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

..

Utils

Use APInt::count{l,r}_{zero,one} (NFC)

2023-02-19 22:04:47 -08:00

CallInterfaces.cpp

[mlir] Update CallInterfaceCallable to use the new casting infra.

2022-10-13 20:33:24 -05:00

CastInterfaces.cpp

…

CMakeLists.txt

[mlir] Add a generic mem2reg implementation.

2023-04-27 06:00:48 +00:00

ControlFlowInterfaces.cpp

[mlir] NFC: work around gcc-aarch64 v8.3 compilation issue in getRegionBranchSuccessorOperands implementation.

2022-12-29 12:03:32 +01:00

CopyOpInterface.cpp

…

DataLayoutInterfaces.cpp

[MLIR][LLVM] Realign allocas to avoid dynamic realignment in inliner.

2023-04-19 18:39:54 +00:00

DerivedAttributeOpInterface.cpp

…

DestinationStyleOpInterface.cpp

[mlir][linalg] Allow some fusion on mixed generics

2022-11-29 15:35:02 +01:00

InferIntRangeInterface.cpp

[mlir] Use std::optional instead of llvm::Optional (NFC)

2023-01-14 01:25:58 -08:00

InferTypeOpInterface.cpp

[mlir][NFC] reifyResultShapes: Add extra error checking

2023-03-10 11:37:54 +01:00

LoopLikeInterface.cpp

[mlir] Add function for checking if a block is inside a loop

2023-02-10 16:14:17 +00:00

Mem2RegInterfaces.cpp

[mlir] Add a generic mem2reg implementation.

2023-04-27 06:00:48 +00:00

ParallelCombiningOpInterface.cpp

…

RuntimeVerifiableOpInterface.cpp

[mlir] Add RuntimeVerifiableOpInterface and transform

2022-12-21 10:57:14 +01:00

ShapedOpInterfaces.cpp

[mlir][interfaces] Add ShapedDimOpInterface

2022-10-03 13:58:52 +09:00

SideEffectInterfaces.cpp

Allow inline of all pure ops from the LLVM dialect.

2022-12-14 15:18:37 +00:00

TilingInterface.cpp

…

ValueBoundsOpInterface.cpp

[mlir][tensor] Add transform to make tensor.pad loop-independent

2023-04-28 11:46:32 +09:00

VectorInterfaces.cpp

Reland "[mlir][Vector] Re-define masking semantics in vector.transfer ops""

2022-11-29 03:36:54 +00:00

ViewLikeInterface.cpp

[mlir] Add loop bounds to scf.foreach_thread.

2023-02-17 08:57:52 +01:00