dpl: support LEF58_CELLEDGESPACINGTABLE #6595
Conversation
Signed-off-by: osamahammad21 <[email protected]>
|
There will be a followup PR for optimizing performance of the checkEdgeSpacing function. This is just an initial attempt. The followup PR would include the following:
|
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
I don't think this is necessary. The search on the pixel grid is already fairly efficient. The extra memory cost versus performance feels heavy. |
Signed-off-by: osamahammad21 <[email protected]>
Signed-off-by: osamahammad21 <[email protected]>
Signed-off-by: osamahammad21 <[email protected]>
|
clang-tidy review says "All clean, LGTM! 👍" |
| : spc(spc_in), is_exact(is_exact_in), except_abutted(except_abutted_in) | ||
| { | ||
| } | ||
| bool operator<(const EdgeSpacingEntry& rhs) const { return spc < rhs.spc; } |
There was a problem hiding this comment.
I only use this for getting the maximum spacing value. It doesn't matter if a < b or b < a if they both end up with the same spacing value.
| @@ -932,6 +933,113 @@ bool Opendp::checkRegionOverlap(const Cell* cell, | |||
| // be fully contained by the cell's bounding box. | |||
| return result.empty(); | |||
| } | |||
| namespace cell_edges { | |||
| Rect transformEdgeRect(const Rect& edge_rect, | |||
There was a problem hiding this comment.
Why can't you just apply the instance transform directly to the edge?
There was a problem hiding this comment.
because x and y are the instance location (lower left corner point of the instance bounding box) not the origin point. applying the inst transform directly is not correct.
Signed-off-by: osamahammad21 <[email protected]>
|
clang-tidy review says "All clean, LGTM! 👍" |
Linked to #3899