Bump OpenROAD: rsz rebuffering estimation-accuracy series#4347
Open
minjukim55 wants to merge 6 commits into
Open
Bump OpenROAD: rsz rebuffering estimation-accuracy series#4347minjukim55 wants to merge 6 commits into
minjukim55 wants to merge 6 commits into
Conversation
Point tools/OpenROAD at the rsz rebuffering-accuracy series (private OR branch secure-rsz-buffering-boost-orfs = ORFS-master OR pin + the series) so ORFS picks up the improved rebuffer delay/slew estimation. OR base = ORFS-master's current OR pin (f5df0085) + only the rsz series, so ORFS CI isolates this change from master drift. Signed-off-by: Minju Kim <mkim@precisioninno.com>
…ries Regenerate rules-base.json for the 13 CI designs whose metrics shifted after the OpenROAD bump (Private ORFS CI build #1). The improved rebuffer delay/slew estimation moved setup-TNS (and routing/area/buffer-count) metrics; goldens regenerated with `make update_rules` (--failing --tighten) from the CI run's metadata.json. Designs: asap7/{aes,ethmac,ethmac_lvt,mock-alu}, gf180/{aes,aes-hybrid}, nangate45/{ariane133,swerv}, sky130hd/{aes,microwatt,riscv32i}, sky130hs/{jpeg,riscv32i}. Note: swerv_wrapper/asap7 also failed CI but is NOT a golden shift -- it SIGSEGVs in PrimaDelayCalc::simulate1 during replace_arith_modules (delay-calc, unrelated to rebuffering); tracked separately. Signed-off-by: Minju Kim <mkim@precisioninno.com>
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Code Review
This pull request updates the tools/OpenROAD submodule to a newer commit and adjusts the baseline design rules (rules-base.json) across various platforms and designs (such as asap7, gf180, nangate45, sky130hd, and sky130hs) to reflect the resulting changes in physical design metrics like timing, area, and wirelength. There are no review comments, so I have no feedback to provide.
Secure CI area metrics exceeded the padded ceilings after the rebuffering estimation-accuracy change. Regenerate with the standard 15% padding (make update_rules): placeopt__design__instance__area 4159 -> 4810 finish__design__instance__area 4236 -> 4886 Signed-off-by: Minju Kim <mkim@precisioninno.com>
Signed-off-by: Minju Kim <mkim@precisioninno.com> # Conflicts: # tools/OpenROAD
Pin tools/OpenROAD at the rsz rebuffering estimation-accuracy series head (The-OpenROAD-Project/OpenROAD#10866, upstream author Martin Poviser). This drops the earlier SparseLU-guard bump so the submodule pin resolves against public OpenROAD for ORFS CI. Signed-off-by: Minju Kim <mkim@precisioninno.com>
openroad-ci
force-pushed
the
secure-rsz-buffering-boost-orfs
branch
from
35b302d to
1cebfb9
Compare
The rsz rebuffering estimation-accuracy series (OpenROAD#10866) shifts asap7/mock-alu finish__timing__setup__tns to -20481.5. Regenerate the rule bound to -20541.5 (period_padding, 20% of the 300ps clock period). Signed-off-by: Minju Kim <mkim@precisioninno.com>
Contributor
Author
|
1 CI crash from OpenSTA - filed a issue The-OpenROAD-Project/OpenROAD#10893 |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Bump
tools/OpenROADto pick up therszrebuffering estimation-accuracyseries (upstream author Martin Povišer): rebuffer's internal delay/slew
estimate is realigned with the model
repair_design/ signoff STA use, plusfixes to the buffered-net cost model (shape-factor location, slew degradation
model, wire-splitting formula, buffered-net delay methodology, junction
placement, buffer-load miscount, long-wire algorithm, zero-RC guards).
The pinned OR commit = ORFS-master's current OR pin + only the rsz series
(cherry-picked onto the same base), so ORFS CI isolates this change from OR
master drift.
Corresponding OpenROAD PR: The-OpenROAD-Project/OpenROAD#10866.
Type of change
Impact
Rebuffering quality only. On a customer design (advanced-node PDK, 5 seeds)
the setup-WS / fmax improved consistently from CTS onward (CTS +8.98%,
GRT +7.08% mean fmax; 5/5 seeds improve at both stages). ORFS flow metric
goldens may shift for affected designs; CI will surface which, and those will
be regenerated in a follow-up commit.
Verification
locally; No-Diffs-In-Ok-Files clean.
Related Issues
Upstream
rszrebuffering-accuracy series; QoR validation tracked internally(private).