To match an interface or trait, users currently have to use the `MatchAny` tag. This tag can be quite problematic for compile time for things like the canonicalizer, as the `MatchAny` patterns may get applied to *every* operation. This revision adds better support by bucketing interface/trait patterns based on which registered operations have them registered. This means that moving forward we will only attempt to match these patterns to operations that have this interface registered. Two simplify defining patterns that match traits and interfaces, two new utility classes have been added: OpTraitRewritePattern and OpInterfaceRewritePattern.
Differential Revision: https://reviews.llvm.org/D98986
This nicely aligns the naming with RewritePatternSet. This type isn't
as widely used, but we keep a using declaration in to help with
downstream consumption of this change.
Differential Revision: https://reviews.llvm.org/D99131
This doesn't change APIs, this just cleans up the many in-tree uses of these
names to use the new preferred names. We'll keep the old names around for a
couple weeks to help transitions.
Differential Revision: https://reviews.llvm.org/D99127
GreedyPatternRewriteDriver was changed from bottom-up traversal to top-down traversal. Not all passes work yet with that change for traversal order. To give some time for fixing, add an option to allow to switch back to bottom-up traversal. Use this option in FusionOfTensorOpsPass which fails otherwise.
Differential Revision: https://reviews.llvm.org/D99059
This updates the codebase to pass the context when creating an instance of
OwningRewritePatternList, and starts removing extraneous MLIRContext
parameters. There are many many more to be removed.
Differential Revision: https://reviews.llvm.org/D99028
This reapplies b5d9a3c / https://reviews.llvm.org/D98609 with a one line fix in
processExistingConstants to skip() when erasing a constant we've already seen.
Original commit message:
1) Change the canonicalizer to walk the function in top-down order instead of
bottom-up order. This composes well with the "top down" nature of constant
folding and simplification, reducing iterations and re-evaluation of ops in
simple cases.
2) Explicitly enter existing constants into the OperationFolder table before
canonicalizing. Previously we would "constant fold" them and rematerialize
them, wastefully recreating a bunch fo constants, which lead to pointless
memory traffic.
Both changes together provide a 33% speedup for canonicalize on some mid-size
CIRCT examples.
One artifact of this change is that the constants generated in normal pattern
application get inserted at the top of the function as the patterns are applied.
Because of this, we get "inverted" constants more often, which is an aethetic
change to the IR but does permute some testcases.
Differential Revision: https://reviews.llvm.org/D99006
This allows for notifying callers when operations/blocks get erased, which is especially useful for the greedy pattern driver. The current greedy pattern driver "throws away" all information on constants in the operation folder because it doesn't know if they get erased or not. By passing in RewriterBase, we can directly track this and prevent the need for the pattern driver to rediscover all of the existing constants. In some situations this cuts the compile time of the canonicalizer in half.
Differential Revision: https://reviews.llvm.org/D98755
When deleting operations in DCE, the algorithm uses a post-order walk of
the IR to ensure that value uses were erased before value defs. Graph
regions do not have the same structural invariants as SSA CFG, and this
post order walk could delete value defs before uses. This problem is
guaranteed to occur when there is a cycle in the use-def graph.
This change stops DCE from visiting the operations and blocks in any
meaningful order. Instead, we rely on explicitly dropping all uses of a
value before deleting it.
Reviewed By: mehdi_amini, rriddle
Differential Revision: https://reviews.llvm.org/D98919
This reverts commit b5d9a3c923.
The commit introduced a memory error in canonicalization/operation
walking that is exposed when compiled with ASAN. It leads to crashes in
some "release" configurations.
We know that all ConstantLike operations have one result and no operands,
so check this first before doing the trait check. This change speeds up
Canonicalize on a CIRCT testcase by ~5%.
Differential Revision: https://reviews.llvm.org/D98615
Two changes:
1) Change the canonicalizer to walk the function in top-down order instead of
bottom-up order. This composes well with the "top down" nature of constant
folding and simplification, reducing iterations and re-evaluation of ops in
simple cases.
2) Explicitly enter existing constants into the OperationFolder table before
canonicalizing. Previously we would "constant fold" them and rematerialize
them, wastefully recreating a bunch fo constants, which lead to pointless
memory traffic.
Both changes together provide a 33% speedup for canonicalize on some mid-size
CIRCT examples.
One artifact of this change is that the constants generated in normal pattern
application get inserted at the top of the function as the patterns are applied.
Because of this, we get "inverted" constants more often, which is an aethetic
change to the IR but does permute some testcases.
Differential Revision: https://reviews.llvm.org/D98609
The current implementation has some inefficiencies that become noticeable when running on large modules. This revision optimizes the code, and updates some out-dated idioms with newer utilities. The main components of this optimization include:
* Add an overload of Block::eraseArguments that allows for O(N) erasure of disjoint arguments.
* Don't process entry block arguments given that we don't erase them at this point.
* Don't track individual operation results, given that we don't erase them. We can just track the parent operation.
Differential Revision: https://reviews.llvm.org/D98309
This makes it easy to compose the distribution computation with
other affine computations.
Reviewed By: mravishankar
Differential Revision: https://reviews.llvm.org/D98171
The current implementation of Value involves a pointer int pair with several different kinds of owners, i.e. BlockArgumentImpl*, Operation *, TrailingOpResult*. This design arose from the desire to save memory overhead for operations that have a very small number of results (generally 0-2). There are, unfortunately, many problematic aspects of the current implementation that make Values difficult to work with or just inefficient.
Operation result types are stored as a separate array on the Operation. This is very inefficient for many reasons: we use TupleType for multiple results, which can lead to huge amounts of memory usage if multi-result operations change types frequently(they do). It also means that simple methods like Value::getType/Value::setType now require complex logic to get to the desired type.
Value only has one pointer bit free, severely limiting the ability to use it in things like PointerUnion/PointerIntPair. Given that we store the kind of a Value along with the "owner" pointer, we only leave one bit free for users of Value. This creates situations where we end up nesting PointerUnions to be able to use Value in one.
As noted above, most of the methods in Value need to branch on at least 3 different cases which is both inefficient, possibly error prone, and verbose. The current storage of results also creates problems for utilities like ValueRange/TypeRange, which want to efficiently store base pointers to ranges (of which Operation* isn't really useful as one).
This revision greatly simplifies the implementation of Value by the introduction of a new ValueImpl class. This class contains all of the state shared between all of the various derived value classes; i.e. the use list, the type, and the kind. This shared implementation class provides several large benefits:
* Most of the methods on value are now branchless, and often one-liners.
* The "kind" of the value is now stored in ValueImpl instead of Value
This frees up all of Value's pointer bits, allowing for users to take full advantage of PointerUnion/PointerIntPair/etc. It also allows for storing more operation results as "inline", 6 now instead of 2, freeing up 1 word per new inline result.
* Operation result types are now stored in the result, instead of a side array
This drops the size of zero-result operations by 1 word. It also removes the memory crushing use of TupleType for operations results (which could lead up to hundreds of megabytes of "dead" TupleTypes in the context). This also allowed restructured ValueRange, making it simpler and one word smaller.
This revision does come with two conceptual downsides:
* Operation::getResultTypes no longer returns an ArrayRef<Type>
This conceptually makes some usages slower, as the iterator increment is slightly more complex.
* OpResult::getOwner is slightly more expensive, as it now requires a little bit of arithmetic
From profiling, neither of the conceptual downsides have resulted in any perceivable hit to performance. Given the advantages of the new design, most compiles are slightly faster.
Differential Revision: https://reviews.llvm.org/D97804
This patch continues detensorizing implementation by detensoring
internal control flow in functions.
In order to detensorize functions, all the non-entry block's arguments
are detensored and branches between such blocks are properly updated to
reflect the detensored types as well. Function entry block (signature)
is left intact.
This continues work towards handling github/google/iree#1159.
Reviewed By: silvas
Differential Revision: https://reviews.llvm.org/D97148
Just a pure method renaming.
It is a preparation step for replacing "memory space as raw integer"
with more generic "memory space as attribute", which will be done in
separate commit.
