grt/cugr: improve via demand calculation by eder-matheus · Pull Request #10802 · The-OpenROAD-Project/OpenROAD

eder-matheus · 2026-07-03T03:04:16Z

Summary

CUGR previously modeled a via's global-routing demand as a min-area-derived stub length inflated by a flat via_multiplier = 2.0 — a proxy with no perpendicular (track-blocking) term. This replaces it with a model built from the via's real enclosure geometry read from the tech:

demand = (enclosure_along / gcell) * ((enclosure_perp + 2*spacing) / pitch)

i.e. "how much of a track's length the pad spans" times "how many tracks it blocks across." This removes the magic via_multiplier on the geometry path and better approximates the resource a via consumes at detailed routing. The via
geometry is extracted once in Design; GridGraph consumes the digested per-layer demand.

Also included:

Via selection ranks candidates (OR_DEFAULT, then LEF-default, then fewest cuts, then smallest enclosure) instead of relying on ODB declaration order.
Spacing falls back to the parallel-run-length value when the plain spacing rule is 0 (advanced techs), so the perpendicular term is not silently dropped.
Warns (GRT-173) when a layer pair has no via and the legacy min-area proxy is used as a fallback.

Type of Change

New feature
Refactoring

Impact

Affects CUGR only (global_route -use_cugr); the default FastRoute path is unchanged. Via-demand magnitudes change per layer, which shifts CUGR routing, congestion, and guides: neutral on uncongested designs, improved on realistically congested designs (lower reported overflow and wirelength, e.g. aes-class designs 58->27 and 101->42 overflow with shorter wirelength), and higher reported congestion only on an extreme 90%-capacity stress test (the model correctly charging fat upper-metal vias under starvation; RRR still converges). Resistance-aware timing is preserved. A new warning (GRT-173) is emitted for techs with no via for a routing-layer pair. All CUGR test goldens were regenerated.

Verification

I have verified that the local build succeeds (./etc/Build.sh).
I have run the relevant tests and they pass.
My code follows the repository's formatting guidelines.
I have included tests to prevent regressions.
I have signed my commits (DCO).

Related Issues

N/A

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

…op dead field Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

…er default Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

…king Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

eder-matheus · 2026-07-03T03:05:01Z

@codex review

gemini-code-assist · 2026-07-03T03:05:34Z

Warning

Gemini encountered an error creating the review. You can try again by commenting /gemini review.

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: ba01782fdc

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-07-03T03:14:17Z

+          lo_dx = r.dx();
+          lo_dy = r.dy();
+        } else if (box->getTechLayer() == upper_tl) {
+          up_dx = r.dx();
+          up_dy = r.dy();


Accumulate via metal extents across all boxes

When the selected tech via has more than one RECT on a routing layer, this loop overwrites lo_dx/lo_dy or up_dx/up_dy with whichever dbBox is visited last. That makes the new via-demand model charge a single rectangle rather than the via's full routing-layer footprint, so compound/default vias can have congestion and cost under-reported in a way that depends on box order. Please union/accumulate the routing-layer bbox before calling viaDemandLength.

Useful? React with 👍 / 👎.

We would support such a non-standard via.

maliberty · 2026-07-03T12:19:54Z

(perp + 2 * spacing) / pitch

Do you want to use ceil? You can't really partially block a track.

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

eder-matheus · 2026-07-03T22:08:14Z

@codex review

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 63bd0c554a

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-07-03T22:14:31Z

+    const std::tuple<bool, bool, int, int64_t> key{
+        !or_default, !via->isDefault(), cuts, enc_area};


Prefer single-cut vias before LEF defaults

When a layer has both a LEF DEFAULT multi-cut via and any single-cut via, this tuple selects the default multi-cut via because !via->isDefault() is ranked before cuts. I checked DRT's initDefaultVias in src/drt/src/io/io_parser_helper.cpp:95-106: it groups candidates by cut count and always chooses from the 1-cut group when available, with getViaRawPriority only ordering inside that group. In that scenario CUGR now charges congestion for a different footprint than detailed routing will use, which can push routes away from vias DRT would actually place.

Useful? React with 👍 / 👎.

eder-matheus added 7 commits July 2, 2026 17:48

grt/cugr: model via demand from real via enclosure geometry

37bf65b

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: update CUGR test goldens for geometry-based via demand

836430a

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: add via_geometry_cugr regression test

512a2b6

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: fall back to drt-style via selection when no default via

87eba10

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: harden via-demand spacing fallback, warn on missing via, dr…

48cf181

…op dead field Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: rank candidate vias by priority instead of arbitrary db-ord…

d5fd3d3

…er default Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: add via_priority_cugr test for db-order-independent via ran…

ba01782

…king Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

github-actions Bot added the size/M label Jul 3, 2026

chatgpt-codex-connector Bot reviewed Jul 3, 2026

View reviewed changes

eder-matheus added 5 commits July 3, 2026 11:26

grt/cugr: union all via boxes per layer for the demand footprint

1d75d37

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: add via_multirect_cugr test for compound-via footprint union

c8ec166

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: include Layers.h directly for MetalLayer (include-cleaner)

6ac1f5e

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: ceil perpendicular track blockage in via demand

3209d3f

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

grt/cugr: update CUGR goldens for integer via track blockage

63bd0c5

Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>

chatgpt-codex-connector Bot reviewed Jul 3, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

grt/cugr: improve via demand calculation#10802

grt/cugr: improve via demand calculation#10802
eder-matheus wants to merge 12 commits into
The-OpenROAD-Project:masterfrom
eder-matheus:grt_cugr_vias

eder-matheus commented Jul 3, 2026

Uh oh!

eder-matheus commented Jul 3, 2026

Uh oh!

gemini-code-assist Bot commented Jul 3, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Jul 3, 2026

Uh oh!

maliberty Jul 3, 2026

Uh oh!

maliberty commented Jul 3, 2026

Uh oh!

eder-matheus commented Jul 3, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot Jul 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

		const std::tuple<bool, bool, int, int64_t> key{
		!or_default, !via->isDefault(), cuts, enc_area};

Uh oh!

Conversation

eder-matheus commented Jul 3, 2026

Summary

Type of Change

Impact

Verification

Related Issues

Uh oh!

eder-matheus commented Jul 3, 2026

Uh oh!

gemini-code-assist Bot commented Jul 3, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Jul 3, 2026

Choose a reason for hiding this comment

Uh oh!

maliberty Jul 3, 2026

Choose a reason for hiding this comment

Uh oh!

maliberty commented Jul 3, 2026

Uh oh!

eder-matheus commented Jul 3, 2026

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot Jul 3, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants