwise access is relevant can now be specified. All other operands are ignored. This is useful because only two particular operands participate in a RaW conflict. Furthermore, the two tensors no longer must be equivalent to rule out conflicts due to elementwise access. Equivalent tensor sets may be formed after an inplace bufferization decision is made. The two tensors are actually not required to be equivalent. The only important thing is that they have "equivalent" indexing into the same base buffer.
Differential Revision: https://reviews.llvm.org/D158428
This revision is needed to support bufferization of `cf.br`/`cf.cond_br`. It will also be useful for better analysis of loop ops.
This revision generalizes `getAliasingOpResults` to `getAliasingValues`. An OpOperand can now not only alias with OpResults but also with BlockArguments. In the case of `cf.br` (will be added in a later revision): a `cf.br` operand will alias with the corresponding argument of the destination block.
If an op does not implement the `BufferizableOpInterface`, the analysis in conservative. It previously assumed that an OpOperand may alias with each OpResult. It now assumes that an OpOperand may alias with each OpResult and each BlockArgument of the entry block.
Differential Revision: https://reviews.llvm.org/D157957
Report an error when trying to bufferize an op that contains unstructured control flow but for ops for which the bufferization implementation does not support unstructured control flow. At the moment, there are no ops for which unstructured control flow is supported.
Differential Revision: https://reviews.llvm.org/D157893
This implication was already done de-facto and there were plenty of users and wrapper functions specifically used to handle the "return-like or RegionBranchTerminatorOpInterface" case. These simply existed due to up until recently missing features in ODS.
With the new capabilities of traits, we can make `ReturnLike` imply `RegionBranchTerminatorOpInterface` and auto generate proper definitions for its methods.
Various occurrences and wrapper methods used for `isa<RegionBranchTerminatorOpInterface>() || hasTrait<ReturnLike>()` have all been removed.
Differential Revision: https://reviews.llvm.org/D157402
Before this change, two equivalent operands that bufferize to a memory read and write, respectively, were always conflicting. This change improves the analysis for ops that bufferize to element-wise access. Such ops can bufferize in-place, because an original element value is not needed anymore after computing and writing an updated element value.
This change allows ops such as the following one to bufferize in-place:
```
%0 = linalg.elemwise_binary {fun = #linalg.binary_fn<add>}
ins(%a, %b : tensor<5xf32>, tensor<5xf32>)
outs(%a : tensor<5xf32>) -> tensor<5xf32>
```
Differential Revision: https://reviews.llvm.org/D156887
The `getEnclosingRepetitiveRegion` functions walk the ancestor regions everytime which can be expensive especially when there are multiple regions inbetween. This commit adds a cache to the bufferization analysis to remember the result of the walk.
Reviewed By: springerm
Differential Revision: https://reviews.llvm.org/D154710
Until now, only `tensor.pad` ops could be bufferized to an allocation. This revision adds support for all bufferizable ops that do not already bufferize to an allocation. (Those still need special handling.)
Differential Revision: https://reviews.llvm.org/D153971
bufferization.to_memref ops are allowed in One-Shot Bufferize, but they are treated conservatively: in the absence of a memref analysis, we have to assume that the result buffer is read and written.
Note: to_memref cannot introduce any future aliases that would have to be considered during One-Shot Bufferize, because only to_tensor ops with the `restrict` attribute are supported. Such tensors are guaranteed to not alias with any other buffer after bufferization.
Differential Revision: https://reviews.llvm.org/D153365
The MLIR classes Type/Attribute/Operation/Op/Value support
cast/dyn_cast/isa/dyn_cast_or_null functionality through llvm's doCast
functionality in addition to defining methods with the same name.
This change begins the migration of uses of the method to the
corresponding function call as has been decided as more consistent.
Note that there still exist classes that only define methods directly,
such as AffineExpr, and this does not include work currently to support
a functional cast/isa call.
Context:
- https://mlir.llvm.org/deprecation/ at "Use the free function variants
for dyn_cast/cast/isa/…"
- Original discussion at https://discourse.llvm.org/t/preferred-casting-style-going-forward/68443
Implementation:
This patch updates all remaining uses of the deprecated functionality in
mlir/. This was done with clang-tidy as described below and further
modifications to GPUBase.td and OpenMPOpsInterfaces.td.
Steps are described per line, as comments are removed by git:
0. Retrieve the change from the following to build clang-tidy with an
additional check:
main...tpopp:llvm-project:tidy-cast-check
1. Build clang-tidy
2. Run clang-tidy over your entire codebase while disabling all checks
and enabling the one relevant one. Run on all header files also.
3. Delete .inc files that were also modified, so the next build rebuilds
them to a pure state.
