Andy Davis [Wed, 4 Dec 2019 14:53:07 +0000 (06:53 -0800)]
Adds support for unrolling single-result vector operations with iterator type lists and indexing maps to a target vector size.
Adds unit tests for unrolling the vector ContractionOp with different iteration orders.
PiperOrigin-RevId:
283747503
Kazuaki Ishizaki [Wed, 4 Dec 2019 12:58:12 +0000 (04:58 -0800)]
minor spelling tweaks
Closes tensorflow/mlir#250
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/250 from kiszk:spelling_tweaks_201911
50fc04443723190b764e824b6fcd2469fecb56e6
PiperOrigin-RevId:
283733032
Smit Hinsu [Wed, 4 Dec 2019 05:25:02 +0000 (21:25 -0800)]
Avoid variable name conflict in MLIR tutorial code snippet
PiperOrigin-RevId:
283682865
Nicolas Vasilache [Wed, 4 Dec 2019 01:51:34 +0000 (17:51 -0800)]
Refactor dependencies to expose Vector transformations as patterns - NFC
This CL refactors some of the MLIR vector dependencies to allow decoupling VectorOps, vector analysis, vector transformations and vector conversions from each other.
This makes the system more modular and allows extracting VectorToVector into VectorTransforms that do not depend on vector conversions.
This refactoring exhibited a bunch of cyclic library dependencies that have been cleaned up.
PiperOrigin-RevId:
283660308
Lei Zhang [Wed, 4 Dec 2019 00:43:40 +0000 (16:43 -0800)]
[spirv] Add spv.GroupNonUniformBallot
This CL also did the following cleanup:
- Moved the test for spv.SubgroupBallotKHR to its own file
- Wrapped generated canonicalization patterns in anonymous namespace
- Updated header comments in SPVOps.td
PiperOrigin-RevId:
283650091
Mahesh Ravishankar [Wed, 4 Dec 2019 00:05:46 +0000 (16:05 -0800)]
Add a pass to legalize operations before lowering to SPIR-V.
Not all StandardOps can be lowered to SPIR-V. For example, subview op
implementation requires use of pointer bitcasts which is not valid
according to SPIR-V spec (or at least is ambiguous about it). Such ops
need to be removed/transformed before lowering to SPIR-V. The
SPIRVLegalizationPass is added a place where such legalizations can be
added. Current implementation folds the subview ops with load/stores
so that the lowering itself does not have to convert a subview op.
PiperOrigin-RevId:
283642981
Sean Silva [Tue, 3 Dec 2019 22:00:36 +0000 (14:00 -0800)]
Make diagnostic a bit clearer.
This prints out in case of any pass failure. Not just a crash.
PiperOrigin-RevId:
283616719
Andy Davis [Tue, 3 Dec 2019 19:55:09 +0000 (11:55 -0800)]
Add CreateMaskOp to the VectorOps dialect.
PiperOrigin-RevId:
283591888
Sean Silva [Tue, 3 Dec 2019 19:23:48 +0000 (11:23 -0800)]
Verifier: Better error message in case of successor operand mismatch.
In particular, print the successor number in the diagnostic.
PiperOrigin-RevId:
283585084
River Riddle [Tue, 3 Dec 2019 19:13:39 +0000 (11:13 -0800)]
Allow analyses to provide a hook 'isInvalidated' to determine if they are truly invalidated.
The hook has the following form:
* `bool isInvalidated(const AnalysisManager::PreservedAnalyses &)`
Given a preserved analysis set, the analysis returns true if it should truly be
invalidated. This allows for more fine-tuned invalidation in cases where an
analysis wasn't explicitly marked preserved, but may be preserved(or
invalidated) based upon other properties; such as analyses sets.
PiperOrigin-RevId:
283582889
Mahesh Ravishankar [Tue, 3 Dec 2019 18:20:37 +0000 (10:20 -0800)]
Convert MemRefType to a linearized array in SPIR-V lowering.
The SPIR-V lowering used nested !spv.arrays to represented
multi-dimensional arrays, with the hope that in-conjunction with the
layout annotations, the shape and layout of memref can be represented
directly. It is unclear though how portable this representation will
end up being. It will rely on driver compilers implementing complex
index computations faithfully. A more portable approach is to use
linearized arrays to represent memrefs and explicitly instantiate all
the index computation in SPIR-V. This gives added benefit that we can
further optimize the generated code in MLIR before generating the
SPIR-V binary.
PiperOrigin-RevId:
283571167
MLIR Team [Tue, 3 Dec 2019 18:11:40 +0000 (10:11 -0800)]
Add Python bindings for affine expressions with binary operators.
PiperOrigin-RevId:
283569325
MLIR Team [Tue, 3 Dec 2019 17:32:16 +0000 (09:32 -0800)]
Add python bindings for ArrayAttr, AffineMapAttr.