The `MemRefType::getMemorySpace` method will return `Attribute` and
become the main API, while `getMemorySpaceAsInt` will be declared as
deprecated and will be replaced in all in-tree dialects (also in separate
commits).
Reviewed By: mehdi_amini, rriddle
Differential Revision: https://reviews.llvm.org/D97476
This also exposed a bug in Dialect loading where it was not correctly identifying identifiers that had the dialect namespace as a prefix.
Differential Revision: https://reviews.llvm.org/D97431
Fixes a bug in affine fusion pipeline where an incorrect fusion is performed
despite a Call Op that potentially modifies memrefs under consideration
exists between source and target.
Fixes part of https://bugs.llvm.org/show_bug.cgi?id=49220
Reviewed By: bondhugula, dcaballe
Differential Revision: https://reviews.llvm.org/D97252
This patch handles defining ops between the source and dest loop nests, and prevents loop nests with `iter_args` from being fused.
If there is any SSA value in the dest loop nest whose defining op has dependence from the source loop nest, we cannot fuse the loop nests.
If there is a `affine.for` with `iter_args`, prevent it from being fused.
Reviewed By: dcaballe, bondhugula
Differential Revision: https://reviews.llvm.org/D97030
This prevents a bug in the pass instrumentation implementation where the main thread would end up with a different pass manager in different runs of the pass.
llvm::parallelTransformReduce does not schedule work on the caller thread, which becomes very costly for
the inliner where a majority of SCCs are small, often ~1 element. The switch to llvm::parallelForEach solves this,
and also aligns the implementation with the PassManager (which realistically should share the same implementation).
This change dropped compile time on an internal benchmark by ~1(25%) second.
Differential Revision: https://reviews.llvm.org/D96086
Affine parallel ops may contain and yield results from MemRefsNormalizable ops in the loop body. Thus, both affine.parallel and affine.yield should have the MemRefsNormalizable trait.
Reviewed By: bondhugula
Differential Revision: https://reviews.llvm.org/D96821
The current implementation of tilePerfectlyNested utility doesn't handle
the non-unit step size. We have added support to perform tiling
correctly even if the step size of the loop to be tiled is non-unit.
Fixes https://bugs.llvm.org/show_bug.cgi?id=49188.
Differential Revision: https://reviews.llvm.org/D97037
This commit fixes a bug in affine fusion pipeline where an
incorrect fusion is performed despite a dealloc op is present
between a producer and a consumer. This is done by creating a
node for dealloc op in the MDG.
Reviewed By: bondhugula, dcaballe
Differential Revision: https://reviews.llvm.org/D97032
This commit introduced a cyclic dependency:
Memref dialect depends on Standard because it used ConstantIndexOp.
Std depends on the MemRef dialect in its EDSC/Intrinsics.h
Working on a fix.
This reverts commit 8aa6c3765b.
Create the memref dialect and move several dialect-specific ops without
dependencies to other ops from std dialect to this dialect.
Moved ops:
AllocOp -> MemRef_AllocOp
AllocaOp -> MemRef_AllocaOp
DeallocOp -> MemRef_DeallocOp
MemRefCastOp -> MemRef_CastOp
GetGlobalMemRefOp -> MemRef_GetGlobalOp
GlobalMemRefOp -> MemRef_GlobalOp
PrefetchOp -> MemRef_PrefetchOp
ReshapeOp -> MemRef_ReshapeOp
StoreOp -> MemRef_StoreOp
TransposeOp -> MemRef_TransposeOp
ViewOp -> MemRef_ViewOp
The roadmap to split the memref dialect from std is discussed here:
https://llvm.discourse.group/t/rfc-split-the-memref-dialect-from-std/2667
Differential Revision: https://reviews.llvm.org/D96425
Separating the AffineMapAccessInterface from AffineRead/WriteOp interface so that dialects which extend Affine capabilities (e.g. PlaidML PXA = parallel extensions for Affine) can utilize relevant passes (e.g. MemRef normalization).
Reviewed By: bondhugula
Differential Revision: https://reviews.llvm.org/D96284
SliceAnalysis originally was developed in the context of affine.for within mlfunc.
It predates the notion of region.
This revision updates it to not hardcode specific ops like scf::ForOp.
When rooted at an op, the behavior of the slice computation changes as it recurses into the regions of the op. This does not support gathering all values transitively depending on a loop induction variable anymore.
Additional variants rooted at a Value are added to also support the existing behavior.
Differential revision: https://reviews.llvm.org/D96702
These properties were useful for a few things before traits had a better integration story, but don't really carry their weight well these days. Most of these properties are already checked via traits in most of the code. It is better to align the system around traits, and improve the performance/cost of traits in general.
Differential Revision: https://reviews.llvm.org/D96088
This makes ignoring a result explicit by the user, and helps to prevent accidental errors with dropped results. Marking LogicalResult as no discard was always the intention from the beginning, but got lost along the way.
Differential Revision: https://reviews.llvm.org/D95841
In dialect conversion infrastructure, source materialization applies as part of
the finalization procedure to results of the newly produced operations that
replace previously existing values with values having a different type.
However, such operations may be created to replace operations created in other
patterns. At this point, it is possible that the results of the _original_
operation are still in use and have mismatching types, but the results of the
_intermediate_ operation that performed the type change are not in use leading
to the absence of source materialization. For example,
%0 = dialect.produce : !dialect.A
dialect.use %0 : !dialect.A
can be replaced with
%0 = dialect.other : !dialect.A
%1 = dialect.produce : !dialect.A // replaced, scheduled for removal
dialect.use %1 : !dialect.A
and then with
%0 = dialect.final : !dialect.B
%1 = dialect.other : !dialect.A // replaced, scheduled for removal
%2 = dialect.produce : !dialect.A // replaced, scheduled for removal
dialect.use %2 : !dialect.A
in the same rewriting, but only the %1->%0 replacement is currently considered.
Change the logic in dialect conversion to look up all values that were replaced
by the given value and performing source materialization if any of those values
is still in use with mismatching types. This is performed by computing the
inverse value replacement mapping. This arguably expensive manipulation is
performed only if there were some type-changing replacements. An alternative
could be to consider all replaced operations and not only those that resulted
in type changes, but it would harm pattern-level composability: the pattern
that performed the non-type-changing replacement would have to be made aware of
the type converter in order to call the materialization hook.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D95626
We could extend this with an interface to allow dialect to perform a type
conversion, but that would make the folder creating operation which isn't
the case at the moment, and isn't necessarily always desirable.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D95991
In dialect conversion, signature conversions essentially perform block argument
replacement and are added to the general value remapping. However, the replaced
values were not tracked, so if a signature conversion was rolled back, the
construction of operand lists for the following patterns could have obtained
block arguments from the mapping and give them to the pattern leading to
use-after-free. Keep track of signature conversions similarly to normal block
argument replacement, and erase such replacements from the general mapping when
the conversion is rolled back.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D95688
Currently, for a scf.parallel (i,j,k) after the loop collapsing to 1D is done, the
IVs would be traversed as for an scf.parallel(k,j,i).
Differential Revision: https://reviews.llvm.org/D95693
This patch adds support for producer-consumer fusion scenarios with
multiple producer stores to the AffineLoopFusion pass. The patch
introduces some changes to the producer-consumer algorithm, including:
* For a given consumer loop, producer-consumer fusion iterates over its
producer candidates until a fixed point is reached.
* Producer candidates are gathered beforehand for each iteration of the
consumer loop and visited in reverse program order (not strictly guaranteed)
to maximize the number of loops fused per iteration.
In general, these changes were needed to simplify the multi-store producer
support and remove some of the workarounds that were introduced in the past
to support more fusion cases under the single-store producer limitation.
This patch also preserves the existing functionality of AffineLoopFusion with
one minor change in behavior. Producer-consumer fusion didn't fuse scenarios
with escaping memrefs and multiple outgoing edges (from a single store).
Multi-store producer scenarios will usually (always?) have multiple outgoing
edges so we couldn't fuse any with escaping memrefs, which would greatly limit
the applicability of this new feature. Therefore, the patch enables fusion for
these scenarios. Please, see modified tests for specific details.
