palaniappan-r's picture
Add markdown sources
d59bb1b verified
# Running times for 6_report as a function of area
Tool: Global Routing
Subcategory: Performance issue
## Conversation
### oharboe
Using https://github.com/The-OpenROAD-Project/OpenROAD-flow-scripts/pull/1211 and executing `python util/plot-running-times.py`, I've plotted some running times against area:
`6_report` running time in seconds below. Note that the x axis is um^2, but that it is multiplied by 1e6 (as can be seen in the lower right corner).
As can be seen, the running time is worse than proportional to area and I imagine that there could be some low hanging fruit as I would expect reporting not to be on the order of detailed routing times.
![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/499880aa-0e07-4772-9e10-73dddcf9ef51)
```
cat /proc/cpuinfo | grep processor | wc -l
12
```
### oharboe
Same for `2_6_pdn`, here running time is proportional to area:
![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/12bf57a3-9670-47d3-a19e-98c35f023ac1)
### oharboe
`5_2_TritonRoute` running times are misleading as there's some sort of problem with DRC errors #3592
However, since I have the data, I include the plot here anyway.
![image](https://github.com/The-OpenROAD-Project/OpenROAD/assets/2798822/80e6d705-994a-408d-9847-461b75cfaa37)
### oharboe
A feature request has been filed to improve reporting times #3599