PiperOrigin-RevId:
283561252
Alex Zinenko [Tue, 3 Dec 2019 14:22:31 +0000 (06:22 -0800)]
Fix ViewOp to have at most one offset operand
As described in the documentation, ViewOp is expected to take an optional
dynamic offset followed by a list of dynamic sizes. However, the ViewOp parser
did not include a check for the offset being a single value and accepeted a
list of values instead.
Furthermore, several tests have been exercising the wrong syntax of a ViewOp,
passing multiple values to the dyanmic stride list, which was not caught by the
parser. The trailing values could have been erronously interpreted as dynamic
sizes. This is likely due to resyntaxing of the ViewOp, with the previous
syntax taking the list of sizes before the offset. Update the tests to use the
syntax with the offset preceding the sizes.
Worse, the conversion of ViewOp to the LLVM dialect assumed the wrong order of
operands with offset in the trailing position, and erronously relied on the
permissive parsing that interpreted trailing dynamic offset values as leading
dynamic sizes. Fix the lowering to use the correct order of operands.
PiperOrigin-RevId:
283532506
Diego Caballero [Tue, 3 Dec 2019 14:09:21 +0000 (06:09 -0800)]
AffineLoopFusion: Prevent fusion of multi-out-edge producer loops
tensorflow/mlir#162 introduced a bug that
incorrectly allowed fusion of producer loops with multiple outgoing
edges. This commit fixes that problem. It also introduces a new flag to
disable sibling loop fusion so that we can test producer-consumer fusion
in isolation.
Closes tensorflow/mlir#259
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/259 from dcaballe:dcaballe/fix_multi_out_edge_producer_fusion
578d5661705fd5c56c555832d5e0528df88c5282
PiperOrigin-RevId:
283531105
Stephan Herhut [Tue, 3 Dec 2019 13:11:20 +0000 (05:11 -0800)]
Extend conversion of SubViewOp to llvm to also support cases where size and stride
are constant (i.e., there are no size and stride operands).
We recently added canonicalization that rewrites constant size and stride operands to
SubViewOp into static information in the type, so these patterns now occur during code
generation.
PiperOrigin-RevId:
283524688
Lei Zhang [Tue, 3 Dec 2019 12:49:20 +0000 (04:49 -0800)]
[spirv] Add spv.SubgroupBallotKHROp
PiperOrigin-RevId:
283522284
Alexander Belyaev [Tue, 3 Dec 2019 09:55:18 +0000 (01:55 -0800)]
[Linalg] Update/fix documentation for linalg.indexed_generic.
PiperOrigin-RevId:
283503642
Alex Zinenko [Tue, 3 Dec 2019 08:26:13 +0000 (00:26 -0800)]
Add linkage support to LLVMFuncOp
A recent commit introduced the Linkage attribute to the LLVM dialect and used
it in the Global Op. Also use it in LLVMFuncOp. As per LLVM Language Reference,
if the linkage attribute is omitted, the function is assumed to have external
linkage.
PiperOrigin-RevId:
283493299
Lei Zhang [Mon, 2 Dec 2019 22:21:42 +0000 (14:21 -0800)]
[spirv] NFC: reorder sections in SPIRVBase.td
Put extensions and capabilities at the very beginning because
they will be referenced later by other definitions.
PiperOrigin-RevId:
283416972
Lei Zhang [Mon, 2 Dec 2019 20:26:34 +0000 (12:26 -0800)]
NFC: use `&&` instead of `and`
PiperOrigin-RevId:
283392575
Aart Bik [Mon, 2 Dec 2019 17:56:58 +0000 (09:56 -0800)]
[VectorOps] Add legality rules to broadcast
PiperOrigin-RevId:
283360101
Lei Zhang [Mon, 2 Dec 2019 17:33:24 +0000 (09:33 -0800)]
[ODS] Generate builders taking unwrapped value and defaults for attributes
Existing builders generated by ODS require attributes to be passed
in as mlir::Attribute or its subclasses. This is okay foraggregate-
parameter builders, which is primarily to be used by programmatic
C++ code generation; it is inconvenient for separate-parameter
builders meant to be called in manually written C++ code because
it requires developers to wrap raw values into mlir::Attribute by
themselves.
This CL extends to generate additional builder methods that
take raw values for attributes and handles the wrapping in the
builder implementation. Additionally, if an attribute appears
late in the arguments list and has a default value, the default
value is supplied in the declaration if possible.
PiperOrigin-RevId:
283355919
Mehdi Amini [Mon, 2 Dec 2019 17:17:51 +0000 (09:17 -0800)]
Generate dialect documentations in the doc folder for every dialect
Also add a mlir-doc build target to general all the docs
PiperOrigin-RevId:
283353529
brett koonce [Mon, 2 Dec 2019 17:12:48 +0000 (09:12 -0800)]
docs: minor spelling tweaks
Closes tensorflow/mlir#262
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/262 from brettkoonce:docs-sp
6833fc8aa41edd02d8bc7c3cbb84211cb8b0334c
PiperOrigin-RevId:
283352765
Denis Khalikov [Mon, 2 Dec 2019 15:58:38 +0000 (07:58 -0800)]
Add missing `>` to the description of std.view.
