Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [sumo-user] Performance Optimization Inquiry for mapDetectors.py and routeSampler.py

1. 
mapDetectors.py: 
- reduce the radius to the minimum value that fits the accuracy of your coordinates (the default of 1000m likely pulls in to many edges)
- you could partition the list of coordinates and run multiple instances in parallel (then combine the results in a post-processing step)
routeSampler.py: 
- reduce the number of candidate routes if possible,
- if you can, avoid optimization as this is much slower than sampling. Partial optimization may also be faster than sampling though I didn't test this in a systematic way.
- if you are solving multiple time intervals, you can use option --threads to compute them in parallel

2. no systematic tests are available. Parallelization for routeSampler, obviously depends on multiple cores.  Both tools are CPU-bound which should give you an obvious way to speed up processing.

3. Personally, I wasn't bothered with processing times of several minutes so my data wouldn't be much use to you.


Am Mo., 21. Apr. 2025 um 13:59 Uhr schrieb Manohara C V via sumo-user <sumo-user@xxxxxxxxxxx>:
Dear SUMO Community,
I am encountering performance issues when working with real-world scenarios using SUMO tools.
  • mapDetectors.py: When processing a high number of detectors (approximately 150), the script takes more than a minute to generate the output xml file.
  • routeSampler.py: Even after restricting the route generation to specific time intervals using -b and -e options, and applying --optimize full, the route creation process is taking over a minute to complete.
Could you please advise on:
  1. Best practices or configuration changes to improve the performance of these scripts?
  2. Recommendations for hardware setup (e.g., number of CPU cores, memory) to achieve faster processing times.
  3. If possible, could you also share your typical system specifications and setup that deliver good performance for similar use cases?
Any guidance to help optimize the response times would be highly appreciated.
Thank you for your support!
Note:  I'm working with a Python script to generate the networks and route files and windows OS.

Thanks & Regards,

Manohara C V

AI Services 


_______________________________________________
sumo-user mailing list
sumo-user@xxxxxxxxxxx
To unsubscribe from this list, visit https://www.eclipse.org/mailman/listinfo/sumo-user

Back to the top