A butterfly parametric dependency defines the dependencies for each element in a big butterfly skeleton. Some algorithms that use butterfly_pd need to know the location of the element in the big butterfly skeleton. An example is DIT FFT algorithm that changes the computation based on the location of the parametric dependency. For the same reason is_position_aware should be specified by user in such cases. In addition, in some cases one would like to read partially from the inputs on a butterfly. That's when the the F (filter) can be used.
More...
|
|
| butterfly_pd (Op op, F f) |
| |
| template<typename Coord , typename Visitor , typename In > |
| void | case_of (Coord const &skeleton_size, Coord const &coord, Visitor &visitor, In &&in_flow) const |
| | If coord is <i, h, ...> it wraps the Op with the following inputs and sends it to the visitor along with the m_op. More...
|
| |
| template<typename Size , typename Coord , typename FlowIndex > |
| std::size_t | consumer_count (Size const &skeleton_size, Coord const &, FlowIndex) const |
| | determines how many of the instances of this parametric dependency will be consuming from a producer with a given coordinate. This is a reverse query as compared to case_of. More...
|
| |
|
Op | get_op () const |
| |
|
Op | get_filter () const |
| |
|
template<typename Coord > |
| int | get_result_id (Coord const &, Coord const &) const |
| |
|
template<typename Coord , typename Span > |
| void | configure (Coord &&, Span &&) |
| |
|
|
static constexpr std::size_t | in_port_size = 1 |
| |
|
static constexpr std::size_t | op_arity = 2 |
| |
template<typename Op, typename F, bool is_position_aware, bool is_reversed = false>
class stapl::skeletons::skeletons_impl::butterfly_pd< Op, F, is_position_aware, is_reversed >
A butterfly parametric dependency defines the dependencies for each element in a big butterfly skeleton. Some algorithms that use butterfly_pd need to know the location of the element in the big butterfly skeleton. An example is DIT FFT algorithm that changes the computation based on the location of the parametric dependency. For the same reason is_position_aware should be specified by user in such cases. In addition, in some cases one would like to read partially from the inputs on a butterfly. That's when the the F (filter) can be used.
- Template Parameters
-
| Op | the workfunction to be applied on each element in the big butterfly skeleton |
| F | the filter to be applied on the input edges |
| is_position_aware | if the computation in the workfunction is dependent on the position of the element |
- See also
- butterfly
◆ case_of()
template<typename Op , typename F , bool is_position_aware, bool is_reversed = false>
template<typename Coord , typename Visitor , typename In >
If coord is <i, h, ...> it wraps the Op with the following inputs and sends it to the visitor along with the m_op.
- in<0>[i]
- in<0>[butterflied_index(i)]
- Parameters
-
| skeleton_size | <n, m, p, ...> where each element is potentially multi-dimensional. |
| coord | <i, j, k, ...> where i < n, j < m, k < p |
| visitor | the information about Op and input is passed so that later this information can be converted to a node in the dependence graph |
| in_flow | a tuple of input flows to consume from |
- See also
- butterflied_index
◆ consumer_count()
template<typename Op , typename F , bool is_position_aware, bool is_reversed = false>
template<typename Size , typename Coord , typename FlowIndex >
determines how many of the instances of this parametric dependency will be consuming from a producer with a given coordinate. This is a reverse query as compared to case_of.
- Template Parameters
-
| FlowIndex | the flow index to which this request is sent |
The documentation for this class was generated from the following file:
1.8.13