Closes tensorflow/mlir#266
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/266 from denis0x0D:sandbox/miss_char
a5f662e1bf103b5009da67d045ee2fcebf822ab0
PiperOrigin-RevId:
283340486
Lei Zhang [Mon, 2 Dec 2019 15:54:23 +0000 (07:54 -0800)]
[DRR] Introduce `$_` to ignore op argument match
Right now op argument matching in DRR is position-based, meaning we need to
specify N arguments for an op with N ODS-declared argument. This can be annoying
when we don't want to capture all the arguments. `$_` is to remedy the situation.
PiperOrigin-RevId:
283339992
Lei Zhang [Mon, 2 Dec 2019 15:51:27 +0000 (07:51 -0800)]
NFC: Update std.subview op to use AttrSizedOperandSegments
This turns a few manually written helper methods into auto-generated ones.
PiperOrigin-RevId:
283339617
JKIsaacLee [Mon, 2 Dec 2019 15:08:31 +0000 (07:08 -0800)]
add missing '>' in Ch-2
add missing '>' in Ch-2
(tensor<2x3xf64)->(tensor<2x3xf64>)
Closes tensorflow/mlir#283
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/283 from JKIsaacLee:patch-1
b69fe8d51e2a540f7efaded159d35b88778ad159
PiperOrigin-RevId:
283333807
Alexander Belyaev [Mon, 2 Dec 2019 14:30:19 +0000 (06:30 -0800)]
Lower linalg.indexed_generic with libcall to LLVM.
PiperOrigin-RevId:
283328994
Alex Zinenko [Mon, 2 Dec 2019 11:27:38 +0000 (03:27 -0800)]
Introduce Linkage attribute to the LLVM dialect
LLVM IR supports linkage on global objects such as global variables and
functions. Introduce the Linkage attribute into the LLVM dialect, backed by an
integer storage. Use this attribute on LLVM::GlobalOp and make it mandatory.
Implement parsing/printing of the attribute and conversion to LLVM IR.
See tensorflow/mlir#277.
PiperOrigin-RevId:
283309328
Jacques Pienaar [Fri, 29 Nov 2019 18:43:24 +0000 (10:43 -0800)]
mlir-tblgen: Dump input records when no generator is set
Follow LLVM's tblgen convention when no generator is set instead of asserting.
PiperOrigin-RevId:
283073690
Jacques Pienaar [Fri, 29 Nov 2019 18:26:23 +0000 (10:26 -0800)]
Fix redundant convert and use NamedAttributeList as value
* Had leftover call that would result in converting to dictionary attr before
being implicitedly converted back to NamedAttributeList;
* NamedAttributeList is value typed, so don't use const reference;
PiperOrigin-RevId:
283072576
JKIsaacLee [Fri, 29 Nov 2019 16:48:22 +0000 (08:48 -0800)]
Fixed typo in Ch-1 of Toy tutorial
Closes tensorflow/mlir#282
PiperOrigin-RevId:
283064785
Denis Khalikov [Thu, 28 Nov 2019 21:27:26 +0000 (13:27 -0800)]
[spirv] Check that operand of `spirv::CompositeExtractOp` is constant while folding.
Closes tensorflow/mlir#281
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/281 from denis0x0D:sandbox/composite_ex_fold
d02d73658bd1b9eaa515eb4e0aee34bc41d4252b
PiperOrigin-RevId:
282971563
Alex Zinenko [Thu, 28 Nov 2019 19:50:47 +0000 (11:50 -0800)]
Split out FunctionLike printing/parsing into FunctionImplementation.{h,cpp}
Helper utilies for parsing and printing FunctionLike Ops are only relevant to
the implementation of the Op, not its definition. They depend on
OpImplementation.h and increase the inclusion footprint of FunctionSupport.h,
and do so only to provide some utilities in the "impl" namespace. Move them to
a separate files, similarly to OpDefinition/OpImplementation distinction, and
make only Op implementations use them while keeping headers cleaner. NFC.
PiperOrigin-RevId:
282964556
Jose Ignacio Gomez [Thu, 28 Nov 2019 09:59:22 +0000 (01:59 -0800)]
[Linalg] Change attribute n_loop_types to iterator
This addresses issue tensorflow/mlir#270. Linalg is updated to take the same form
of iterator_types than vector contraction.
Closes tensorflow/mlir#280
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/280 from tetuante:PRissue270
d26d88d090d3765d3b9884bfabdd023143f27287
PiperOrigin-RevId:
282905396
Lei Zhang [Wed, 27 Nov 2019 23:55:07 +0000 (15:55 -0800)]
NFC: A few cleanups for SPIRVLowering
Updated comments and used static instead of anonymous namspace
to hide functions to be consistent with the existing codebase.
