Re: [sumo-user] Optimizing sumo simulations by configuring enabling infr

[Jakob Erdmann <namdre.sumo@xxxxxxxxx>]

No Literature, this is all from personal experience:

- Loading input files from ramdisk can help (but won't have a large impact for long-running simulations)

- routing can be parallelized with option --device.rerouting.threads   and potentially sped up further with other routing algorithms: https://sumo.dlr.de/docs/Simulation/Routing.html#routing_algorithms

- Large Caches and high single-core performance will help

- option --threads is rather experimental and the speedup is low: using 16 cores can give a modest speedup (i.e. 2-3)

Am Mi., 5. Juni 2024 um 17:55 Uhr schrieb Hector A Martinez via sumo-user <sumo-user@xxxxxxxxxxx>:

Good morning SUMO users,

 

Has there been any prior work to improve the performance of sumo microscopic simulation by configuring the underlying infrastructure (RAM, CPU, Disk, networking, etc.) configuration either via parallel processing or using containers vs. VMs, etc.?

 

We are interested in large regions of simulation and we need to find a way to improve the performance of the simulation for quick turn analysis.

 

Any literature on optimal architectures to speed up the SUMO performance that can be shared will be greatly appreciated. Cheers!

 

Hector A. Martinez, P.E.

MITRE | Solving Problems for a Safer World (R)

813.207.5365

 

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