This op is useful for rank specialization of broadcasts. Kernel Generator
needs to generate one kernel for each rank, so if we can minimize the rank
of the broadcast shape, we can support more cases with the same number of
special-cased kernels.
PiperOrigin-RevId: 360137827
This pattern only works for normal convolutions. It does not work for depthwise
convolutions. The Linalg conv ops are defined with static rank, so it only
supports 1d/2d/3d cases, which are the most typical cases.
This also refactors out the same check in lmhlo.conv lowering.
PiperOrigin-RevId: 359503527
- XLA:HLO -> LMHLO conversion drops all token arguments and return values, however
custom calls that users write still expect to get buffer pointers for these token types.
- To be able to support this, add an optional call target argument mapping attribute to
LMHLO custom calls. When this attribute is present, it indicates the number of
arguments and returns that the custom call expects and also indicates which LMHLO
arg() or output() maps to which arg or result number of the custom call.
PiperOrigin-RevId: 358826664
This just blows up everything to ranked (up to 6) and is probably quite slow.
This is sufficient to make kernelgen compile SelectV2.
PiperOrigin-RevId: 358777728
Verification of HLO_BroadcastInDimOp was previously failing or crashing if the
operand had a dynamic shape or was unranked. Update the verification code to
allow the operand to be unranked or have dynamic shape.
PiperOrigin-RevId: 358056793
A shape that contains exactly one element is effectively a scalar. This leads
to a speedup in cases where we have a binary op with one operand that is
effectively a scalar, because we can use the fast path.
PiperOrigin-RevId: 357515552
This is being done by just removing the approximation and lowering to atan2 lib calls later to make the implementation the same as XLA. Note that if the approximation is brought back later, it can be fixed by changing the IR checking `less-than(X, 0)` to `less-than(copysign(X, 1), 0)`
PiperOrigin-RevId: 356253941
- Use a common base class to for AllReduce, AllGather, and AllToAll in the ODS spec.
- Add basic verification for replica groups attribute.
PiperOrigin-RevId: 354969654
In IREE, we use indexed generic op to handle the initial value. However, we
lower it to a generic op that carries an init_tensor here, and leave the handle
of initialization problem to later passes.
PiperOrigin-RevId: 354294807
If mhlo.reshape is not purely collapsing some consecutive operand
dimensions into result dimensions, we will generate two linalg
reshape op for it: the first one collapses all operand dimensions
into one dimension, and the second one expands it to all result
dimensions. For this case, the number of collapsed/expanded dimensions
should be coming strictly from the operand/result. It is different
from the case where we can generate one linalg reshape. For that case,
the reassociation map should have rank equal to the largest among
operand/result shape.
PiperOrigin-RevId: 354293826
Also generate the kernels for all types of casts between signed int and float types.
This requires some adaptations to our build macros so that we can also specify the
output type of a kernel.
PiperOrigin-RevId: 354067727
Allow for relative tolerance in unary kernel tests. In case of the cosh kernels,
this allows to accept an observed difference of 5.6e-8 between the kernel and
the `std::cosh` reference (32829984.568665262 vs. 32829984.568665318) in one of
the test cases.
PiperOrigin-RevId: 351983698
We prototyped the lowering from mhlo.dot to linalg.matmul in IREE. Since Linalg
now supports matmul in tensors world, we can move the lowering logic to tensors
world, and upstream to legalize_to_linalg.cc. The patch lowers the mhlo.dot to
the linalg.matmul/matvec/dot in tensors world.
PiperOrigin-RevId: 351184911
This updates the tests to no longer rely on tensor_store. Once all users of this behavior have adopted, the tensor_store support will be removed.
PiperOrigin-RevId: 348624899
For floating point operations, this uses std.pow.
For integer operations, this lowers to a loop.
This adds a dependency on scf.
PiperOrigin-RevId: 348537232
These are failing for complex types. Complex types require special handling. We have a fallback lowering for these ops so we can disable complex element types for now.
PiperOrigin-RevId: 348205002
Shape inference in case of ops with complex element types need to use the element type of complex as the result element type and not the full operand type.
Before:
"mhlo.abs"(%arg0) : (tensor<4xcomplex<f32>>) -> tensor<4xtensor<4xcomplex<f32>>>
After:
"mhlo.abs"(%arg0) : (tensor<4xcomplex<f32>>) -> tensor<4xf32>
PiperOrigin-RevId: 348123967
It can happen that a lowering for a certain type is not implemented yet.
We should not segfault in such a case, but instead return a failure().
PiperOrigin-RevId: 347801106
- Add this attribute to match the corresponding XLA HLO attribute on convolution
operations.