Reviewed By: andydavis1, bondhugula
Differential Revision: https://reviews.llvm.org/D92876
This extracts the implementation of getType, setType, and getBody from
FunctionSupport.h into the mlir::impl namespace and defines them
generically in FunctionSupport.cpp. This allows them to be used
elsewhere for any FunctionLike ops that use FunctionType for their
type signature.
Using the new helpers, FuncOpSignatureConversion is generalized to
work with all such FunctionLike ops. Convenience helpers are added to
configure the pattern for a given concrete FunctionLike op type.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D95021
This patch adds support for producer-consumer fusion scenarios with
multiple producer stores to the AffineLoopFusion pass. The patch
introduces some changes to the producer-consumer algorithm, including:
* For a given consumer loop, producer-consumer fusion iterates over its
producer candidates until a fixed point is reached.
* Producer candidates are gathered beforehand for each iteration of the
consumer loop and visited in reverse program order (not strictly guaranteed)
to maximize the number of loops fused per iteration.
In general, these changes were needed to simplify the multi-store producer
support and remove some of the workarounds that were introduced in the past
to support more fusion cases under the single-store producer limitation.
This patch also preserves the existing functionality of AffineLoopFusion with
one minor change in behavior. Producer-consumer fusion didn't fuse scenarios
with escaping memrefs and multiple outgoing edges (from a single store).
Multi-store producer scenarios will usually (always?) have multiple outgoing
edges so we couldn't fuse any with escaping memrefs, which would greatly limit
the applicability of this new feature. Therefore, the patch enables fusion for
these scenarios. Please, see modified tests for specific details.
Reviewed By: andydavis1, bondhugula
Differential Revision: https://reviews.llvm.org/D92876
Add a check if regions do not implement the RegionBranchOpInterface. This is not
allowed in the current deallocation steps. Furthermore, we handle edge-cases,
where a single region is attached and the parent operation has no results.
This fixes: https://bugs.llvm.org/show_bug.cgi?id=48575
Differential Revision: https://reviews.llvm.org/D94586
This revision adds a new `replaceOpWithIf` hook that replaces uses of an operation that satisfy a given functor. If all uses are replaced, the operation gets erased in a similar manner to `replaceOp`. DialectConversion support will be added in a followup as this requires adjusting how replacements are tracked there.
Differential Revision: https://reviews.llvm.org/D94632
This corrects the last 2 issues caught by tests when causing dialect
conversion rollbacks to occur.
Differential Revision: https://reviews.llvm.org/D94623
getDynOperands behavior is commonly used in a number of passes. Refactored to
use a helper function and avoid code reuse.
Differential Revision: https://reviews.llvm.org/D94340
This revision adds a new `initialize(MLIRContext *)` hook to passes that allows for them to initialize any heavy state before the first execution of the pass. A concrete use case of this is with patterns that rely on PDL, given that PDL is compiled at run time it is imperative that compilation results are cached as much as possible. The first use of this hook is in the Canonicalizer, which has the added benefit of reducing the number of expensive accesses to the context when collecting patterns.
Differential Revision: https://reviews.llvm.org/D93147
This class used to serve a few useful purposes:
* Allowed containing a null DictionaryAttr
* Provided some simple mutable API around a DictionaryAttr
The first of which is no longer an issue now that there is much better caching support for attributes in general, and a cache in the context for empty dictionaries. The second results in more trouble than it's worth because it mutates the internal dictionary on every action, leading to a potentially large number of dictionary copies. NamedAttrList is a much better alternative for the second use case, and should be modified as needed to better fit it's usage as a DictionaryAttrBuilder.
Differential Revision: https://reviews.llvm.org/D93442
This better matches the rest of the infrastructure, is much simpler, and makes it easier to move these types to being declaratively specified.
Differential Revision: https://reviews.llvm.org/D93432
The current condition implies that the target materialization will be
called even if the type is the new operand type is legal, but slightly
different. For example, if there is a bufferization pattern that changes
memref layout, then target materialization for an illegal type
(TensorType) would be called.
Differential Revision: https://reviews.llvm.org/D93126
Now that passes have support for running nested pipelines, the inliner can now allow for users to provide proper nested pipelines to use for optimization during inlining. This revision also changes the behavior of optimization during inlining to optimize before attempting to inline, which should lead to a more accurate cost model and prevents the need for users to schedule additional duplicate cleanup passes before/after the inliner that would already be run during inlining.
Differential Revision: https://reviews.llvm.org/D91211
This reverts commit 0d48d265db.
This reapplies the following commit, with a fix for CAPI/ir.c:
[mlir] Start splitting the `tensor` dialect out of `std`.
This starts by moving `std.extract_element` to `tensor.extract` (this
mirrors the naming of `vector.extract`).
Curiously, `std.extract_element` supposedly works on vectors as well,
and this patch removes that functionality. I would tend to do that in
separate patch, but I couldn't find any downstream users relying on
this, and the fact that we have `vector.extract` made it seem safe
enough to lump in here.
This also sets up the `tensor` dialect as a dependency of the `std`
dialect, as some ops that currently live in `std` depend on
`tensor.extract` via their canonicalization patterns.
Part of RFC: https://llvm.discourse.group/t/rfc-split-the-tensor-dialect-from-std/2347/2
Differential Revision: https://reviews.llvm.org/D92991
This starts by moving `std.extract_element` to `tensor.extract` (this
mirrors the naming of `vector.extract`).
Curiously, `std.extract_element` supposedly works on vectors as well,
and this patch removes that functionality. I would tend to do that in
separate patch, but I couldn't find any downstream users relying on
this, and the fact that we have `vector.extract` made it seem safe
enough to lump in here.
This also sets up the `tensor` dialect as a dependency of the `std`
dialect, as some ops that currently live in `std` depend on
`tensor.extract` via their canonicalization patterns.
Part of RFC: https://llvm.discourse.group/t/rfc-split-the-tensor-dialect-from-std/2347/2
Differential Revision: https://reviews.llvm.org/D92991
OperationFolder currently uses ConstantOp as a backup when trying to materialize a constant after an operation is folded. This dependency isn't really useful or necessary given that dialects can/should provide a `materializeConstant` implementation.
Fixes PR#44866
Differential Revision: https://reviews.llvm.org/D92980
This fixes a subtle bug where SCCP could incorrectly optimize a private callable while waiting for its arguments to be resolved.
Fixes PR#48457
Differential Revision: https://reviews.llvm.org/D92976
This is part of a larger refactoring the better congregates the builtin structures under the BuiltinDialect. This also removes the problematic "standard" naming that clashes with the "standard" dialect, which is not defined within IR/. A temporary forward is placed in StandardTypes.h to allow time for downstream users to replaced references.
Differential Revision: https://reviews.llvm.org/D92435
Memrefs with affine_map in the results of normalizable operation were
not normalized by `--normalize-memrefs` option. This patch normalizes
them.
Differential Revision: https://reviews.llvm.org/D88719
Extended promote buffers to stack pass to support dynamically shaped allocas.
The conversion is limited by the rank of the underlying tensor.
An option is added to the pass to adjust the given rank.
Differential Revision: https://reviews.llvm.org/D91969
Given that OpState already implicit converts to Operator*, this seems reasonable.
The alternative would be to add more functions to OpState which forward to Operation.
Reviewed By: rriddle, ftynse
Differential Revision: https://reviews.llvm.org/D92266
- Address TODO in scf-bufferize: the argument materialization issue is
now fixed and the code is now in Transforms/Bufferize.cpp
- Tighten up finalizing-bufferize to avoid creating invalid IR when
operand types potentially change
- Tidy up the testing of func-bufferize, and move appropriate tests
to a new finalizing-bufferize.mlir
- The new stricter checking in finalizing-bufferize revealed that we
needed a DimOp conversion pattern (found when integrating into npcomp).
Previously, the converion infrastructure was blindly changing the
operand type during finalization, which happened to work due to
DimOp's tensor/memref polymorphism, but is generally not encouraged
(the new pattern is the way to tell the conversion infrastructure that
it is legal to change that type).
