Jaehyun Kim

rsz: repair_design crashes druing global placement in hierarchical flow

2

### Describe the bug In hierarchical flow (OPENROAD_HIERARCHICAL=1), `nangate45/bp_multi_top` crashes during global placement. **Problematic designs** - `nangate45/bp_multi_top` - `nangate45/black_parrot` **bp_multi_top - crash log** ``` /workspace/clean/OpenROAD-flow-scripts/flow/scripts/flow.sh 3_3_place_gp global_place Running global_place.tcl, stage...

### Describe the bug `asap7/cva6` design failed with `[ERROR PSM-0069] Check connectivity failed on VSS.` error in `final_report` stage. ### How to reproduce OpenROAD-flow-scripts/flow$ make DESIGN_CONFIG=./designs/asap7/cva6/config.mk SYNTH_WRAPPED_OPERATORS=1 ### Error log...

### Describe the bug Resizer has multiple buffering logics here and there (e.g., `rebuffer, makeRepeater, splitLoadMove, cloneMove, ...`). In a hierarchical flow, I found that those buffering logics can break...

### Describe the bug net1 -- buf -- net2 When a `remove_buffers buf` is executed for the `buf`, the `net1` and `net2` are merged into `net1` and the resulting RC...

### Describe the bug If hierarchical flow (OPENROAD_HIERARCHICAL=1) is enabled, it gives more restrictions on buffering related to `dbModNet`. It results in different buffering outcome and different QoR compared to...

### Describe the bug The current `repair_tie_fanout` command inserts new TIE cells for each load pin and remove the original TIE cell. This causes an undriven hierarchical input and `dbModNet`...