- A true value indicates a reversal of the corresponding kernel spatial dimension.
- Since XLA builder does not support this attribute, use a custom HLO converted to map
from mlir::mhlo::ConvOp to XLA.
PiperOrigin-RevId: 346891737
Casting negative s32 number to u64 directly will have leading 1s in the representation which is not what we want to get a single u64 out of two s32 seeds. Fixed this by first getting unsigned number of the same bit-width.
PiperOrigin-RevId: 345902167
Two tensors having the same SSA-value isn't sufficient for equality for floating types, as `NaN != NaN`. As written this causes `tf.IsNan` to [miscompile](https://github.com/google/iree/issues/4061).
PiperOrigin-RevId: 345730640
Casting negative s32 number to u64 directly will have leading 1s in the representation which is not what we want to get a single u64 out of two s32 seeds. Fixed this by first getting unsigned number of the same bit-width.
PiperOrigin-RevId: 345618958
Casting negative s32 number to u64 directly will have leading 1s in the representation which is not what we want to get a single u64 out of two s32 seeds. Fixed this by first getting unsigned number of the same bit-width.
PiperOrigin-RevId: 345605910
- Split out enum definitions in hlo dialect into a separate .td file (similar to structs)
and generate enum decl/defs for these enums.
- Also split out the LHLO GPU enums into a separate .td file and generate enum
decl/defs for these enums as well.
- Remove unused dialect from ConvolutionAttributes and generate lhlo_gpu enums.
- Add appropriate namespace for all the enums.
PiperOrigin-RevId: 345277240
Casting negative s32 number to u64 directly will have leading 1s in the representation which is not what we want to get a single u64 out of two s32 seeds. Fixed this by first getting unsigned number of the same bit-width.
PiperOrigin-RevId: 345239817
Casting negative s32 number to u64 directly will have leading 1s in the representation which is not what we want to get a single u64 out of two s32 seeds. Fixed this by first getting unsigned number of the same bit-width.
PiperOrigin-RevId: 345227848
Use constant to generate the correct assertion message. This avoids
confusion when lowering the max rank specialization for debugging.
PiperOrigin-RevId: 344769021
Boolean element values should be fetched as an unsigned integer and not signed integer which would return -1 for true.
Added to a TODO to handle unsigned types correctly as well as we don't seem to be using unsigned types.
PiperOrigin-RevId: 343927564
M_PI and other math constants (used in chlo_legalize_hlo_patterns.td)
are not part of the C++ standard and must be enabled on MSVC
(similar to _GNU_SOURCE adding glibc symbols to posix headers).
PiperOrigin-RevId: 342432987
- Extend MHLO CustomCall to have multiple tensors as results.
- Extend LHLO CustomCall to have multiple memrefs for output operands.
- Fix HLO->LHLO and XLA HLO->LHLO mapping for CustomCall to setup the
operand_segment_sizes attribute correctly.
PiperOrigin-RevId: 342067762
This specific pattern can be replaced with the shape
passed to dynamic_reshape. This is implemented as a
canonicalization on mhlo.dynamic_reshape to fit in
the infrastructure of canonicalization.
PiperOrigin-RevId: 342009365
- Extract code to create result memref's into a ConvertResults function.
- Also fix a bug when using reifyReturnTypes: use correct index for result_shape instead
of always using the first element.
PiperOrigin-RevId: 341852227
The conversion had a bug in computation of strides and sizes args for std.memref_reinterpret_cast. The previous version also relied on linalg::ReshapeOp to do broadcasting when the rank of the output was higher than the rank of the input. Now the broadcasting is entirely done via descriptor modification and linalg::ReshapeOp was replaced with CopyOp.
PiperOrigin-RevId: 341379871
This is to match with HLO semantics and general dimension semantics in MLIR.
Also,
* Define minimal verifier for these ops.
* Add folder for SetDimensionSize op on static shaped dimension.
* Fix assumption of ranked shape in GetDimensionSize op.
PiperOrigin-RevId: 341150923
Lowerings that depended on operations between real and complex types may
not infer the correct intermediate type. Removing these operations as
they are not technically legally generated operations. Updated tests
to validate this.
PiperOrigin-RevId: 341128903
Previously this started at rank 2 after checking for scalars and equal shapes. This resulted in cases such as <1xf32> + <2xf32> being treated as impossible.
PiperOrigin-RevId: 341043965
Imported from GitHub PR https://github.com/tensorflow/tensorflow/pull/44405
Splitting #43857 by top-level directories.