PiperOrigin-RevId:
282847784
Lei Zhang [Wed, 27 Nov 2019 22:12:32 +0000 (14:12 -0800)]
[spirv] NFC: Add getZero() and getOne() static method to ConstantOp
Getting constant zero or one is very common so it merits a special handy
method on spirv::ConstantOp itself.
PiperOrigin-RevId:
282832572
Lei Zhang [Wed, 27 Nov 2019 21:46:20 +0000 (13:46 -0800)]
[spirv] Add folders for spv.IAdd and spv.IMul
Adding zero and multiplying one can be common when generating code
for index calculation.
This CL also sorted canonicalize.mlir to alphabetical order.
PiperOrigin-RevId:
282828055
Aart Bik [Wed, 27 Nov 2019 19:58:08 +0000 (11:58 -0800)]
Fixed typo in Toy tutorial (second var e -> var f)
PiperOrigin-RevId:
282810649
Nicolas Vasilache [Wed, 27 Nov 2019 15:31:41 +0000 (07:31 -0800)]
Implement Linalg to loops lowering as a pattern
This CL rewrites the linalg ops to loops transformations as patterns that can be targeted directly from Tablegen. Reliance on OpFolder is removed and to cope with it we introduce local folding patterns that are applied greedily.
PiperOrigin-RevId:
282765550
Aart Bik [Wed, 27 Nov 2019 03:52:02 +0000 (19:52 -0800)]
[VectorOps] Refine BroadcastOp in VectorOps dialect
Since second argument is always fully overwritten and
shape is define in "to" clause, it is not needed.
Also renamed "into" to "to" now that arg is dropped.
PiperOrigin-RevId:
282686475
Jacques Pienaar [Tue, 26 Nov 2019 23:30:04 +0000 (15:30 -0800)]
Add create method that takes equivalent of OperationState with NamedAttributeList
This method is close to creating an OperationState first and then unpacking it
but avoids creating the OperationState and takes a NamedAttributeList for
attributes rather than array of NamedAttribute (to enable reusing an already
created NamedAttributeList).
Reuse this new method via create that takes OperationState. I'll update inferReturnTypes in follow up to also take NamedAttributeList and so a build method that uses both inferReturnTypes and create can reuse the same list.
PiperOrigin-RevId:
282651642
Aart Bik [Tue, 26 Nov 2019 22:43:03 +0000 (14:43 -0800)]
[VectorOps] Add a BroadcastOp to the VectorOps dialect
PiperOrigin-RevId:
282643305
David Truby [Tue, 26 Nov 2019 19:37:46 +0000 (11:37 -0800)]
Add OpenMP dialect to the dialect registry
Closes tensorflow/mlir#244
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/244 from DavidTruby:openmp
30e2638ee678188575dd5aeb3f7fa51d93369f5f
PiperOrigin-RevId:
282607397
Mahesh Ravishankar [Tue, 26 Nov 2019 18:10:59 +0000 (10:10 -0800)]
Misc changes to lowering to SPIR-V.
These changes to SPIR-V lowering while adding support for lowering
SUbViewOp, but are not directly related.
- Change the lowering of MemRefType to
!spv.ptr<!spv.struct<!spv.array<...>[offset]>, ..>
This is consistent with the Vulkan spec.
- To enable testing a simple pattern of lowering functions is added to
ConvertStandardToSPIRVPass. This is just used to convert the type of
the arguments of the function. The added function lowering itself is
not meant to be the way functions are eventually lowered into SPIR-V
dialect.
PiperOrigin-RevId:
282589644
Nicolas Vasilache [Tue, 26 Nov 2019 16:47:14 +0000 (08:47 -0800)]
Automated rollback of commit
d60133f89bb08341718bb3132b19bc891f7d4f4d
PiperOrigin-RevId:
282574110
Nicolas Vasilache [Tue, 26 Nov 2019 15:38:33 +0000 (07:38 -0800)]
Relax restriction on affine_apply dim and symbol operands
The affine_apply operation is currently "doubly" affine and conflates two things:
1. it applies an affine map to a list of values of type `index` that are defined as either dim or symbol
2. it restricts (and propagates constraints on) the provenance of dims and symbols to a small subset of ops for which more restrictive polyhedral constraints apply.
Point 2. is related to the ability to form so-called static control parts and is related to dependence analysis and legality of transformations.
Point 1. however is completely independent, the only local implication of dims and symbol for affine_apply is that dims compose while symbols concatenate as well as the structural constraint that dims may not be multiplied.
The properties of composition and canonicalization in affine_apply are more generally useful. This CL relaxes the verifier on affine_apply so it can be used more generally.
The relevant affine.for/if/load/store op verifiers already implement the dim and symbol checking.
