Is this a tools issue or a configuration issue? I...
# openlane
Is this a tools issue or a configuration issue? If it's a tools issue, has it been reported yet?
Copy code
Update congestion history type 2
[INFO] iteration 19, enlarge 205, costheight 132, threshold 0 via cost 0 
[INFO] log_coef 0.4180[ERROR]: during executing: "openroad -exit /openLANE_flow/scripts/openroad/or_route.tcl |& tee >&@stdout /project/openlane/decred_top/runs/decred_top/logs/routing/fastroute.log"
[ERROR]: Exit code: 1
[ERROR]: Last 10 lines:
child killed: segmentation violation

[ERROR]: Please check openroad  log file
[ERROR]: Dumping to /project/openlane/decred_top/runs/decred_top/error.log
your design is probably too dense
most of the time something will crash instead of telling you
reproducing segfault is very hard on different systems, you'd have to provide the input files / scripts and backtrace for a dev to have any chance of tracking it down.
The source is here: To reproduce, follow the directions in the
Heh yeah, usually they want a bit more than that, like the exact Open ROAD script and input file not the whol eproject and "follow" readme.
it's unlikely that the OpenROAD devs are going tthrough the whole OpenLane install and PDK install just to reproduce your issue.
The offending script is referenced in the github issue I opened. The project I provided has all of the necessary inputs that are not already standardized by the environment required for all work we're doing in this slack project (skywater-pdk/openlane). I think everyone would agree that a tool should not just "run out of memory" without any indication of the fundamental problem (at a minimum). I'm not saying I don't have a configuration issue, I just don't see it. Finally, the failure occurs with a
so I really hope this is not a density issue.
(1) It's not running out of memory,
segmentation violation
is not an OOM kill (2) All I'm saying is that what's crashing here is
. It should never segfault, under any circumstances. But that tool is something that just happens to be used by OpenLANE, it's independent from it. The OpenROAD devs are probably the one that should be debugging this issues. but those guys don't use OpenLANE, they don't have anything to do with the 'standardized environment' . If a crash happenned in
, it'd be the same thing, it would need to be reported and fixed by the
developers and you wouldn't expect them to go and setup OpenLANE just to reproduce your issue, you'd need to provide a minimal test case using just
to reproduce it.
If the developer that addresses this problem needs more, I'd be happy to provide whatever they request.