Expand description
tensor dialect.
The tensor dialect is intended to hold core tensor creation and
manipulation ops, which are not strongly associated with any particular
other dialect or domain abstraction. The primary smoke test of this is ops
that make sense for any tensor element type.
We leave it to other dialects to hold the vast swath of possible computations one might want to do on a tensor.
The tensor type is (for better or for worse) used to represent all kinds
of things, and supports an open-ended set of element types. Examples:
- representing large, dense aggregations of primitive types, suitable for high-performance numerical computing.
- representing shapes in the
shapedialect, which consist of small 1D tensors ofindexdata type. - representing aggregations of strings or “variant” types.
- representing large, sparse aggregations of primitive types, suitable for high-performance numerical computing.
Thus, for the tensor dialect, we prefer for now to constrain the
scope as much as possible. The expectation is that at some point
in the future, the tensor dialect’s scope may be broadened through a
careful discussion of the tradeoffs.
The tensor type is actually a builtin type (it lives in the builtin
dialect), and does not live in this dialect.
Structs
- A
bitcastoperation. Tensor bitcast operation. - A builder for a
bitcastoperation. - A
castoperation. Tensor cast operation. - A builder for a
castoperation. - A
collapse_shapeoperation. Operation to produce a tensor with a smaller rank. - A builder for a
collapse_shapeoperation. - A
dimoperation. Dimension index operation. - A builder for a
dimoperation. - An
emptyoperation. Empty tensor operation. - A builder for an
emptyoperation. - An
expand_shapeoperation. Operation to produce a tensor with a higher rank. - A builder for an
expand_shapeoperation. - An
extractoperation. Element extraction operation. - A builder for an
extractoperation. - An
extract_sliceoperation. Extract slice operation. - A builder for an
extract_sliceoperation. - A
from_elementsoperation. Tensor from elements operation.. - A builder for a
from_elementsoperation. - A
gatheroperation. Gather a subset of a tensor at specified indices. - A builder for a
gatheroperation. - A
generateoperation. Creates a dynamically sized tensor from elements. - A builder for a
generateoperation. - An
insertoperation. Element insertion operation. - A builder for an
insertoperation. - An
insert_sliceoperation. Insert_slice operation. - A builder for an
insert_sliceoperation. - A
packoperation. Tensor pack operation. - A builder for a
packoperation. - A
padoperation. Tensor pad operation. - A builder for a
padoperation. - A
parallel_insert_sliceoperation. Specify the tensor slice update of a single thread of a parent ParallelCombiningOpInterface op. . - A builder for a
parallel_insert_sliceoperation. - A
rankoperation. Rank operation. - A builder for a
rankoperation. - A
reshapeoperation. Tensor reshape operation. - A builder for a
reshapeoperation. - A
scatteroperation. Scatter a tensor into a destination tensor at specified indices. - A builder for a
scatteroperation. - A
splatoperation. Tensor splat or broadcast operation. - A builder for a
splatoperation. - An
unpackoperation. Tensor unpack operation. - A builder for an
unpackoperation. - A
yieldoperation. Yield a value from a region. - A builder for a
yieldoperation.
Functions
- Creates a
bitcastoperation. - Creates a
castoperation. - Creates a
collapse_shapeoperation. - Creates a
dimoperation. - Creates an
emptyoperation. - Creates an
expand_shapeoperation. - Creates an
extractoperation. - Creates an
extract_sliceoperation. - Creates a
from_elementsoperation. - Creates a
gatheroperation. - Creates a
generateoperation. - Creates an
insertoperation. - Creates an
insert_sliceoperation. - Creates a
packoperation. - Creates a
padoperation. - Creates a
parallel_insert_sliceoperation. - Creates a
rankoperation. - Creates a
reshapeoperation. - Creates a
scatteroperation. - Creates a
splatoperation. - Creates an
unpackoperation. - Creates a
yieldoperation.