See this thread for the related discussion: https://groups.google.com/a/tensorflow.org/g/mlir/c/HkwCbV8D9N0/m/8srUNrX6CAAJ
PiperOrigin-RevId:
282562517
Andrew Anderson [Tue, 26 Nov 2019 01:53:20 +0000 (17:53 -0800)]
Some minor corrections and improvements to LangRef
Some productions in the LangRef were using undefined terminals and non-terminals, which have been added to the EBNF.
The dialect type and dialect attribute productions matched precisely the same structure and have been deduplicated.
The production for ssa-id was ambiguous but the fix is trivial (merging the leading '%') and has been applied.
Closes tensorflow/mlir#265
PiperOrigin-RevId:
282470892
Lei Zhang [Tue, 26 Nov 2019 01:26:16 +0000 (17:26 -0800)]
Add support for AttrSizedOperandSegments/AttrSizedResultSegments
Certain operations can have multiple variadic operands and their size
relationship is not always known statically. For such cases, we need
a per-op-instance specification to divide the operands into logical
groups or segments. This can be modeled by attributes.
This CL introduces C++ trait AttrSizedOperandSegments for operands and
AttrSizedResultSegments for results. The C++ trait just guarantees
such size attribute has the correct type (1D vector) and values
(non-negative), etc. It serves as the basis for ODS sugaring that
with ODS argument declarations we can further verify the number of
elements match the number of ODS-declared operands and we can generate
handy getter methods.
PiperOrigin-RevId:
282467075
Nicolas Vasilache [Mon, 25 Nov 2019 23:56:06 +0000 (15:56 -0800)]
Use vector.InsertStridedSlice in Vector -> Vector unrolling
This CL uses the recently added op to finish the implementation of Vector -> Vector unrolling by replacing the "fake join op" by a series of InsertStridedSliceOp.
Test is updated accordingly
PiperOrigin-RevId:
282451126
Nicolas Vasilache [Mon, 25 Nov 2019 23:36:45 +0000 (15:36 -0800)]
Add a vector.InsertStridedSliceOp
This new op is the counterpart of vector.StridedSliceOp and will be used for in the pattern rewrites for vector unrolling.
PiperOrigin-RevId:
282447414
MLIR Team [Mon, 25 Nov 2019 22:44:20 +0000 (14:44 -0800)]
Allow LLVM::ExtractElementOp to have non-i32 indices.
Also change the text format a bit, so that indices are braced by squares.
PiperOrigin-RevId:
282437095
Ben Vanik [Mon, 25 Nov 2019 22:31:12 +0000 (14:31 -0800)]
Make std.divis and std.diviu support ElementsAttr folding.
PiperOrigin-RevId:
282434465
Mahesh Ravishankar [Mon, 25 Nov 2019 21:19:19 +0000 (13:19 -0800)]
NFC: Actually expose the implementation of createGPUToSPIRVLoweringPass.
A mismatch in the function declaration and function definition,
prevented the implementation of the createGPUToSPIRVLoweringPass from
being exposed.
PiperOrigin-RevId:
282419815
Mahesh Ravishankar [Mon, 25 Nov 2019 20:56:46 +0000 (12:56 -0800)]
Add missing rule to generate SPIR-V ABI Attribute using tblgen to CMake.
PiperOrigin-RevId:
282415592
Andy Davis [Mon, 25 Nov 2019 20:39:30 +0000 (12:39 -0800)]
Update VectorContractionOp to take iterator types and index mapping attributes compatible with linalg ops.
PiperOrigin-RevId:
282412311
Christian Sigg [Mon, 25 Nov 2019 20:30:22 +0000 (12:30 -0800)]
Changing directory shortcut for CPU/GPU runner utils.
Moving cuda-runtime-wrappers.so into subdirectory to match libmlir_runner_utils.so.
Provide parent directory when running test and load .so from subdirectory.
PiperOrigin-RevId:
282410749
Lei Zhang [Mon, 25 Nov 2019 19:29:21 +0000 (11:29 -0800)]
De-duplicate EnumAttr overrides by defining defaults
EnumAttr should provide meaningful defaults so concrete instances
do not need to duplicate the fields.
PiperOrigin-RevId:
282398431
Mahesh Ravishankar [Mon, 25 Nov 2019 18:38:31 +0000 (10:38 -0800)]
Introduce attributes that specify the final ABI for a spirv::ModuleOp.
To simplify the lowering into SPIR-V, while still respecting the ABI
requirements of SPIR-V/Vulkan, split the process into two
1) While lowering a function to SPIR-V (when the function is an entry
point function), allow specifying attributes on arguments and
function itself that describe the ABI of the function.
2) Add a pass that materializes the ABI described in the function.
Two attributes are needed.
1) Attribute on arguments of the entry point function that describe
the descriptor_set, binding, storage class, etc, of the
spv.globalVariable this argument will be replaced by
2) Attribute on function that specifies workgroup size, etc. (for now
only workgroup size).
Add the pass -spirv-lower-abi-attrs to materialize the ABI described
by the attributes.