The rewrite logic has an optimization to drop a cast operation after
rewriting block arguments if the cast operation has no users. This is
unsafe as there might be a pending rewrite that replaced the cast operation
itself and hence would trigger a second free.
Instead, do not remove the casts and leave it up to a later canonicalization
to do so.
Differential Revision: https://reviews.llvm.org/D92184
This enables partial bufferization that includes function signatures. To test this, this
change also makes the func-bufferize partial and adds a dedicated finalizing-bufferize pass.
Differential Revision: https://reviews.llvm.org/D92032
Block merging in MLIR will incorrectly merge blocks with operations whose values are used outside of that block. This change forbids this behavior and provides a test where it is illegal to perform such a merge.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D91745
These utilities are more closely associated with the buffer
optimizations and buffer deallocation than with the dialect conversion
stuff in Bufferize.h. So move them out.
This makes Bufferize.h very easy to understand and completely focused on
dialect conversion.
Differential Revision: https://reviews.llvm.org/D91563
Refactoring/clean-up step needed to add support for producer-consumer fusion
with multi-store producer loops and, in general, to implement more general
loop fusion strategies in Affine. It introduces the following changes:
- AffineLoopFusion pass now uses loop fusion utilities more broadly to compute
fusion legality (canFuseLoops utility) and perform the fusion transformation
(fuseLoops utility).
- Loop fusion utilities have been extended to deal with AffineLoopFusion
requirements and assumptions while preserving both loop fusion utilities and
AffineLoopFusion current functionality within a unified implementation.
'FusionStrategy' has been introduced for this purpose and, in the future, it
will allow us to have a single loop fusion core implementation that will produce
different fusion outputs depending on the strategy used.
- Improve separation of concerns for legality and profitability analysis:
'isFusionProfitable' no longer filters out illegal scenarios that 'canFuse'
didn't detect, or the other way around. 'canFuse' now takes loop dependences
into account to determine the fusion loop depth (producer-consumer fusion only).
- As a result, maximal fusion now doesn't require any profitability analysis.
- Slices are now computed only once and reused across the legality, profitability
and fusion transformation steps (producer-consumer).
- Refactor some utilities and remove redundant copies of them.
This patch is NFCI and should preserve the existing functionality of both the
AffineLoopFusion pass and the affine fusion utilities.
Reviewed By: andydavis1, bondhugula
Differential Revision: https://reviews.llvm.org/D90798
These includes have been deprecated in favor of BuiltinDialect.h, which contains the definitions of ModuleOp and FuncOp.
Differential Revision: https://reviews.llvm.org/D91572
Some rewriters take more iterations to converge, add a parameter to overwrite
the built-in maximum iteration count.
Fix PR48073.
Reviewed By: mehdi_amini
Differential Revision: https://reviews.llvm.org/D91553
This replaces the old type decomposition logic that was previously mixed
into bufferization, and makes it easily accessible.
This also deletes TestFinalizingBufferize, because after we remove the type
decomposition, it doesn't do anything that is not already provided by
func-bufferize.
Differential Revision: https://reviews.llvm.org/D90899
The index type does not have a bitsize and hence the size of corresponding allocations cannot be computed. Instead, the promotion pass now has an explicit option to specify the size of index.
Differential Revision: https://reviews.llvm.org/D91360
The previous logic for inlining a region A with N blocks into region B
would produce incorrect results on rollback for N greater than 1. This
rollback logic would leave blocks 1..N in region B and only move block 0
to region A.
The new inlining action recording stores the block move actions from N-1
to 0. Now on roll back, block 0 is moved to region A and then 1..N is
appended to the list of blocks in region A.
Differential Revision: https://reviews.llvm.org/D91185
This functionality is superceded by BufferResultsToOutParams pass (see
https://reviews.llvm.org/D90071) for users the require buffers to be
out-params. That pass should be run immediately after all tensors are gone from
the program (before buffer optimizations and deallocation insertion), such as
immediately after a "finalizing" bufferize pass.
The -test-finalizing-bufferize pass now defaults to what used to be the
`allowMemrefFunctionResults=true` flag. and the
finalizing-bufferize-allowed-memref-results.mlir file is moved
to test/Transforms/finalizing-bufferize.mlir.
Differential Revision: https://reviews.llvm.org/D90778
This is the most basic possible finalizing bufferization pass, which I
also think is sufficient for most new use cases. The more concentrated
nature of this pass also greatly clarifies the invariants that it
requires on its input to safely transform the program (see the
pass description in Passes.td).
With this pass, I have now upstreamed practically all of the
bufferizations from npcomp (the exception being std.constant, which can
be upstreamed when std.global_memref lands:
https://llvm.discourse.group/t/rfc-global-variables-in-mlir/2076/16 )
Differential Revision: https://reviews.llvm.org/D90205
This pass allows removing getResultConversionKind from
BufferizeTypeConverter. This pass replaces the AppendToArgumentsList
functionality. As far as I could tell, the only use of this functionlity
is to perform the transformation that is implemented in this pass.
Future patches will remove the getResultConversionKind machinery from
BufferizeTypeConverter, but sending this patch for individual review for
clarity.
Differential Revision: https://reviews.llvm.org/D90071
Often times the legality of inlining can change depending on if the callable is going to be inlined in-place, or cloned. For example, some operations are not allowed to be duplicated and can only be inlined if the original callable will cease to exist afterwards. The new `wouldBeCloned` flag allows for dialects to hook into this when determining legality.
Differential Revision: https://reviews.llvm.org/D90360
In certain situations it isn't legal to inline a call operation, but this isn't something that is possible(at least not easily) to prevent with the current hooks. This revision adds a new hook so that dialects with call operations that shouldn't be inlined can prevent it.
Differential Revision: https://reviews.llvm.org/D90359
This class represents a rewrite pattern list that has been frozen, and thus immutable. This replaces the uses of OwningRewritePatternList in pattern driver related API, such as dialect conversion. When PDL becomes more prevalent, this API will allow for optimizing a set of patterns once without the need to do this per run of a pass.
Differential Revision: https://reviews.llvm.org/D89104
There are several pieces of pattern rewriting infra in IR/ that really shouldn't be there. This revision moves those pieces to a better location such that they are easier to evolve in the future(e.g. with PDL). More concretely this revision does the following:
* Create a Transforms/GreedyPatternRewriteDriver.h and move the apply*andFold methods there.
The definitions for these methods are already in Transforms/ so it doesn't make sense for the declarations to be in IR.
* Create a new lib/Rewrite library and move PatternApplicator there.
This new library will be focused on applying rewrites, and will also include compiling rewrites with PDL.
Differential Revision: https://reviews.llvm.org/D89103
The Pattern class was originally intended to be used for solely matching operations, but that use never materialized. All of the pattern infrastructure uses RewritePattern, and the infrastructure for pure matching(Matchers.h) is implemented inline. This means that this class isn't a useful abstraction at the moment, so this revision refactors it to solely encapsulate the "metadata" of a pattern. The metadata includes the various state describing a pattern; benefit, root operation, etc. The API on PatternApplicator is updated to now operate on `Pattern`s as nothing special from `RewritePattern` is necessary.
This refactoring is also necessary for the upcoming use of PDL patterns alongside C++ rewrite patterns.
Differential Revision: https://reviews.llvm.org/D86258
Added optimization pass to convert heap-based allocs to stack-based allocas in
buffer placement. Added the corresponding test file.
Differential Revision: https://reviews.llvm.org/D89688
Before this change, we would run `maxIterations` if the first iteration changed the op.
After this change, we exit the loop as soon as an iteration hasn't changed the op.
Assuming that we have reached a fixed point when an iteration doesn't change the op, this doesn't affect correctness.
Reviewed By: rriddle
Differential Revision: https://reviews.llvm.org/D89981
A "structural" type conversion is one where the underlying ops are
completely agnostic to the actual types involved and simply need to update
their types. An example of this is scf.if -- the scf.if op and the
corresponding scf.yield ops need to update their types accordingly to the
TypeConverter, but otherwise don't care what type conversions are happening.
To test the structural type conversions, it is convenient to define a
bufferize pass for a dialect, which exercises them nicely.
