This is loading from 2-D memref, in addition to D139655 where we
load from 1-D memref cases.
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D140136
This is now possible with transpose semantics on subgroup MMA
load ops.
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D139655
Add support for loading, computing, and storing `gpu.subgroup` WMMA ops
in transpose mode as well. Update the GPU to NVVM lowerings to support
`transpose` mode and update integration tests as well.
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D139021
This patch mechanically replaces None with std::nullopt where the
compiler would warn if None were deprecated. The intent is to reduce
the amount of manual work required in migrating from Optional to
std::optional.
This is part of an effort to migrate from llvm::Optional to
std::optional:
https://discourse.llvm.org/t/deprecating-llvm-optional-x-hasvalue-getvalue-getvalueor/63716
Enables transposed gpu.subgroup_mma_load_matrix and updates the lowerings in Vector to GPU and GPU to SPIRV. Needed to enable B transpose matmuls lowering to wmma ops.
Taken over from author: stanley-nod <stanley@nod-labs.com>
Reviewed By: ThomasRaoux, antiagainst
Differential Revision: https://reviews.llvm.org/D138770
This patch handles native `mma.sync` sizes and enables issuing `ldmatrix` on
largest possible tiles for matrixB. It requires handling
`vector.extract_strided_slice` from vector to ngpu lowering.
Differential Revision: https://reviews.llvm.org/D135749
The ConvertVectorToGpu pass implementation contained a small private
support library for performing various calculations during conversion
between `vector` and `nvgpu.mma.sync` and `nvgpu.ldmatrix` operations.
The support library is moved under `Dialect/NVGPU/Utils` because the
functions have wider utility. Some documentation comments are added or
improved.
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D135303
This is the first step in replacing interator_type from strings with enums in Vector and Linalg dialect. This change adds IteratorTypeAttr and uses it in ContractionOp.
To avoid breaking all the tests, print/parse code has conversion between string and enum for now.
There is a shared code in StructuredOpsUtils.h that expects iterator types to be strings. To break this dependancy, this change forks helper function `isParallelIterator` and `isReductionIterator` to utils in both dialects and adds `getIteratorTypeNames()` to support backward compatibility with StructuredGenerator.
In the later changes, I plan to add a similar enum attribute to Linalg.
Differential Revision: https://reviews.llvm.org/D133696
The patch introduces the required changes to update the pass declarations and definitions to use the new autogenerated files and allow dropping the old infrastructure.
Reviewed By: mehdi_amini, rriddle
Differential Review: https://reviews.llvm.org/D132838
The patch introduces the required changes to update the pass declarations and definitions to use the new autogenerated files and allow dropping the old infrastructure.
Reviewed By: mehdi_amini, rriddle
Differential Review: https://reviews.llvm.org/D132838
Adds optional attribute to support tensor cores on F32 datatype by lowering to `mma.sync` with TF32 operands. Since, TF32 is not a native datatype in LLVM we are adding `tf32Enabled` as an attribute to allow the IR to be aware of `MmaSyncOp` datatype. Additionally, this patch adds placeholders for nvgpu-to-nvgpu transformation targeting higher precision tf32x3.
For mma.sync on f32 input using tensor cores there are two possibilites:
(a) tf32 (1 `mma.sync` per warp-level matrix-multiply-accumulate)
(b) tf32x3 (3 `mma.sync` per warp-level matrix-multiply-accumulate)
Typically, tf32 tensor core acceleration comes at a cost of accuracy from missing precision bits. While f32 has 23 precision bits, tf32 has only 10 precision bits. tf32x3 aims to recover the precision bits by splitting each operand into two tf32 values and issue three `mma.sync` tensor core operations.
Reviewed By: ThomasRaoux
Differential Revision: https://reviews.llvm.org/D130294
This patch removes the `type` field from `Attribute` along with the
`Attribute::getType` accessor.
Going forward, this means that attributes in MLIR will no longer have
types as a first-class concept. This patch lays the groundwork to
incrementally remove or refactor code that relies on generic attributes
being typed. The immediate impact will be on attributes that rely on
`Attribute` containing a type, such as `IntegerAttr`,
`DenseElementsAttr`, and `ml_program::ExternAttr`, which will now need
to define a type parameter on their storage classes. This will save
memory as all other attribute kinds will no longer contain a type.
Moreover, it will not be possible to generically query the type of an
attribute directly. This patch provides an attribute interface
`TypedAttr` that implements only one method, `getType`, which can be
used to generically query the types of attributes that implement the
interface. This interface can be used to retain the concept of a "typed
attribute". The ODS-generated accessor for a `type` parameter
automatically implements this method.
Next steps will be to refactor the assembly formats of certain operations
that rely on `parseAttribute(type)` and `printAttributeWithoutType` to
remove special handling of type elision until `type` can be removed from
the dialect parsing hook entirely; and incrementally remove uses of
`TypedAttr`.
