Finish the dense FIFO placement without keeping the legalizer’s full wirelength cost.
Hardware context: Legalized standard-cell datapaths. Check placement legality, wirelength and estimated timing together.
The FIFO core is fixed at the L4.3 70% utilization request. Global placement finds short nets, but legalization moves overlapping cells onto legal sites and stretches the wiring.
Choose the target placement density and whether to run improve_placement followed by optimize_mirroring after legalization. Finish with a passing legality check, HPWL at most 6500 µm and estimated maximum slack at least 0.1 ns.
Requirements
| Requirement | Limit | Unit |
|---|---|---|
| Detailed-placement legality check passed | = 1 | count |
| Final placement HPWL | ≤ 6500 | µm |
| Estimated maximum slack | ≥ 0.1 | ns |
Live analysis: Live FIFO placement at TT and a 5 ns clock, through detailed placement. The core stays fixed. Placement RC estimates cover the internal timed paths; the example retains electrical violations, so this task is a placement milestone, not sign-off.
Full specification
| JSON field | Allowed values | Meaning |
|---|---|---|
| density | 0.71 to 0.9 | A soft density objective above the 70% utilization request. |
| improve | false, true | Run local placement improvement and cell mirroring after legalization. |
Live FIFO placement at TT and a 5 ns clock, through detailed placement. The core stays fixed. Placement RC estimates cover the internal timed paths; the example retains electrical violations, so this task is a placement milestone, not sign-off.
Numeric checks use inclusive bounds. A missing measurement always fails. Only the documented fields are accepted.
Examples
Example 1
Improve the legal placement- Input
A candidate configuration for the fixed circuit and stated environment.
- Output
Detailed-placement legality check passed: 1 Final placement HPWL: 6291.9305 µm Estimated maximum slack: 0.1775 ns
A legal placement can still miss its wirelength target.
Example 2
Stop after legalization- Input
{ "density": 0.71, "improve": false }- Output
Detailed-placement legality check passed: 1 Final placement HPWL: 6821.6395 µm Estimated maximum slack: 0.1654 ns
A legal placement can still miss its wirelength target.
Example 3
Change density alone- Input
{ "density": 0.8, "improve": false }- Output
Detailed-placement legality check passed: 1 Final placement HPWL: 6573.9155 µm Estimated maximum slack: 0.404 ns
A legal placement can still miss its wirelength target.
Submission rules
- Submit one JSON object with every documented configuration field. The supplied circuit and external environment stay fixed.
- Run measures your configuration. Submit applies the same requirements and records the result.
- The stated stage and cell corner define this exercise; passing does not establish full chip sign-off.
Hints
Hint 1
Legalization and placement improvement are separate passes.
Hint 2
A lower target density is not a hard occupancy limit.
Hint 3
Check the final HPWL after the optional local passes.
What is checked: Live FIFO placement at TT and a 5 ns clock, through detailed placement. The core stays fixed. Placement RC estimates cover the internal timed paths; the example retains electrical violations, so this task is a placement milestone, not sign-off.
Run and Submit use the same measured requirements. Run is a preview; Submit records an attempt. The fixed circuit and allowed settings use the pinned OpenROAD/OpenSTA engine.
More problems on this topic