Differential Revision: https://reviews.llvm.org/D89757
The current BufferPlacement transformation contains several concepts for
hoisting allocations. However, more advanced hoisting techniques should not be
integrated into the BufferPlacement transformation. Hence, this CL refactors the
current BufferPlacement pass into three separate pieces: BufferDeallocation and
BufferAllocation(Loop)Hoisting. Moreover, it extends the hoisting functionality
by allowing to move allocations out of loops.
Differential Revision: https://reviews.llvm.org/D87756
This transforms the symbol lookups to O(1) from O(NM), greatly speeding up both passes. For a large MLIR module this shaved seconds off of the compilation time.
Differential Revision: https://reviews.llvm.org/D89522
This revision contains two optimizations related to symbol checking:
* Optimize SymbolOpInterface to only check for a name attribute if the operation is an optional symbol.
This removes an otherwise unnecessary attribute lookup from a majority of symbols.
* Add a new SymbolTableCollection class to represent a collection of SymbolTables.
This allows for perfoming non-flat symbol lookups in O(1) time by caching SymbolTables for symbol table operations. This class is very useful for algorithms that operate on multiple symbol tables, either recursively or not.
Differential Revision: https://reviews.llvm.org/D89505
The opposite of tensor_to_memref is tensor_load.
- Add some basic tensor_load/tensor_to_memref folding.
- Add source/target materializations to BufferizeTypeConverter.
- Add an example std bufferization pattern/pass that shows how the
materialiations work together (more std bufferization patterns to come
in subsequent commits).
- In coming commits, I'll document how to write composable
bufferization passes/patterns and update the other in-tree
bufferization passes to match this convention. The populate* functions
will of course continue to be exposed for power users.
The naming on tensor_load/tensor_to_memref and their pretty forms are
not very intuitive. I'm open to any suggestions here. One key
observation is that the memref type must always be the one specified in
the pretty form, since the tensor type can be inferred from the memref
type but not vice-versa.
With this, I've been able to replace all my custom bufferization type
converters in npcomp with BufferizeTypeConverter!
Part of the plan discussed in:
https://llvm.discourse.group/t/what-is-the-strategy-for-tensor-memref-conversion-bufferization/1938/17
Differential Revision: https://reviews.llvm.org/D89437
This revision introduces support for buffer allocation for any named linalg op.
To avoid template instantiating many ops, a new ConversionPattern is created to capture the LinalgOp interface.
Some APIs are updated to remain consistent with MLIR style:
`OwningRewritePatternList * -> OwningRewritePatternList &`
`BufferAssignmentTypeConverter * -> BufferAssignmentTypeConverter &`
Differential revision: https://reviews.llvm.org/D89226
This revision also inserts an end-to-end test that lowers tensors to buffers all the way to executable code on CPU.
Differential revision: https://reviews.llvm.org/D88998
The documentation for the NormalizeMemRefs pass and the associated MemRefsNormalizable
traits was confusing and not on the website. This update clarifies the language
around the difference between a MemRef Type, an operation that accesses the value of
MemRef Type, and better documents the limitations of the current implementation.
This patch also includes some basic debugging information for the pass so people
might have a chance of figuring out why it doesn't work on their code.
Differential Revision: https://reviews.llvm.org/D88532
Normalizing memrefs failed when a caller of symbolic use in a function
can not be casted to `CallOp`. This patch avoids the failure by checking
the result of the casting. If the caller can not be casted to `CallOp`,
it is skipped.
Differential Revision: https://reviews.llvm.org/D87746
Add support to tile affine.for ops with parametric sizes (i.e., SSA
values). Currently supports hyper-rectangular loop nests with constant
lower bounds only. Move methods
- moveLoopBody(*)
- getTileableBands(*)
- checkTilingLegality(*)
- tilePerfectlyNested(*)
- constructTiledIndexSetHyperRect(*)
to allow reuse with constant tile size API. Add a test pass -test-affine
-parametric-tile to test parametric tiling.
Differential Revision: https://reviews.llvm.org/D87353
The current BufferPlacement transformation cannot handle loops properly. Buffers
passed via backedges will not be freed automatically introducing memory leaks.
This CL adds support for loops to overcome these limitations.
Differential Revision: https://reviews.llvm.org/D85513
Currently, there is no option to allow for unrolling a loop up to a specific factor (specified by the user).
The code for doing that is there and there are benefits when unrolling is done to smaller loops (smaller than the factor specified).
Reviewed By: bondhugula
Differential Revision: https://reviews.llvm.org/D87111
In this PR, the users of BufferPlacement can configure
BufferAssginmentTypeConverter. These new configurations would give the user more
freedom in the process of converting function signature, and return and call
operation conversions.
These are the new features:
- Accepting callback functions for decomposing types (i.e. 1 to N type
conversion such as unpacking tuple types).
- Defining ResultConversionKind for specifying whether a function result
with a certain type should be appended to the function arguments list or
should be kept as function result. (Usage:
converter.setResultConversionKind<MemRefType>(AppendToArgumentList))
- Accepting callback functions for composing or decomposing values (i.e. N
to 1 and 1 to N value conversion).
Differential Revision: https://reviews.llvm.org/D85133
This reverts commit 94f5d24877 because
of failing the following tests:
MLIR :: Dialect/Linalg/tensors-to-buffers.mlir
MLIR :: Transforms/buffer-placement-preparation-allowed-memref-results.mlir
MLIR :: Transforms/buffer-placement-preparation.mlir
In this PR, the users of BufferPlacement can configure
BufferAssginmentTypeConverter. These new configurations would give the user more
freedom in the process of converting function signature, and return and call
operation conversions.
These are the new features:
- Accepting callback functions for decomposing types (i.e. 1 to N type
conversion such as unpacking tuple types).
- Defining ResultConversionKind for specifying whether a function result
with a certain type should be appended to the function arguments list or
should be kept as function result. (Usage:
converter.setResultConversionKind<MemRefType>(AppendToArgumentList))
- Accepting callback functions for composing or decomposing values (i.e. N
to 1 and 1 to N value conversion).
Differential Revision: https://reviews.llvm.org/D85133
When dealing with dialects that will results in function calls to
external libraries, it is important to be able to handle maps as some
dialects may require mapped data. Before this patch, the detection of
whether normalization can apply or not, operations are compared to an
explicit list of operations (`alloc`, `dealloc`, `return`) or to the
presence of specific operation interfaces (`AffineReadOpInterface`,
`AffineWriteOpInterface`, `AffineDMAStartOp`, or `AffineDMAWaitOp`).
This patch add a trait, `MemRefsNormalizable` to determine if an
operation can have its `memrefs` normalized.
This trait can be used in turn by dialects to assert that such
operations are compatible with normalization of `memrefs` with
nontrivial memory layout specification. An example is given in the
literal tests.
Differential Revision: https://reviews.llvm.org/D86236
This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects are only loaded explicitly
on demand:
- the Parser is lazily loading Dialects in the context as it encounters them
during parsing. This is the only purpose for registering dialects and not load
them in the context.
- Passes are expected to declare the dialects they will create entity from
(Operations, Attributes, or Types), and the PassManager is loading Dialects into
the Context when starting a pipeline.
This changes simplifies the configuration of the registration: a compiler only
need to load the dialect for the IR it will emit, and the optimizer is
self-contained and load the required Dialects. For example in the Toy tutorial,
the compiler only needs to load the Toy dialect in the Context, all the others
(linalg, affine, std, LLVM, ...) are automatically loaded depending on the
optimization pipeline enabled.
To adjust to this change, stop using the existing dialect registration: the
global registry will be removed soon.
1) For passes, you need to override the method:
virtual void getDependentDialects(DialectRegistry ®istry) const {}
and registery on the provided registry any dialect that this pass can produce.
Passes defined in TableGen can provide this list in the dependentDialects list
field.
2) For dialects, on construction you can register dependent dialects using the
provided MLIRContext: `context.getOrLoadDialect<DialectName>()`
This is useful if a dialect may canonicalize or have interfaces involving
another dialect.
