|
MLIR 22.0.0git
|
Converts narrow integer or float types that are not supported by the target hardware to wider types. More...
#include "mlir/Dialect/Arith/Transforms/NarrowTypeEmulationConverter.h"
Public Member Functions | |
| NarrowTypeEmulationConverter (unsigned targetBitwidth) | |
| unsigned | getLoadStoreBitwidth () const |
Converts narrow integer or float types that are not supported by the target hardware to wider types.
Currently, we only handle power-of-two integer types and convert them to wider integers that are equal or larger than 8 bits.
Definition at line 20 of file NarrowTypeEmulationConverter.h.
|
explicit |
Definition at line 27 of file EmulateNarrowType.cpp.
References mlir::Type::getContext().
|
inline |
Definition at line 24 of file NarrowTypeEmulationConverter.h.
Referenced by mlir::memref::populateMemRefNarrowTypeEmulationConversions().