This change makes the SPIRVBasicTypeConverter class unnecessary and is
removed, further simplifying the SPIR-V lowering path.
PiperOrigin-RevId:
282387587
Mahesh Ravishankar [Mon, 25 Nov 2019 18:10:58 +0000 (10:10 -0800)]
Allow memref_cast from static strides to dynamic strides.
Memref_cast supports cast from static shape to dynamic shape
memrefs. The same should be true for strides as well, i.e a memref
with static strides can be casted to a memref with dynamic strides.
PiperOrigin-RevId:
282381862
Nicolas Vasilache [Mon, 25 Nov 2019 16:46:37 +0000 (08:46 -0800)]
Add vector.insertelement op
This is the counterpart of vector.extractelement op and has the same
limitations at the moment (static I64IntegerArrayAttr to express position).
This restriction will be filterd in the future.
LLVM lowering will be added in a subsequent commit.
PiperOrigin-RevId:
282365760
Alex Zinenko [Mon, 25 Nov 2019 15:59:52 +0000 (07:59 -0800)]
Introduce gpu.func
Introduce a new function-like operation to the GPU dialect to provide a
placeholder for the execution semantic description and to add support for GPU
memory hierarchy. This aligns with the overall goal of the dialect to expose
the common abstraction layer for GPU devices, in particular by providing an
MLIR unit of semantics (i.e. an operation) for memory modeling.
This proposal has been discussed in the mailing list:
https://groups.google.com/a/tensorflow.org/d/msg/mlir/RfXNP7Hklsc/MBNN7KhjAgAJ
As decided, the "convergence" aspect of the execution model will be factored
out into a new discussion and therefore is not included in this commit. This
commit only introduces the operation but does not hook it up with the remaining
flow. The intention is to develop the new flow while keeping the old flow
operational and do the switch in a simple, separately reversible commit.
PiperOrigin-RevId:
282357599
Ben Vanik [Mon, 25 Nov 2019 02:50:54 +0000 (18:50 -0800)]
Support folding of StandardOps with DenseElementsAttr.
PiperOrigin-RevId:
282270243
Lei Zhang [Sat, 23 Nov 2019 14:08:50 +0000 (06:08 -0800)]
NFC: Wire up DRR settings for SPIR-V canonicalization patterns
This CL added necessary files and settings for using DRR to
write SPIR-V canonicalization patterns and also converted the
patterns for spv.Bitcast and spv.LogicalNot.
PiperOrigin-RevId:
282132786
Lei Zhang [Sat, 23 Nov 2019 14:03:40 +0000 (06:03 -0800)]
[spirv] NFC: rename test files and sort tests inside
PiperOrigin-RevId:
282132339
Uday Bondhugula [Sat, 23 Nov 2019 05:47:47 +0000 (21:47 -0800)]
Make isValidSymbol more powerful
The check in isValidSymbol, as far as a DimOp result went, checked if
the dim op was on a top-level memref. However, any alloc'ed, view, or
subview memref would be fine as long as the corresponding dimension of
that memref is either a static one or was in turn created using a valid
symbol in the case of dynamic dimensions.
Reported-by: Jose Gomez
Signed-off-by: Uday Bondhugula <uday@polymagelabs.com>
Closes tensorflow/mlir#252
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/252 from bondhugula:symbol
7b57dc394df9375e651f497231c6e4525a32a662
PiperOrigin-RevId:
282097114
River Riddle [Sat, 23 Nov 2019 01:51:23 +0000 (17:51 -0800)]
NFC: Remove unnecessarily guarded tablegen includes.
Support for including a file multiple times was added in tablegen, removing the need for these extra guards. This is because we already insert c/c++ style header guards within each of the specific .td files.
PiperOrigin-RevId:
282076728
Nicolas Vasilache [Fri, 22 Nov 2019 22:56:36 +0000 (14:56 -0800)]
Fix Windows Build
PiperOrigin-RevId:
282048102
Denis Khalikov [Fri, 22 Nov 2019 19:49:22 +0000 (11:49 -0800)]
[spirv] Add a canonicalizer for `spirv::LogicalNotOp`.
Add a canonicalizer for `spirv::LogicalNotOp`.
Converts:
* spv.LogicalNot(spv.IEqual(...)) -> spv.INotEqual(...)
* spv.LogicalNot(spv.INotEqual(...)) -> spv.IEqual(...)
* spv.LogicalNot(spv.LogicalEqual(...)) -> spv.LogicalNotEqual(...)
* spv.LogicalNot(spv.LogicalNotEqual(...)) -> spv.LogicalEqual(...)
Also moved the test for spv.IMul to arithemtic tests.
Closes tensorflow/mlir#256
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/256 from denis0x0D:sandbox/canon_logical_not
76ab5787b2c777f948c8978db061d99e76453d44
PiperOrigin-RevId:
282012356
Mahesh Ravishankar [Fri, 22 Nov 2019 19:41:29 +0000 (11:41 -0800)]
Add more canonicalizations for SubViewOp.
