This commit relaxes the verifier of
`bufferization.materialize_in_destination` such that mixed
static/dynamic dimensions are allowed for the source and destination
operands. E.g., `tensor<5xf32>` and `tensor<?xf32>` are now compatible,
but it is assumed that the dynamic dimension is `5` at runtime.
This commit fixes#91265.
Extend `bufferization.materialize_in_destination` to support memref
destinations. This op can now be used to indicate that a tensor
computation should materialize in a given buffer (that may have been
allocated by another component/runtime). The op still participates in
"empty tensor elimination".
Example:
```mlir
func.func @test(%out: memref<10xf32>) {
%t = tensor.empty() : tensor<10xf32>
%c = linalg.generic ... outs(%t: tensor<10xf32>) -> tensor<10xf32>
bufferization.materialize_in_destination %c in restrict writable %out : (tensor<10xf32>, memref<10xf32>) -> ()
return
}
```
After "empty tensor elimination", the above IR can bufferize without an
allocation:
```mlir
func.func @test(%out: memref<10xf32>) {
linalg.generic ... outs(%out: memref<10xf32>)
return
}
```
This change also clarifies the meaning of the `restrict` unit attribute
on `bufferization.to_tensor` ops.
This change allows supporting operations for which we don't get precise aliasing information without the need to insert clone operations. E.g., `arith.select`.
Reviewed By: springerm
Differential Revision: https://reviews.llvm.org/D156992
The dealloc operation deallocates each of the given memrefs if there is no alias
to that memref in the list of retained memrefs and the corresponding
condition value is set. This condition can be used to indicate and pass on
ownership of memref values (or in other words, the responsibility of
deallocating that memref). If two memrefs alias each other, only one will be
deallocated to avoid double free situations.
The memrefs to be deallocated must be the originally allocated memrefs,
however, the memrefs to be retained may be arbitrary memrefs.
Returns a list of conditions corresponding to the list of memrefs which
indicates the new ownerships, i.e., if the memref was deallocated the
ownership was dropped (set to 'false') and otherwise will be the same as the
input condition.
Differential Revision: https://reviews.llvm.org/D155467
This operation is a "copy" operation on tensors. It is guaranteed to bufferize to a memcpy. This is different from "tensor.insert_slice", which may fold away.
Note: There is a symmetry between certain tensor, bufferization and memref ops:
* `tensor.empty`, `bufferization.alloc_tensor`, `memref.alloc`
* (none), `bufferization.dealloc_tensor`, `memref.dealloc`
* `tensor.insert_slice`, `bufferization.copy_tensor`, `memref.copy`
Tensor ops can generally canonicalize/fold away, while bufferization dialect ops can be used when a certain side effect is expected to materialize; so they do not fold away.
Differential Revision: https://reviews.llvm.org/D153552
This commit is part of the migration of towards the new STEA syntax/design. In particular, this commit includes the following changes:
* Renaming compiler-internal functions/methods:
* `SparseTensorEncodingAttr::{getDimLevelType => getLvlTypes}`
* `Merger::{getDimLevelType => getLvlType}` (for consistency)
* `sparse_tensor::{getDimLevelType => buildLevelType}` (to help reduce confusion vs actual getter methods)
* Renaming external facets to match:
* the STEA parser and printer
* the C and Python bindings
* PyTACO
However, the actual renaming of the `DimLevelType` itself (along with all the "dlt" names) will be handled in a separate commit.
Reviewed By: aartbik
Differential Revision: https://reviews.llvm.org/D150330
`restrict` is similar to the C++ restrict keyword. Results of `to_tensor` that have the `restrict` attribute are guaranteed to not alias any other `to_tensor` result (after bufferization).
Note: Since `to_memref` ops are not supported by One-Shot Bufferize and all bufferizable ops follow DPS rules (i.e., the buffer of the result is the buffer of an operand or an alias thereof), the buffer of a `to_tensor` op that has the `restrict` attribute is always an entirely "new" buffer that is not aliasing with the future buffer of any tensor value in the entire program. This makes such `to_tensor` ops "safe" from a bufferization perspective; they cannot cause RaW conflicts.
Differential Revision: https://reviews.llvm.org/D144021
MemRef has been accepting a general Attribute as memory space for
a long time. This commits updates bufferization side to catch up,
which allows downstream users to plugin customized symbolic memory
space. This also eliminates quite a few `getMemorySpaceAsInt`
calls, which is deprecated.
Reviewed By: springerm
Differential Revision: https://reviews.llvm.org/D138330
This op used to belong to the sparse dialect, but there are use cases for dense bufferization as well. (E.g., when a tensor alloc is returned from a function and should be deallocated at the call site.) This change moves the op to the bufferization dialect, which now has an `alloc_tensor` and a `dealloc_tensor` op.
Differential Revision: https://reviews.llvm.org/D129985
If `copy` is specified, the newly allocated buffer is initialized with the given contents. Also add an optional `escape` attribute to indicate whether the buffer of the tensor may be returned from the parent block (aka. "escape") after bufferization.
This change is in preparation of connecting One-Shot Bufferize to the sparse compiler.
Differential Revision: https://reviews.llvm.org/D126570
