restore dpl: remove use of disallow_one_site_gaps #6718
Conversation
…l-one-site" This reverts commit 1b67316. Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
eder-matheus
requested review from
maliberty and
gadfort
and removed request for
maliberty
| @@ -305,8 +304,7 @@ void Grid::visitCellBoundaryPixels( | |||
| } | |||
| } | |||
| if (!have_obstructions) { | |||
| const auto grid_rect | |||
| = padded ? gridCoveringPadded(&cell) : gridCovering(&cell); | |||
| const auto grid_rect = gridCovering(&cell); | |||
There was a problem hiding this comment.
@maliberty @gadfort Here's the change to remove the padding from 1-site gap check.
There was a problem hiding this comment.
does it make sense to add a test for this behavior?
There was a problem hiding this comment.
I'll try to reproduce it with a public PDK.
There was a problem hiding this comment.
I would imagine, a DEF with 2 inverters spaced 3 sites apart with a padding of 1 would be sufficient and check_placement (just need to remove all the x1 cells from the platforms)
|
clang-tidy review says "All clean, LGTM! 👍" |
|
@gadfort This fix was not enough for all designs. I'll go back to it in the weekend. |
switched to draft until you have this ready to go |
Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
|
@maliberty I've updated DPO by adding a new struct to represent the blockages. With this struct, we can choose between the original and the padded position of a blockage. It worked in the design I was testing, and I started a secure-ci now. |
Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
eder-matheus
force-pushed
the
restore_dpl
branch
from
b8d386f to
ee74f25
Compare
Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
Signed-off-by: Eder Monteiro <[email protected]>
|
@maliberty I believe this is finally ready to merge. The last issue I had in the secure-ci is finally fixed with the latest commits. |
No description provided.