How do you check there's physical room for ECO cells before fixing?

Short Answer

Run report_cell_feasible_space (ICC2) on the region around your violations before PrimeTime adds any buffers. It reports free placeable sites, grouped by width, after subtracting existing cells, placement blockages and macros. If there are no gaps wide enough for the cells you plan to add, fix with sizing, use PrimeTime's open-site mode, or make room first.

Technical Explanation

• Feasible space is placeable site area minus occupied area. Occupied means existing cells (ECO cells excluded), placement blockages and macros.
• By default the whole block is reported. -voltage_areas restricts it to named voltage areas and -boundary to one rectangle or rectilinear polygon; the two options are mutually exclusive. With -boundary, only space inside voltage areas within the boundary counts.
• Spaces are grouped by width, and anything narrower than the smallest existing cell is reported as less than minimum width. Designs with several site definitions get a separate breakdown for each.
• The report assumes every non-ECO cell is legal. Non-ECO cells off the site grid are treated as free space, collected in $ecoNonLegalCellsForSpaceCheck and flagged with a warning. ECO cells are ignored unless their eco_change_status is eco_legalized.
• Fillers occupy sites. -ignore_filler_references treats unfixed fillers of the listed references as free; fixed fillers always count as occupied.
• Why it matters: the default place_eco_cells (ICC2) mode uses free sites only, and a cell with no free site nearby can be displaced a long way. Knowing the space first lets you choose sizing over buffering, or fix_eco_timing -physical_mode open_site (PT), which only makes changes where room already exists. Physically aware modes need a PrimeTime-ADV license (PT UG, ECO chapter).
• A buffer must sit in the voltage area of its net's power domain, so free space in a neighbouring voltage area does not help.

What To Check

• Free space by width inside the violation region, not only the whole-block total.
• The widest gaps against the width of the buffers or delay cells you intend to use.
• Whether fillers were counted as occupied or ignored.
• Any warning about non-legal cells, which makes the numbers unreliable.
• The voltage area the fixes must go into.

Healthy, Suspicious & Hard-stop Results

• Healthy (illustrative): The violation region has 120 gaps of four sites or wider, enough for the planned 30 buffers.
• Suspicious (illustrative): Most free space is in slivers below minimum width, or appears only when fillers are ignored.
• Hard stop: No gap in the region is wide enough for the smallest delay cell, or the non-legal-cell warning fires and the report cannot be trusted.

What The Interviewer Is Testing

• Checks space in the region and voltage area that matter, not the global number.
• Knows what the report counts as occupied, including fillers and ECO cells.
• Can say how the answer changes the fixing method.

Follow-up Question & Model Response

"Why does the report ignore ECO cells unless they are eco_legalized?"

Candidate Model Response: ECO cells that are not yet legalized only have tentative locations, and many may overlap existing cells or each other. Counting them would mark sites as taken that may end up free, or miss the sites they will really take. Once legalized, their locations are final and they are counted like any other cell. So run the report after legalizing one batch to see what room is left for the next.