3) For loading IR, dialect that can be in the input file must be explicitly
registered with the context. `MlirOptMain()` is taking an explicit registry for
this purpose. See how the standalone-opt.cpp example is setup:
mlir::DialectRegistry registry;
registry.insert<mlir::standalone::StandaloneDialect>();
registry.insert<mlir::StandardOpsDialect>();
Only operations from these two dialects can be in the input file. To include all
of the dialects in MLIR Core, you can populate the registry this way:
mlir::registerAllDialects(registry);
4) For `mlir-translate` callback, as well as frontend, Dialects can be loaded in
the context before emitting the IR: context.getOrLoadDialect<ToyDialect>()
Differential Revision: https://reviews.llvm.org/D85622
This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects are only loaded explicitly
on demand:
- the Parser is lazily loading Dialects in the context as it encounters them
during parsing. This is the only purpose for registering dialects and not load
them in the context.
- Passes are expected to declare the dialects they will create entity from
(Operations, Attributes, or Types), and the PassManager is loading Dialects into
the Context when starting a pipeline.
This changes simplifies the configuration of the registration: a compiler only
need to load the dialect for the IR it will emit, and the optimizer is
self-contained and load the required Dialects. For example in the Toy tutorial,
the compiler only needs to load the Toy dialect in the Context, all the others
(linalg, affine, std, LLVM, ...) are automatically loaded depending on the
optimization pipeline enabled.
To adjust to this change, stop using the existing dialect registration: the
global registry will be removed soon.
1) For passes, you need to override the method:
virtual void getDependentDialects(DialectRegistry ®istry) const {}
and registery on the provided registry any dialect that this pass can produce.
Passes defined in TableGen can provide this list in the dependentDialects list
field.
2) For dialects, on construction you can register dependent dialects using the
provided MLIRContext: `context.getOrLoadDialect<DialectName>()`
This is useful if a dialect may canonicalize or have interfaces involving
another dialect.
3) For loading IR, dialect that can be in the input file must be explicitly
registered with the context. `MlirOptMain()` is taking an explicit registry for
this purpose. See how the standalone-opt.cpp example is setup:
mlir::DialectRegistry registry;
registry.insert<mlir::standalone::StandaloneDialect>();
registry.insert<mlir::StandardOpsDialect>();
Only operations from these two dialects can be in the input file. To include all
of the dialects in MLIR Core, you can populate the registry this way:
mlir::registerAllDialects(registry);
4) For `mlir-translate` callback, as well as frontend, Dialects can be loaded in
the context before emitting the IR: context.getOrLoadDialect<ToyDialect>()
Differential Revision: https://reviews.llvm.org/D85622
This changes the behavior of constructing MLIRContext to no longer load globally
registered dialects on construction. Instead Dialects are only loaded explicitly
on demand:
- the Parser is lazily loading Dialects in the context as it encounters them
during parsing. This is the only purpose for registering dialects and not load
them in the context.
- Passes are expected to declare the dialects they will create entity from
(Operations, Attributes, or Types), and the PassManager is loading Dialects into
the Context when starting a pipeline.
This changes simplifies the configuration of the registration: a compiler only
need to load the dialect for the IR it will emit, and the optimizer is
self-contained and load the required Dialects. For example in the Toy tutorial,
the compiler only needs to load the Toy dialect in the Context, all the others
(linalg, affine, std, LLVM, ...) are automatically loaded depending on the
optimization pipeline enabled.
To adjust to this change, stop using the existing dialect registration: the
global registry will be removed soon.
1) For passes, you need to override the method:
virtual void getDependentDialects(DialectRegistry ®istry) const {}
and registery on the provided registry any dialect that this pass can produce.
Passes defined in TableGen can provide this list in the dependentDialects list
field.
2) For dialects, on construction you can register dependent dialects using the
provided MLIRContext: `context.getOrLoadDialect<DialectName>()`
This is useful if a dialect may canonicalize or have interfaces involving
another dialect.
3) For loading IR, dialect that can be in the input file must be explicitly
registered with the context. `MlirOptMain()` is taking an explicit registry for
this purpose. See how the standalone-opt.cpp example is setup:
mlir::DialectRegistry registry;
mlir::registerDialect<mlir::standalone::StandaloneDialect>();
mlir::registerDialect<mlir::StandardOpsDialect>();
Only operations from these two dialects can be in the input file. To include all
of the dialects in MLIR Core, you can populate the registry this way:
mlir::registerAllDialects(registry);
4) For `mlir-translate` callback, as well as frontend, Dialects can be loaded in
the context before emitting the IR: context.getOrLoadDialect<ToyDialect>()
This changes the behavior of constructing MLIRContext to no longer load globally registered dialects on construction. Instead Dialects are only loaded explicitly on demand:
- the Parser is lazily loading Dialects in the context as it encounters them during parsing. This is the only purpose for registering dialects and not load them in the context.
- Passes are expected to declare the dialects they will create entity from (Operations, Attributes, or Types), and the PassManager is loading Dialects into the Context when starting a pipeline.
This changes simplifies the configuration of the registration: a compiler only need to load the dialect for the IR it will emit, and the optimizer is self-contained and load the required Dialects. For example in the Toy tutorial, the compiler only needs to load the Toy dialect in the Context, all the others (linalg, affine, std, LLVM, ...) are automatically loaded depending on the optimization pipeline enabled.
Differential Revision: https://reviews.llvm.org/D85622
This changes the behavior of constructing MLIRContext to no longer load globally registered dialects on construction. Instead Dialects are only loaded explicitly on demand:
- the Parser is lazily loading Dialects in the context as it encounters them during parsing. This is the only purpose for registering dialects and not load them in the context.
- Passes are expected to declare the dialects they will create entity from (Operations, Attributes, or Types), and the PassManager is loading Dialects into the Context when starting a pipeline.
This changes simplifies the configuration of the registration: a compiler only need to load the dialect for the IR it will emit, and the optimizer is self-contained and load the required Dialects. For example in the Toy tutorial, the compiler only needs to load the Toy dialect in the Context, all the others (linalg, affine, std, LLVM, ...) are automatically loaded depending on the optimization pipeline enabled.
-- This commit handles the returnOp in memref map layout normalization.
-- An initial filter is applied on FuncOps which helps us know which functions can be
a suitable candidate for memref normalization which doesn't lead to invalid IR.
-- Handles memref map normalization for external function assuming the external function
is normalizable.
Differential Revision: https://reviews.llvm.org/D85226
This diff attempts to resolve the TODO in `getOpIndexSet` (formerly
known as `getInstIndexSet`), which states "Add support to handle IfInsts
surronding `op`".
Major changes in this diff:
1. Overload `getIndexSet`. The overloaded version considers both
`AffineForOp` and `AffineIfOp`.
2. The `getInstIndexSet` is updated accordingly: its name is changed to
`getOpIndexSet` and its implementation is based on a new API `getIVs`
instead of `getLoopIVs`.
3. Add `addAffineIfOpDomain` to `FlatAffineConstraints`, which extracts
new constraints from the integer set of `AffineIfOp` and merges it to
the current constraint system.
4. Update how a `Value` is determined as dim or symbol for
`ValuePositionMap` in `buildDimAndSymbolPositionMaps`.
Differential Revision: https://reviews.llvm.org/D84698
Always define a remapping for the memref replacement (`indexRemap`)
with the proper number of inputs, including all the `outerIVs`, so that
the number of inputs and the operands provided for the map don't mismatch.
Reviewed By: bondhugula, andydavis1
Differential Revision: https://reviews.llvm.org/D85177
Remove use of iterator::difference_type to know where to insert a
moved or erased block during undo actions.
Differential Revision: https://reviews.llvm.org/D85066
-- Introduces a pass that normalizes the affine layout maps to the identity layout map both within and across functions by rewriting function arguments and call operands where necessary.
-- Memref normalization is now implemented entirely in the module pass '-normalize-memrefs' and the limited intra-procedural version has been removed from '-simplify-affine-structures'.
-- Run using -normalize-memrefs.
-- Return ops are not handled and would be handled in the subsequent revisions.
