8 Commits

Author SHA1 Message Date
Matthias Springer
77124386fe [mlir][tensor] Add transform to make tensor.pad loop-independent
Add a transform to make `tensor.pad` and `tensor.empty` ops independent of SCF loop IVs. Such ops can then be hoisted.

E.g.:
```
scf.for %iv = %lb to %ub step %step {
  %high = affine.apply affine_map<(d0)[s0] -> (s0 - d0)> (%i)[%ub]
  %p = tensor.pad %t low[5] high[%high] ...
  ...
}
```
Is transformed to:
```
%high_new = affine.apply affine_map<()[s0, s1] -> (-s0 + s1)> ()[%lb, %ub]
%p_hoistable = tensor.pad %t low[5] high[%high_new]
%dim = tensor.dim %t, %c0
%size = affine.apply affine_map<(d0)[s0, s1] -> (-d0 + s0 + s1 + 5)>(%iv)[%ub, %dim]
%slice = tensor.extract_slice %p_hoistable [0] [%size] [1]
```

Differential Revision: https://reviews.llvm.org/D143910
2023-04-28 11:46:32 +09:00
Matthias Springer
eabb6ccdc8 [mlir][linalg] Replace getUpperBoundForIndex implementation
Use `reifyValueBound` instead, which is more general not hard-coded to a specific list supported ops.

Also add a `closedUB` parameter to the ValueBoundsOpInterface API.

Differential Revision: https://reviews.llvm.org/D146356
2023-04-11 10:32:34 +09:00
Matthias Springer
d8804ecd76 [mlir][Interfaces] Better error handling for addBound
Do not assert if a bound could not be added. This may or may not cause an error. We have the appropriate error handling in `computeBound` (which returns a `LogicalResult`).

Differential Revision: https://reviews.llvm.org/D147769
2023-04-11 10:09:52 +09:00
Kazu Hirata
af6ad7ac9a [mlir] Use DenseMap::contains (NFC) 2023-04-08 15:44:18 -07:00
Matthias Springer
c3f5fd76d6 [mlir][SCF] ValueBoundsOpInterface: Support scf.for results and iter_args
If an `scf.for` loop yields an equal index-typed value or a shaped value with the same dimension sizes (in comparison to the corresponding iter_arg), bounds can be computed for the iter_arg and the OpResult of the `scf.for` op.

Differential Revision: https://reviews.llvm.org/D146306
2023-04-07 11:57:44 +09:00
Matthias Springer
0dc9087ac7 [mlir][Interfaces] ValueBoundsOpInterface: Compute constant bounds
Add a helper function that computes a constant (`int64_t`) bound. The `stopCondition` is optional: If none is provided, the traversal continues until a constant bound could be computed.

Differential Revision: https://reviews.llvm.org/D146296
2023-04-07 11:35:02 +09:00
Matthias Springer
041bc485bf [mlir][Interfaces] ValueBoundsOpInterface: Support LB and UB bounds
This change also adds support for `affine.min` and `affine.max` ops.

Differential Revision: https://reviews.llvm.org/D145787
2023-04-07 10:48:19 +09:00
Matthias Springer
8c885658ed [mlir][Interfaces] Add ValueBoundsOpInterface
Ops can implement this interface to specify lower/upper bounds for their result values and block arguments. Bounds can be specified for:
* Index-type values
* Dimension sizes of shapes values

The bounds are added to a constraint set. Users can query this constraint set to compute bounds wrt. to a user-specified set of values. Only EQ bounds are supported at the moment.

This revision also contains interface implementations for various tensor dialect ops, which illustrates how to implement this interface.

Differential Revision: https://reviews.llvm.org/D145681
2023-04-06 02:57:14 +02:00