open-source-silicon.dev

I had to switch to using `.lef` rather than `.magic.lef` because of the access point issue in the `skywater-pdk` master ... but known issue ...

I tried cherry-picking the ROUTING_CORE commit to get more routing cores but ... doesn't seem to do anything.

<@U017E7L119N> Did setting threads:NNN do anything for you for TritonRoute ?

allowed me to generate a lot of heat and converge faster…

Mmm ... here I still see only 1 core being used :confused:

And I checked that `tritonRoute.param` has the field set.

But even in the sources, that value defaults to `8`

What repo / commit are you running from ?

```/OpenROAD-flow/flow ] $ git show
commit 3f559a4c9aeb6868fe75c80a5cf554620b2280f8 (HEAD -&gt; master, origin/master, origin/HEAD)```


```tools/TritonRoute ] $ git show
commit ad4b80c34b22e578e61802c6ac19b7f903476e48 (HEAD, origin/openroad)```
&lt;//github.com/The-OpenROAD-Project/TritonRoute.git&gt;

ok, this is where I added:
```openlane ] $ cat scripts/tritonRoute.param
lef:${MERGED_LEF_UNPADDED}
def:${CURRENT_DEF}
guide:${fastroute_tmp_file_tag}.guide
output:${tritonRoute_result_file_tag}.def
outputTA:${tritonRoute_tmp_file_tag}_TA.def
outputguide:${tritonRoute_tmp_file_tag}.guide
outputDRC:${tritonRoute_report_file_tag}.drc
outputMaze:${tritonRoute_tmp_file_tag}_maze.log
optimizationMode:${ROUTING_STRATEGY}
threads:16
verbose:1```

Yup switching to TritonRoute14 strategy (which used the OR repo latest version), then I get multicore usage.

i guess I got really lucky, because I only noticed it was slow from trying openroad’s msaligane early branch

will build 14 aswell to test differnt strategies