```
ninja -C $BUILD_DIR clang-tidy
run-clang-tidy -clang-tidy-binary=$BUILD_DIR/bin/clang-tidy -checks='-*,misc-cast-functions'\
-header-filter=mlir/ mlir/* -fix
rm -rf $BUILD_DIR/tools/mlir/**/*.inc
```
Differential Revision: https://reviews.llvm.org/D151542
Instead of passing traversal options as a long list of arguments, store them in a TraversalConfig object and pass that object.
Differential Revision: https://reviews.llvm.org/D143927
The MLIR classes Type/Attribute/Operation/Op/Value support
cast/dyn_cast/isa/dyn_cast_or_null functionality through llvm's doCast
functionality in addition to defining methods with the same name.
This change begins the migration of uses of the method to the
corresponding function call as has been decided as more consistent.
Note that there still exist classes that only define methods directly,
such as AffineExpr, and this does not include work currently to support
a functional cast/isa call.
Caveats include:
- This clang-tidy script probably has more problems.
- This only touches C++ code, so nothing that is being generated.
Context:
- https://mlir.llvm.org/deprecation/ at "Use the free function variants
for dyn_cast/cast/isa/…"
- Original discussion at https://discourse.llvm.org/t/preferred-casting-style-going-forward/68443
Implementation:
This first patch was created with the following steps. The intention is
to only do automated changes at first, so I waste less time if it's
reverted, and so the first mass change is more clear as an example to
other teams that will need to follow similar steps.
Steps are described per line, as comments are removed by git:
0. Retrieve the change from the following to build clang-tidy with an
additional check:
https://github.com/llvm/llvm-project/compare/main...tpopp:llvm-project:tidy-cast-check
1. Build clang-tidy
2. Run clang-tidy over your entire codebase while disabling all checks
and enabling the one relevant one. Run on all header files also.
3. Delete .inc files that were also modified, so the next build rebuilds
them to a pure state.
4. Some changes have been deleted for the following reasons:
- Some files had a variable also named cast
- Some files had not included a header file that defines the cast
functions
- Some files are definitions of the classes that have the casting
methods, so the code still refers to the method instead of the
function without adding a prefix or removing the method declaration
at the same time.
```
ninja -C $BUILD_DIR clang-tidy
run-clang-tidy -clang-tidy-binary=$BUILD_DIR/bin/clang-tidy -checks='-*,misc-cast-functions'\
-header-filter=mlir/ mlir/* -fix
rm -rf $BUILD_DIR/tools/mlir/**/*.inc
git restore mlir/lib/IR mlir/lib/Dialect/DLTI/DLTI.cpp\
mlir/lib/Dialect/Complex/IR/ComplexDialect.cpp\
mlir/lib/**/IR/\
mlir/lib/Dialect/SparseTensor/Transforms/SparseVectorization.cpp\
mlir/lib/Dialect/Vector/Transforms/LowerVectorMultiReduction.cpp\
mlir/test/lib/Dialect/Test/TestTypes.cpp\
mlir/test/lib/Dialect/Transform/TestTransformDialectExtension.cpp\
mlir/test/lib/Dialect/Test/TestAttributes.cpp\
mlir/unittests/TableGen/EnumsGenTest.cpp\
mlir/test/python/lib/PythonTestCAPI.cpp\
mlir/include/mlir/IR/
```
Differential Revision: https://reviews.llvm.org/D150123
`restrict` is similar to the C++ restrict keyword. Results of `to_tensor` that have the `restrict` attribute are guaranteed to not alias any other `to_tensor` result (after bufferization).
Note: Since `to_memref` ops are not supported by One-Shot Bufferize and all bufferizable ops follow DPS rules (i.e., the buffer of the result is the buffer of an operand or an alias thereof), the buffer of a `to_tensor` op that has the `restrict` attribute is always an entirely "new" buffer that is not aliasing with the future buffer of any tensor value in the entire program. This makes such `to_tensor` ops "safe" from a bufferization perspective; they cannot cause RaW conflicts.
Differential Revision: https://reviews.llvm.org/D144021
`getAliasingOpOperands`/`getAliasingOpResults` now encodes OpOperand/OpResult, buffer relation and a degree of certainty. E.g.:
```
// aliasingOpOperands(%r) = {(%t, EQUIV, DEFINITE)}
// aliasingOpResults(%t) = {(%r, EQUIV, DEFINITE)}
%r = tensor.insert %f into %t[%idx] : tensor<?xf32>
// aliasingOpOperands(%r) = {(%t0, EQUIV, MAYBE), (%t1, EQUIV, MAYBE)}
// aliasingOpResults(%t0) = {(%r, EQUIV, MAYBE)}
// aliasingOpResults(%t1) = {(%r, EQUIV, MAYBE)}
%r = arith.select %c, %t0, %t1 : tensor<?xf32>
```
`BufferizableOpInterface::bufferRelation` is removed, as it is now part of `getAliasingOpOperands`/`getAliasingOpResults`.