Depending on which of the offsets, sizes, or strides are constant, the
subview op can be canonicalized in different ways. Add such
canonicalizations, which generalize the existing approach of
canonicalizing subview op only if all of offsets, sizes and shapes are
constants.
PiperOrigin-RevId:
282010703
Lucy Fox [Fri, 22 Nov 2019 18:42:57 +0000 (10:42 -0800)]
Small formatting fix in Tutorial Ch2.
PiperOrigin-RevId:
281998069
Jean-Michel Gorius [Fri, 22 Nov 2019 15:52:02 +0000 (07:52 -0800)]
Unify vector op names with other dialects.
Change vector op names from VectorFooOp to Vector_FooOp and from
vector::VectorFooOp to vector::FooOp.
Closes tensorflow/mlir#257
COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/mlir/pull/257 from Kayjukh:master
dfc3a0e04114885aaec8740d5951d6984d6e1577
PiperOrigin-RevId:
281967461
Lucy Fox [Fri, 22 Nov 2019 01:15:24 +0000 (17:15 -0800)]
Add more detail about locations in Chapter 2 of tutorial.
Resolves issue 241 (tensorflow/mlir#241).
PiperOrigin-RevId:
281867192
Nicolas Vasilache [Thu, 21 Nov 2019 23:19:52 +0000 (15:19 -0800)]
Move Linalg Transforms that are actually Conversions - NFC
PiperOrigin-RevId:
281844602
River Riddle [Thu, 21 Nov 2019 22:34:03 +0000 (14:34 -0800)]
Add support for using the ODS result names as the Asm result names for multi-result operations.
This changes changes the OpDefinitionsGen to automatically add the OpAsmOpInterface for operations with multiple result groups using the provided ODS names. We currently just limit the generation to multi-result ops as most single result operations don't have an interesting name(result/output/etc.). An example is shown below:
// The following operation:
def MyOp : ... {
let results = (outs AnyType:$first, Variadic<AnyType>:$middle, AnyType);
}
// May now be printed as:
%first, %middle:2, %0 = "my.op" ...
PiperOrigin-RevId:
281834156
Christian Sigg [Thu, 21 Nov 2019 19:16:02 +0000 (11:16 -0800)]
Change CUDA tests to use print_memref.
Swap dimensions in all-reduce-op test.
PiperOrigin-RevId:
281791744
River Riddle [Thu, 21 Nov 2019 17:57:43 +0000 (09:57 -0800)]
NFC: Add wrappers around DenseIntElementsAttr/DenseFPElementsAttr::get to avoid the need to cast.
This avoids the need to cast back to the derived type when calling get, i.e. removes the need to do DenseIntElementsAttr::get(...).cast<DenseIntElementsAttr>().
PiperOrigin-RevId:
281772163
Nicolas Vasilache [Thu, 21 Nov 2019 16:46:14 +0000 (08:46 -0800)]
Fix OSS builds - NFC
PiperOrigin-RevId:
281757979
Nicolas Vasilache [Thu, 21 Nov 2019 15:08:41 +0000 (07:08 -0800)]
Drop unused function - NFC
PiperOrigin-RevId:
281741923
Nicolas Vasilache [Thu, 21 Nov 2019 14:29:41 +0000 (06:29 -0800)]
Split Linalg declarative patterns from specific test patterns - NFC
This will make it easier to scale out test patterns and build specific passes that do not interfere with independent testing.
PiperOrigin-RevId:
281736335
Benjamin Kramer [Thu, 21 Nov 2019 14:01:35 +0000 (06:01 -0800)]
Add missing include after LLVM
049043b598ef5b12a5894c0c22db8608be70f517
PiperOrigin-RevId:
281732683
Alex Zinenko [Thu, 21 Nov 2019 12:24:52 +0000 (04:24 -0800)]
Don't force newline before function attributes
Due to legacy reasons, a newline character followed by two spaces was always
inserted before the attributes of the function Op in pretty form. This breaks
formatting when functions are nested in some other operations. Don't print the
newline and just put the attributes on the same line, which is also more
consistent with module Op. Line breaking aware of indentation can be introduced
separately into the parser if deemed useful.
PiperOrigin-RevId:
281721793
Nicolas Vasilache [Thu, 21 Nov 2019 12:24:13 +0000 (04:24 -0800)]
Fix OSS build
Add include of ADT/SmallVector.h.
Fixes tensorflow/mlir#254.
PiperOrigin-RevId:
281721705
Aart Bik [Thu, 21 Nov 2019 05:22:13 +0000 (21:22 -0800)]
Fixed typo in 2-d tiled layout
PiperOrigin-RevId:
281671097
River Riddle [Thu, 21 Nov 2019 03:20:04 +0000 (19:20 -0800)]
NFC: Use Region::getBlocks to fix build failure with drop_begin.
