#include <ckpttn/FMKWayGainCalc.hpp>
template<typename Gnl>
FMKWayGainCalc class
| Template parameters | |
|---|---|
| Gnl | |
Public types
- using ret_info = std::vector<std::vector<int>>
Constructors, destructors, conversion operators
- FMKWayGainCalc(const Gnl& hyprgraph, std::uint8_t num_parts)
- Constructs a new FMKWayGainCalc object.
Public functions
- auto init(std::span<const std::uint8_t> part) -> int -> auto
- Initializes the FMKWayGainCalc object.
- auto update_move_init() -> void -> auto
- Resets the delta gain vector to 0.
- auto update_move_2pin_net(std::span<const std::uint8_t> part, const MoveInfo<node_t>& move_info) -> node_t -> auto
- Updates the gain for a 2-pin net after a move.
- void init_idx_vec(const node_t& v, const node_t& net)
- Initializes the index vector for a given vertex and net.
- auto update_move_3pin_net(std::span<const std::uint8_t> part, const MoveInfo<node_t>& move_info) -> ret_info -> auto
- Updates the gain for a 3-pin net after a move.
- auto update_move_general_net(std::span<const std::uint8_t> part, const MoveInfo<node_t>& move_info) -> ret_info -> auto
- Updates the gain for a general net after a move.
Public variables
- FMPmr::vector<int> delta_gain_w
- FMPmr::vector<node_t> idx_vec
- bool special_handle_2pin_nets
Function documentation
template<typename Gnl>
FMKWayGainCalc<Gnl>::
Constructs a new FMKWayGainCalc object.
| Parameters | |
|---|---|
| hyprgraph in | The netlist. |
| num_parts in | The number of partitions. |
template<typename Gnl>
auto FMKWayGainCalc<Gnl>::
Initializes the FMKWayGainCalc object.
| Parameters | |
|---|---|
| part in | The partition to initialize. |
| Returns | The total cost after initialization. |
This function resets the total cost, initializes the vertex list and init gain list to 0, and then calls the _init_gain function for each net in the hypergraph.
template<typename Gnl>
auto FMKWayGainCalc<Gnl>::
Resets the delta gain vector to 0.
This function is used to initialize the delta gain vector before updating the gains for a move.
template<typename Gnl>
auto FMKWayGainCalc<Gnl>::
Updates the gain for a 2-pin net after a move.
| Parameters | |
|---|---|
| part in | The current partition. |
| move_info in | The information about the move that was performed. |
| Returns | The updated gain for the 2-pin net. |
This function updates the gain for a 2-pin net after a move has been performed. It takes the current partition and the move information as input, and returns the updated gain for the net.
template<typename Gnl>
void FMKWayGainCalc<Gnl>::
Initializes the index vector for a given vertex and net.
| Parameters | |
|---|---|
| v in | The vertex to initialize the index vector for. |
| net in | The net that the vertex belongs to. |
This function is used to initialize the index vector for a vertex within a net. It is likely an implementation detail of the FMKWayGainCalc class.
template<typename Gnl>
auto FMKWayGainCalc<Gnl>::
Updates the gain for a 3-pin net after a move.
| Parameters | |
|---|---|
| part in | The current partition. |
| move_info in | The information about the move that was performed. |
| Returns | The updated gain for the 3-pin net. |
This function updates the gain for a 3-pin net after a move has been performed. It takes the current partition and the move information as input, and returns the updated gain for the net.
template<typename Gnl>
auto FMKWayGainCalc<Gnl>::
Updates the gain for a general net after a move.
| Parameters | |
|---|---|
| part in | |
| move_info in | |
| Returns | The updated gain for the general net. |
This function updates the gain for a general net (with any number of pins) after a move has been performed. It takes the current partition and the move information as input, and returns the updated gain for the net.