Signed-off-by: Abhishek Varma <abhishek.varma@polymagelabs.com>
Differential Revision: https://reviews.llvm.org/D84490
- Add getArgumentTypes() to Region (missed from before)
- Adopt Region argument API in `hasMultiplyAddBody`
- Fix 2 typos in comments
Differential Revision: https://reviews.llvm.org/D84807
The MemRefDataFlow pass does store to load forwarding
only for affine store/loads. This patch updates the pass
to use affine read/write interface which enables vector
forwarding.
Reviewed By: dcaballe, bondhugula, ftynse
Differential Revision: https://reviews.llvm.org/D84302
This revision adds support for much deeper type conversion integration into the conversion process, and enables auto-generating cast operations when necessary. Type conversions are now largely automatically managed by the conversion infra when using a ConversionPattern with a provided TypeConverter. This removes the need for patterns to do type cast wrapping themselves and moves the burden to the infra. This makes it much easier to perform partial lowerings when type conversions are involved, as any lingering type conversions will be automatically resolved/legalized by the conversion infra.
To support this new integration, a few changes have been made to the type materialization API on TypeConverter. Materialization has been split into three separate categories:
* Argument Materialization: This type of materialization is used when converting the type of block arguments when calling `convertRegionTypes`. This is useful for contextually inserting additional conversion operations when converting a block argument type, such as when converting the types of a function signature.
* Source Materialization: This type of materialization is used to convert a legal type of the converter into a non-legal type, generally a source type. This may be called when uses of a non-legal type persist after the conversion process has finished.
* Target Materialization: This type of materialization is used to convert a non-legal, or source, type into a legal, or target, type. This type of materialization is used when applying a pattern on an operation, but the types of the operands have not yet been converted.
Differential Revision: https://reviews.llvm.org/D82831
AllocOp is updated in normalizeMemref(AllocOp allocOp), but, when the
AllocOp has `alignment` attribute, it was ignored and updated AllocOp
does not have `alignment` attribute. This patch fixes it.
Differential Revision: https://reviews.llvm.org/D83656
Some dialects have semantics which is not well represented by common
SSA structures with dominance constraints. This patch allows
operations to declare the 'kind' of their contained regions.
Currently, two kinds are allowed: "SSACFG" and "Graph". The only
difference between them at the moment is that SSACFG regions are
required to have dominance, while Graph regions are not required to
have dominance. The intention is that this Interface would be
generated by ODS for existing operations, although this has not yet
been implemented. Presumably, if someone were interested in code
generation, we might also have a "CFG" dialect, which defines control
flow, but does not require SSA.
The new behavior is mostly identical to the previous behavior, since
registered operations without a RegionKindInterface are assumed to
contain SSACFG regions. However, the behavior has changed for
unregistered operations. Previously, these were checked for
dominance, however the new behavior allows dominance violations, in
order to allow the processing of unregistered dialects with Graph
regions. One implication of this is that regions in unregistered
operations with more than one op are no longer CSE'd (since it
requires dominance info).
I've also reorganized the LangRef documentation to remove assertions
about "sequential execution", "SSA Values", and "Dominance". Instead,
the core IR is simply "ordered" (i.e. totally ordered) and consists of
"Values". I've also clarified some things about how control flow
passes between blocks in an SSACFG region. Control Flow must enter a
region at the entry block and follow terminator operation successors
or be returned to the containing op. Graph regions do not define a
notion of control flow.
see discussion here:
https://llvm.discourse.group/t/rfc-allowing-dialects-to-relax-the-ssa-dominance-condition/833/53
Differential Revision: https://reviews.llvm.org/D80358
Up until now, there has been an implicit agreement that when an operation is marked as
"erased" all uses of that operation's results are guaranteed to be removed during conversion. How this works in practice is that there is either an assert/crash/asan failure/etc. This revision adds support for properly detecting when an erased operation has dangling users, emits and error and fails the conversion.
Differential Revision: https://reviews.llvm.org/D82830
- Arguments of the first block of a region are considered region arguments.
- Add API on Region class to deal with these arguments directly instead of
using the front() block.
- Changed several instances of existing code that can use this API
- Fixes https://bugs.llvm.org/show_bug.cgi?id=46535
Differential Revision: https://reviews.llvm.org/D83599
Summary:
Almost all uses of these iterators, including implicit ones, really
only need the const variant (as it should be). The only exception is
in NewGVN, which changes the order of dominator tree child nodes.
Change-Id: I4b5bd71e32d71b0c67b03d4927d93fe9413726d4
Reviewers: arsenm, RKSimon, mehdi_amini, courbet, rriddle, aartbik
Subscribers: wdng, Prazek, hiraditya, kuhar, rogfer01, rriddle, jpienaar, shauheen, antiagainst, nicolasvasilache, arpith-jacob, mgester, lucyrfox, aartbik, liufengdb, stephenneuendorffer, Joonsoo, grosul1, vkmr, Kayjukh, jurahul, msifontes, cfe-commits, llvm-commits
Tags: #clang, #mlir, #llvm
Differential Revision: https://reviews.llvm.org/D83087
ViewLikeOpInterfaces introduce new aliases that need to be added to the alias
list. This is necessary to place deallocs in the right positions.
Differential Revision: https://reviews.llvm.org/D83044
This pass removes redundant dialect-independent Copy operations in different
situations like the following:
%from = ...
%to = ...
... (no user/alias for %to)
copy(%from, %to)
... (no user/alias for %from)
dealloc %from
use(%to)
Differential Revision: https://reviews.llvm.org/D82757
Summary: The current BufferPlacement implementation does not support
nested region control flow. This CL adds support for nested regions via
the RegionBranchOpInterface and the detection of branch-like
(ReturnLike) terminators inside nested regions.
Differential Revision: https://reviews.llvm.org/D81926
Summary: The patch fixes an off by one error in the method collapseParallelLoops. It ensures the same normalized bound is used for the computation of the division and the remainder.
Reviewers: herhut
Reviewed By: herhut
Subscribers: mehdi_amini, rriddle, jpienaar, shauheen, antiagainst, nicolasvasilache, arpith-jacob, mgester, lucyrfox, aartbik, liufengdb, stephenneuendorffer, Joonsoo, grosul1, Kayjukh, jurahul, msifontes
Tags: #mlir
Differential Revision: https://reviews.llvm.org/D82634
When there is a mix of affine load/store and non-affine operations (e.g. std.load, std.store),
affine-loop-fusion ignores the present of non-affine ops, thus changing the program semantics.
E.g. we have a program of three affine loops operating on the same memref in which one of them uses std.load and std.store, as follows.
```
affine.for
affine.store %1
affine.for
std.load %1
std.store %1
affine.for
affine.load %1
affine.store %1
```
affine-loop-fusion will produce the following result which changed the program semantics:
```
affine.for
std.load %1
std.store %1
affine.for
affine.store %1
affine.load %1
affine.store %1
```
This patch is to fix the above problem by checking non-affine users of the memref that are between the source and destination nodes of interest.
Differential Revision: https://reviews.llvm.org/D82158
This revision removes the TypeConverter parameter passed to the apply* methods, and instead moves the responsibility of region type conversion to patterns. The types of a region can be converted using the 'convertRegionTypes' method, which acts similarly to the existing 'applySignatureConversion'. This method ensures that all blocks within, and including those moved into, a region will have the block argument types converted using the provided converter.
This has the benefit of making more of the legalization logic controlled by patterns, instead of being handled explicitly by the driver. It also opens up the possibility to support multiple type conversions at some point in the future.
This revision also adds a new utility class `FailureOr<T>` that provides a LogicalResult friendly facility for returning a failure or a valid result value.
Differential Revision: https://reviews.llvm.org/D81681
Traditionally patterns have always had the root operation kind hardcoded to a specific operation name. This has worked well for quite some time, but it has certain limitations that make it undesirable. For example, some lowering have the same implementation for many different operations types with a few lowering entire dialects using the same pattern implementation. This problem has led to several "solutions":
a) Provide a template implementation to the user so that they can instantiate it for each operation combination, generally requiring the inclusion of the auto-generated operation definition file.
b) Use a non-templated pattern that allows for providing the name of the operation to match
- No one ever does this, because enumerating operation names can be cumbersome and so this quickly devolves into solution a.
