cost view, cost aware resynthesis problem, cost generic resub

[codecov-commenter]

Codecov Report

Merging #554 (8b954af) into master (826a2f0) will increase coverage by 0.03%. The diff coverage is 86.95%.

:exclamation: Current head 8b954af differs from pull request most recent head ad8bb59. Consider uploading reports for the commit ad8bb59 to get more accurate results

@@            Coverage Diff             @@
##           master     #554      +/-   ##
==========================================
+ Coverage   82.77%   82.81%   +0.03%     
==========================================
  Files         148      155       +7     
  Lines       18535    19151     +616     
==========================================
+ Hits        15343    15860     +517     
- Misses       3192     3291      +99     

Impacted Files	Coverage Δ
include/mockturtle/views/cost_view.hpp	84.21% <84.21%> (ø)
...e/algorithms/experimental/boolean_optimization.hpp	62.26% <100.00%> (ø)
...lude/mockturtle/utils/recursive_cost_functions.hpp	100.00% <100.00%> (ø)
include/mockturtle/properties/aqfpcost.hpp	87.50% <0.00%> (-12.50%)	:arrow_down:
...ude/mockturtle/algorithms/aqfp/detail/dag_cost.hpp	98.61% <0.00%> (-1.39%)	:arrow_down:
...de/mockturtle/algorithms/aqfp/buffer_insertion.hpp	78.53% <0.00%> (-0.04%)	:arrow_down:
include/mockturtle/algorithms/aqfp/detail/dag.hpp	100.00% <0.00%> (ø)
...ude/mockturtle/algorithms/aqfp/detail/dag_util.hpp	100.00% <0.00%> (ø)
.../mockturtle/algorithms/aqfp/detail/partial_dag.hpp	100.00% <0.00%> (ø)
...ude/mockturtle/algorithms/aqfp/aqfp_node_resyn.hpp	56.84% <0.00%> (ø)
... and 6 more

Help us with your feedback. Take ten seconds to tell us how you rate us.

[Nozidoali]

A cost-generic resubstitution algorithm with a customizable cost function.

• Defined a new class of cost functions called "recursive cost function". Users can describe the optimization target by defining a new cost function.
• Added a "cost view" of the network that is able to calculate the cost of both the entire network and the local window. It can be used in a resubstitution algorithm to evaluate the quality of a resubstitute candidate.
• Implemented a "cost-aware resynthesis solver" that solves the resynthesis problem under functionality constraints and returns the dependency circuit with the lowest cost. It is an extension of the previous "xag_resyn" solver.
• Developed a "cost-generic-resubstitution" optimization script. When calling the resubstitution algorithm, the user can specify the cost function to reduce and the script will optimize the network accordingly. Currently, only networks with base type xag_network are supported.