This change allows for better analysis, in particular wrt. equivalence. This allows additional optimizations and better error checking (which is sometimes overly conservative). Examples:
* EmptyTensorElimination can eliminate `tensor.empty` inside `scf.if` blocks. This requires a modeling of equivalence: It is not a per-OpResult property anymore. Instead, it can be specified for each OpOperand and OpResult. This is important because `tensor.empty` may be eliminated only if all values on the SSA use-def chain to the final consumer (`tensor.insert_slice`) are equivalent.
* The detection of "returning allocs from a block" can be improved. (Addresses a TODO in `assertNoAllocsReturned`.) This allows us to bufferize IR such as "yielding a `tensor.extract_slice` result from an `scf.if` branch", which currently fails to bufferize because the alloc detection is too conservative.
* Better bufferization of loops. Aliases of the iter_arg can be yielded (even if they are not equivalent) without having to realloc and copy the entire buffer on each iteration.
The above-mentioned examples are not yet implemented with this change. This change just improves the BufferizableOpInterface, its implementations and related helper functions, so that better aliasing information is available for each op.
Differential Revision: https://reviews.llvm.org/D142129
This is in preparation of reusing the same AnalysisState for tensor.empty elimination and One-Shot Bufferize (to address performance bottlenecks).
Differential Revision: https://reviews.llvm.org/D143379
There is no longer a need to keep the two separate. This is in preparation of reusing the same AnalysisState for tensor.empty elimination and One-Shot Bufferize (to address performance bottlenecks).
Differential Revision: https://reviews.llvm.org/D143313
The previous strategy was too complex and faulty. Op dominance cannot be used to rule out RaW conflicts due to op ordering if the reading op and the conflicting writing op are in a sub repetitive region of the closest enclosing repetitive region of the definition of the read value.
Differential Revision: https://reviews.llvm.org/D143087
Reading from tensor.empty or bufferization.alloc_tensor (without copy) cannot cause a conflict because these ops do not specify the contents of their result tensors.
Differential Revision: https://reviews.llvm.org/D143183
* `getAliasingOpOperand` => `getAliasingOpOperands`
* `getAliasingOpResult` => `getAliasingOpResults`
Also a few minor code cleanups and better documentation.
Differential Revision: https://reviews.llvm.org/D142979
The previous lingo was confusing. There are no writes on tensors. There are only definitions.
Also some minor cleanup and better documentation.
Differential Revision: https://reviews.llvm.org/D141790
The name of the method was confusing. It is bufferizesToMemoryWrite, but from the perspective of OpResults.
`bufferizesToMemoryWrite(OpResult)` now supports ops with regions that do not have aliasing OpOperands (such as `scf.if`). These ops no longer need to implement `isMemoryWrite`.
Differential Revision: https://reviews.llvm.org/D141684
These functions are generally useful and not specific to One-Shot Analysis. Move them to `BufferizableOpInterface.h` and make them public.
Differential Revision: https://reviews.llvm.org/D141685
The op is not bufferizable but should be analyzable (for `EliminateEmptyTensors`, which uses the bufferization infrastructure).
Also improve debugging functionality and error messages.
Also adds a missing pass to the sparse pipeline. (tensor.empty should be replaced with bufferization.alloc_tensor, but it sometimes used to work without depending on how the tensor.empty is used. Now we always fail explicitly.)
This patch mechanically replaces None with std::nullopt where the
compiler would warn if None were deprecated. The intent is to reduce
the amount of manual work required in migrating from Optional to
std::optional.
This is part of an effort to migrate from llvm::Optional to
std::optional:
https://discourse.llvm.org/t/deprecating-llvm-optional-x-hasvalue-getvalue-getvalueor/63716
The previous error message was confusing. Also improve code documentation and some minor code cleanups.
Differential Revision: https://reviews.llvm.org/D138902
`DialectAnalysisState` is now `OneShotAnalysisState::Extension`.
This state extension mechanism is needed only for One-Shot Analysis, so it is moved from `BufferizableOpInterface.h` to `OneShotAnalysis.h`.
Extensions are now identified via TypeIDs instead of StringRefs. The API of state extensions is cleaned up and follows the same pattern as other extension mechanisms in MLIR (e.g., `transform::TransformState::Extension`).
Also delete some dead code.
Differential Revision: https://reviews.llvm.org/D135051