Reviewed By: lattner, rriddle, jpienaar
Differential Revision: https://reviews.llvm.org/D130092
This one required more changes than ideal due to overlapping generated name
with different return types. Changed getIndexingMaps to getIndexingMapsArray to
move it out of the way/highlight that it returns (more expensively) a
SmallVector and uses the prefixed name for the Attribute.
Differential Revision: https://reviews.llvm.org/D129919
For the conversion to nvgpu `mma.sync` and `ldmatrix` pathways, the code
was missing support for the `i4` data type. While fixing this, another
bug was discoverd that caused the number of ldmatrix tiles calculated for
certain operand types and configurations to be incorrect. This change
fixes both issues and adds additional tests.
Differential Revision: https://reviews.llvm.org/D128074
This aligns the SCF dialect file layout with the majority of the dialects.
Reviewed By: jpienaar
Differential Revision: https://reviews.llvm.org/D128049
This change adds a transformation and pass to the NvGPU dialect that
attempts to optimize reads/writes from a memref representing GPU shared
memory in order to avoid bank conflicts. Given a value representing a
shared memory memref, it traverses all reads/writes within the parent op
and, subject to suitable conditions, rewrites all last dimension index
values such that element locations in the final (col) dimension are
given by
`newColIdx = col % vecSize + perm[row](col/vecSize,row)`
where `perm` is a permutation function indexed by `row` and `vecSize`
is the vector access size in elements (currently assumes 128bit
vectorized accesses, but this can be made a parameter). This specific
transformation can help optimize typical distributed & vectorized accesses
common to loading matrix multiplication operands to/from shared memory.
Differential Revision: https://reviews.llvm.org/D127457
This changes adds the option to lower to NvGpu dialect ops during the
VectorToGPU convsersion pass. Because this transformation reuses
existing VectorToGPU logic, a seperate VectorToNvGpu conversion pass is
not created. The option `use-nvgpu` is added to the VectorToGPU pass.
When this is true, the pass will attempt to convert slices rooted at
`vector.contract` operations into `nvgpu.mma.sync` ops, and
`vector.transfer_read` ops are converted to either `nvgpu.ldmatrix` or
one or more `vector.load` operations. The specific data loaded will
depend on the thread id within a subgroup (warp). These index
calculations depend on data type and shape of the MMA op
according to the downstream PTX specification. The code for supporting
these details is separated into `NvGpuSupport.cpp|h`.
Differential Revision: https://reviews.llvm.org/D122940
This has been on _Both for a couple of weeks. Flip usages in core with
intention to flip flag to _Prefixed in follow up. Needed to add a couple
of helper methods in AffineOps and Linalg to facilitate a pure flag flip
in follow up as some of these classes are used in templates and so
sensitive to Vector dialect changes.
Differential Revision: https://reviews.llvm.org/D122151
These passes generally don't rely on any special aspects of FuncOp, and moving allows
for these passes to be used in many more situations. The passes that obviously weren't
relying on invariants guaranteed by a "function" were updated to be generic pass, the
rest were updated to be FunctionOpinterface InterfacePasses.
The test updates are NFC switching from implicit nesting (-pass -pass2) form to
the -pass-pipeline form (generic passes do not implicitly nest as op-specific passes do).
Differential Revision: https://reviews.llvm.org/D121190
This reduces the dependencies of the MLIRVector target and makes the dialect consistent with other dialects.
Differential Revision: https://reviews.llvm.org/D118533
BlockArguments gained the ability to have locations attached a while ago, but they
have always been optional. This goes against the core tenant of MLIR where location
information is a requirement, so this commit updates the API to require locations.
Fixes#53279
Differential Revision: https://reviews.llvm.org/D117633
The only benefit of FunctionPass is that it filters out function
declarations. This isn't enough to justify carrying it around, as we can
simplify filter out declarations when necessary within the pass. We can
also explore with better scheduling primitives to filter out declarations
at the pipeline level in the future.
The definition of FunctionPass is left intact for now to allow time for downstream
users to migrate.
Differential Revision: https://reviews.llvm.org/D117182
This revision adds 0-d vector support to vector.transfer ops.
In the process, numerous cleanups are applied, in particular around normalizing
and reducing the number of builders.
Reviewed By: ThomasRaoux, springerm
Differential Revision: https://reviews.llvm.org/D114803
Support load with broadcast, elementwise divf op and remove the
hardcoded restriction on the vector size. Picking the right size should
be enfored by user and will fail conversion to llvm/spirv if it is not
supported.
Differential Revision: https://reviews.llvm.org/D113618
This predates the templated variant, and has been simply forwarding
to getSplatValue<Attribute> for some time. Removing this makes the
API a bit more uniform, and also helps prevent users from thinking
it is "cheap".
Precursor: https://reviews.llvm.org/D110200
Removed redundant ops from the standard dialect that were moved to the
`arith` or `math` dialects.
Renamed all instances of operations in the codebase and in tests.
Reviewed By: rriddle, jpienaar
Differential Revision: https://reviews.llvm.org/D110797