Copybara import of the project:
--
fa5da7d5478649d11321dcac9f867b0a57e4798a by Dmitry Volodin <mr.molkree@gmail.com>:
fix typos in compiler dir
--
4d3c9f047f7ecb8ab299f1bf28a86fd39096eee7 by Dmitry Volodin <mr.molkree@gmail.com>:
fix one test as "atleast" in it comes from Bazel
--
9440ebaaa9fc4a735f7f72f0c8f0de4ec58afbd6 by Dmitry Volodin <mr.molkree@gmail.com>:
a bit more
PiperOrigin-RevId: 340819994
The slice indices must be rank-1 and have the same number of elements of the
rank of the operand. Give reasonable error messages for violations of these
requirements instead of a misleading error message that the types of the
indices don't all match.
PiperOrigin-RevId: 340660822
Imported from GitHub PR https://github.com/tensorflow/tensorflow/pull/44499
The file `sink_constants_to_control_flow.cc` includes the header
`PassDetail.h`, which itself includes `mhlo_passes.h.inc`. The latter is
not guaranteed to be already generated since there was no dependency set
to MLIRMhloPassIncGen.
Copybara import of the project:
--
0ff51ccc88c1ba049eb2e9555afb54079bea39c9 by Marius Brehler <marius.brehler@iml.fraunhofer.de>:
Add missing dep on MLIRMhloPassIncGen target
The file `sink_constants_to_control_flow.cc` includes the header
`PassDetail.h`, which itself includes `mhlo_passes.h.inc`. The latter is
not guaranteed to be already generated since there was no dependency set
to MLIRMhloPassIncGen.
PiperOrigin-RevId: 340485068
Additionally:
- Forward listeners through new if/else op builders.
This corrects an error that led to incomplete legalization of broadcasted op
lowering.
- Use OpConversionPattern to ensure up to date operand values are used.
PiperOrigin-RevId: 339838833
Doesn't support tensors right now, as it's somewhat hairy to support both at
the same time. Since we use a generic lowering the result is messy
and needs a mem2reg pass to eliminate extra load/store/allocas.
PiperOrigin-RevId: 339562971
If unspecified, `compare_type` is FLOAT for float element types, SIGNED for signed element types and UNSIGNED for unsigned element types. compare_type can be TOTALORDER for float element types.
- Added import and export support the attribute.
- Restricted legalization from HLO to TF to the default compare types.
- Updated existing usage of the CompareOp
PiperOrigin-RevId: 339099219
As described in mlir/Transforms/Bufferize.h, patterns that don't need the special methods on a BufferizeTypeConverter should use a regular OpConversionPattern.
PiperOrigin-RevId: 338424819
This https://reviews.llvm.org/D89254 diff introduced implicit matching between same name arguments. Modify usages accordingly.
PiperOrigin-RevId: 338090110
XLA HLO concat does not accept scalars, so fail verification if this occurs. Avoids segfault when accessing an empty output shape.
PiperOrigin-RevId: 337618167
The fusion heuristic identifies the root of a fusion by checking whether an
output of a linalg operation is a function result. It did not consider outputs
flowing through aliasing operations (like casts).
PiperOrigin-RevId: 337479910
- Introduce operations in a new lmhlo_gpu dialect that map to GPU library function calls
in the XLA:GPU backend.
- Add basic unit tests as well.
PiperOrigin-RevId: 337132166
Legalize `atan2` analogously to XLA. `atan2` is first reduced to `atan` on the
interval [-1, 1] and subsequently approximated. This CL also adds e2e tests for
trigonometric approximations.
PiperOrigin-RevId: 334794336
- And add conversion from MHLO CustomCall to LHLO CustomCall
- According to XLA documentation, the called function should not be side effecting,
so marking the argument MemRefs as MemRead.
PiperOrigin-RevId: 334737196
A non globally registered pass should define `getName()` in order to generate correct crash reproducers.
This is something we get "for free" when using the TableGen generated base class.
We should also migrate the other passes to the same mechanism and remove the static
global registration.
PiperOrigin-RevId: 332976907
Benjamin Kramer
Adrian Kuegel
Christian Sigg
Hanhan Wang
Geoffrey Martin-Noble
A. Unique TensorFlower
Rahul Joshi
Richard Uhler
Alexander Belyaev
Tres Popp
Stephan Herhut
Mahesh Ravishankar
Lei Zhang
Tim Shen
Jacques Pienaar
Smit Hinsu
River Riddle
Phoenix Meadowlark
Lucy Fox
Ben Vanik
Robert Suderman
Dmitry Volodin
Marius Brehler
Sean Silva
Nicolas Vasilache
Thomas Joerg
Roman Dzhabarov
Chao Xie
Ahmed S. Taei
Mehdi Amini