Excessive run time for building a large network from GMNS

4 views
Skip to first unread message

Smith, Scott B (Volpe)

unread,
Dec 17, 2025, 7:18:50 AM12/17/25
to aequi...@googlegroups.com

I made a GMNS node-link network from the Freight Analysis Framework (FAF5) network,  around 500k links and 300k nodes. 

AequilibraE 1.4.2 seems to be hanging on running project.network.create_from_gmns

 

Observed behavior:

For the first hour, it was consuming about 2 GB of RAM, and 10-20% of CPU

For at least some of the next 16 hours, it was consuming about 200k of RAM, and 10-20% of CPU

I killed the process after 17 hours. 

 

My computer:

Dell Latitude 7430

Intel(R) Core(TM) i7-1265U   1.80 GHz

16 GB RAM

About 20 GB free disk space

 

Is it normal to take this long to build a large network?  Thanks.

 

  • Scott

Scott B. Smith, Ph.D.

Room 444, Technology Analysis and Strategy Division (V-336)

Volpe National Transportation Systems Center | U.S. Department of Transportation

220 Binney St, Cambridge, MA 02142

( +1 617.494.2588 office; +1 857 998 2234 mobile | * scott...@dot.gov | 8 www.volpe.dot.gov

Advancing transportation innovation for the public good

 

 

Pedro Camargo

unread,
Dec 17, 2025, 6:07:34 PM12/17/25
to aequi...@googlegroups.com
Hey Scott,

I suspect that the issue here is coming from the spatial triggers that maintain links and nodes mutually consistent.

In these cases (I have built a model from the cadastral networks for the entire US), the solution is to import nodes and then links, removing all triggers before the insertion of links.

This would probably not work when importing from GMNS, but network import would take a few minutes in that case.

When going the no-trigger route, however, you forgo the guarantee of network consistency, so the a_node and b_node fields in your links may be wrong.

Cheers,
Pedro


---- On Wed, 17 Dec 2025 23:18:43 +1100 'Smith, Scott B (Volpe)' via AequilibraE <aequi...@googlegroups.com> wrote ---

--
You received this message because you are subscribed to the Google Groups "AequilibraE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aequilibrae...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/aequilibrae/EA3PR09MB12132E647138D8BADB4ED970BEFABA%40EA3PR09MB12132.namprd09.prod.outlook.com.


Scott Smith

unread,
Dec 18, 2025, 5:52:46 AM12/18/25
to AequilibraE
Thank you, Pedro.  I suspected the issue was the spatial triggers.   - Scott
Reply all
Reply to author
Forward
0 new messages