PiperOrigin-RevId:
281656603
River Riddle [Thu, 21 Nov 2019 02:19:38 +0000 (18:19 -0800)]
Add a document detailing operation traits, how to define them, and the current list.
Traits are an important piece of operation definition, but don't really have a good documentation presence at the moment.
PiperOrigin-RevId:
281649025
MLIR Team [Thu, 21 Nov 2019 00:30:14 +0000 (16:30 -0800)]
Correctly parse empty affine maps.
Previously the test case crashes / produces an error.
PiperOrigin-RevId:
281630540
River Riddle [Wed, 20 Nov 2019 23:23:52 +0000 (15:23 -0800)]
Merge DCE and unreachable block elimination into a new utility 'simplifyRegions'.
This moves the different canonicalizations of regions into one place and invokes them in the fixed-point iteration of the canonicalizer.
PiperOrigin-RevId:
281617072
Andy Davis [Wed, 20 Nov 2019 22:43:15 +0000 (14:43 -0800)]
Add VectorContractionOp to the VectorOps dialect.
PiperOrigin-RevId:
281605471
Mahesh Ravishankar [Wed, 20 Nov 2019 21:43:20 +0000 (13:43 -0800)]
Verify subview op result has dynamic shape, when sizes are specified.
If the sizes are specified as arguments to the subview op, then the
shape must be dynamic as well.
PiperOrigin-RevId:
281591608
MLIR Team [Wed, 20 Nov 2019 20:46:54 +0000 (12:46 -0800)]
missing outer index %i in search_body
PiperOrigin-RevId:
281580028
Sean Silva [Wed, 20 Nov 2019 19:48:57 +0000 (11:48 -0800)]
Add multi-level DCE pass.
This is a simple multi-level DCE pass that operates pretty generically on
the IR. Its key feature compared to the existing peephole dead op folding
that happens during canonicalization is being able to delete recursively
dead cycles of the use-def graph, including block arguments.
PiperOrigin-RevId:
281568202
Mahesh Ravishankar [Wed, 20 Nov 2019 19:16:37 +0000 (11:16 -0800)]
Changes to SubViewOp to make it more amenable to canonicalization.
The current SubViewOp specification allows for either all offsets,
shape and stride to be dynamic or all of them to be static. There are
opportunities for more fine-grained canonicalization based on which of
these are static. For example, if the sizes are static, the result
memref is of static shape. The specification of SubViewOp is modified
to allow on or more of offsets, shapes and strides to be statically
specified. The verification is updated to ensure that the result type
of the subview op is consistent with which of these are static and
which are dynamic.
PiperOrigin-RevId:
281560457
Nicolas Vasilache [Wed, 20 Nov 2019 18:54:45 +0000 (10:54 -0800)]
Implement unrolling of vector ops to finer-grained vector ops as a pattern.
This CL uses the pattern rewrite infrastructure to implement a simple VectorOps -> VectorOps legalization strategy to unroll coarse-grained vector operations into finer grained ones.
The transformation is written using local pattern rewrites to allow composition with other rewrites. It proceeds by iteratively introducing fake cast ops and cleaning canonicalizing or lowering them away where appropriate.
This is an example of writing transformations as compositions of local pattern rewrites that should enable us to make them significantly more declarative.
PiperOrigin-RevId:
281555100
River Riddle [Wed, 20 Nov 2019 18:19:01 +0000 (10:19 -0800)]
Add a new OpAsmOpInterface to allow for ops to directly hook into the AsmPrinter.
This interface provides more fine-grained hooks into the AsmPrinter than the dialect interface, allowing for operations to define the asm name to use for results directly on the operations themselves. The hook is also expanded to enable defining named result "groups". Get a special name to use when printing the results of this operation.
The given callback is invoked with a specific result value that starts a
result "pack", and the name to give this result pack. To signal that a
result pack should use the default naming scheme, a None can be passed
in instead of the name.
For example, if you have an operation that has four results and you want
to split these into three distinct groups you could do the following:
setNameFn(getResult(0), "first_result");
setNameFn(getResult(1), "middle_results");
setNameFn(getResult(3), ""); // use the default numbering.
This would print the operation as follows:
%first_result, %middle_results:2, %0 = "my.op" ...
PiperOrigin-RevId:
281546873
Nicolas Vasilache [Wed, 20 Nov 2019 18:06:35 +0000 (10:06 -0800)]
Add StridedMemRef<>::operator[] - NFC
This operator is used for internal debugging purposes.
PiperOrigin-RevId:
281544152
Alexander Belyaev [Wed, 20 Nov 2019 14:20:08 +0000 (06:20 -0800)]
Fix the comment to Region block iterators.
PiperOrigin-RevId:
281506693
Alexander Belyaev [Wed, 20 Nov 2019 13:37:49 +0000 (05:37 -0800)]
Fix 'the the' typo.
PiperOrigin-RevId:
281501234