This revision removes the restriction that patterns have a hardcoded root type, and allows for a class patterns that could match "any" operation type. The major downside of root-agnostic patterns is that they make certain pattern analyses more difficult, so it is still very highly encouraged that an operation specific pattern be used whenever possible.
Differential Revision: https://reviews.llvm.org/D82066
This class enables for abstracting more of the details for the rewrite process, and will allow for clients to apply specific cost models to the pattern list. This allows for DialectConversion and the GreedyPatternRewriter to share the same underlying matcher implementation. This also simplifies the plumbing necessary to support dynamic patterns.
Differential Revision: https://reviews.llvm.org/D81985
We previously weren't properly updating the SCC iterator when nodes were removed, leading to asan failures in certain situations. This commit adds a CallGraphSCC class and defers operation deletion until inlining has finished.
Differential Revision: https://reviews.llvm.org/D81984
Similarly to `scf::ForOp`, introduce additional `function_ref` arguments to
`::build` functions of SCF `ParallelOp` and `ReduceOp`. The provided functions
will be called to construct the body of the respective operations while
constructing the operation itself. Exercise them in LoopUtils.
Differential Revision: https://reviews.llvm.org/D81872
It is quite common for the same type to be converted many types throughout the conversion process, and there isn't any good reason why we aren't caching that result. Especially given that we currently use identity conversion to signify legality. This revision also adds a few additional helpers to TypeConverter.
Differential Revision: https://reviews.llvm.org/D81679
allocations cannot be moved freely and can remain in divergent control flow.
The current BufferPlacement pass does not support allocation nodes that carry
additional dependencies (like in the case of dynamic shaped types). These
allocations can often not be moved freely and in turn might remain in divergent
control-flow branches. This requires a different strategy with respect to block
arguments and aliases. This CL adds additinal functionality to support
allocation nodes in divergent control flow while avoiding memory leaks.
Differential Revision: https://reviews.llvm.org/D79850
This patch changes the fusion algorithm so that after fusing two loop nests
we revisit previously visited nodes so that they are considered again for
fusion in the context of the new fused loop nest.
Reviewed By: bondhugula
Differential Revision: https://reviews.llvm.org/D81609
This parameter gives the developers the freedom to choose their desired function
signature conversion for preparing their functions for buffer placement. It is
introduced for BufferAssignmentFuncOpConverter, and also for
BufferAssignmentReturnOpConverter, and BufferAssignmentCallOpConverter to adapt
the return and call operations with the selected function signature conversion.
If the parameter is set, buffer placement won't also deallocate the returned
buffers.
Differential Revision: https://reviews.llvm.org/D81137
This revision adds a helper function to hoist vector.transfer_read /
vector.transfer_write pairs out of immediately enclosing scf::ForOp
iteratively, if the following conditions are true:
1. The 2 ops access the same memref with the same indices.
2. All operands are invariant under the enclosing scf::ForOp.
3. No uses of the memref either dominate the transfer_read or are
dominated by the transfer_write (i.e. no aliasing between the write and
the read across the loop)
To improve hoisting opportunities, call the `moveLoopInvariantCode` helper
function on the candidate loop above which to hoist. Hoisting the transfers
results in scf::ForOp yielding the value that originally transited through
memory.
This revision additionally exposes `moveLoopInvariantCode` as a helper in
LoopUtils.h and updates SliceAnalysis to support return scf::For values and
allow hoisting across multiple scf::ForOps.
Differential Revision: https://reviews.llvm.org/D81199
This patch enables affine loop fusion for loops with affine vector loads
and stores. For that, we only had to use affine memory op interfaces in
LoopFusionUtils.cpp and Utils.cpp so that vector loads and stores are
also taken into account.
Reviewed By: andydavis1, ftynse
Differential Revision: https://reviews.llvm.org/D80971
Dialect conversion infrastructure supports 1->N type conversions by requiring
individual conversions to provide facilities to generate operations
retrofitting N values into 1 of the original type when N > 1. This
functionality can also be used to materialize explicit "cast"-like operations,
but it did not support 1->1 type conversions until now. Modify TypeConverter to
support materialization of cast operations for 1-1 conversions.
This also makes materialization specification more extensible following the
same pattern as type conversions. Instead of overloading a virtual function,
users or subclasses of TypeConversion can now register type-specific
materialization callbacks that will be called in order for the given type.
Differential Revision: https://reviews.llvm.org/D79729
Add BufferAssignmentCallOpConverter as a pattern rewriter for Buffer
Placement. It matches the signature of the caller operation with the callee
after rewriting the callee with FunctionAndBlockSignatureConverter.
Differential Revision: https://reviews.llvm.org/D80785
Buffer placement can now operates on functions that return buffers. These
buffers escape from the deallocation phase of buffer placement.
Differential Revision: https://reviews.llvm.org/D80696
PatternRewriter has support for erasing a Block from its parent region, but
this feature has not been implemented for ConversionPatternRewriter that needs
to keep track of and be able to undo block actions. Introduce support for
undoing block erasure in the ConversionPatternRewriter by marking all the ops
it contains for erasure and by detaching the block from its parent region. The
detached block is stored in the action description and is not actually deleted
until the rewrites are applied.
Differential Revision: https://reviews.llvm.org/D80135
Dialect conversion infrastructure may roll back op creation by erasing the
operations in the reverse order of their creation. While this guarantees uses
of values will be deleted before their definitions, this does not guarantee
that a parent operation will not be deleted before its child. (This may happen
in case of block inlining or if child operations, such as terminators, are
created in the parent's `build` function before the parent itself.) Handle the
parent/child relationship between ops by removing all child ops from the blocks
before erasing the parent. The child ops remain live, detached from a block,
and will be safely destroyed in their turn, which may come later than that of
the parent.
Differential Revision: https://reviews.llvm.org/D80134
This patch introduces interfaces for read and write ops with affine
restrictions. I used `read`/`write` intead of `load`/`store` for the
interfaces so that they can also be implemented by dma ops.
For now, they are only implemented by affine.load, affine.store,
affine.vector_load and affine.vector_store.
For testing purposes, this patch also migrates affine loop fusion and
required analysis to use the new interfaces. No other changes are made
beyond that.
Co-authored-by: Alex Zinenko <zinenko@google.com>
Reviewed By: bondhugula, ftynse
Differential Revision: https://reviews.llvm.org/D79829
Making these two converters more generic. FunctionAndBlockSignatureConverter now
moves only memref results (after type conversion) to the function argument and
keeps other legal function results unchanged. NonVoidToVoidReturnOpConverter is
renamed to NoBufferOperandsReturnOpConverter. It removes only the buffer
operands from the operands of the converted ReturnOp and inserts CopyOps to copy
each buffer to the target function argument.
Differential Revision: https://reviews.llvm.org/D79329
Summary:
This makes a common pattern of
`dyn_cast_or_null<OpTy>(v.getDefiningOp())` more concise.
Differential Revision: https://reviews.llvm.org/D79681
This dialect contains various structured control flow operaitons, not only
loops, reflect this in the name. Drop the Ops suffix for consistency with other
dialects.
Note that this only moves the files and changes the C++ namespace from 'loop'
to 'scf'. The visible IR prefix remains the same and will be updated
separately. The conversions will also be updated separately.
Differential Revision: https://reviews.llvm.org/D79578
The list of destination load ops while evaluating producer-consumer
fusion wasn't being maintained as a set, and as such, duplicate load ops
were being added to it. Although this is harmless correctness-wise, it's
a killer efficiency-wise and it prevents interesting/useful fusions
(including for eg. reshapes into a matmul). The reason the latter
fusions would be missed is that a slice union would be unnecessarily
needed due to the duplicate load ops on a memref added to the 'dst
loads' list. Since slice union is unimplemented for the local var case,
a single destination load op that leads to local vars (like a floordiv /
mod producing fusion), a common case, would not get fused due to an
unnecessary union being tried with itself. (The union would actually be
the same thing but we would bail out.)
Besides the above, this would also significantly speed up fusion as all
the unnecessary slice computations / unions, checks, etc. due to the
duplicates go away.
Differential Revision: https://reviews.llvm.org/D79547
River Riddle