Jump to content
TUFLOW Forum

eirianc

Members
  • Content Count

    6
  • Joined

  • Last visited

Community Reputation

0 Neutral

About eirianc

  • Rank
    Member

Recent Profile Visitors

912 profile views
  1. Our office IT has recently moved to cloud servers. Since the move there have been a number of times where TUFLOW (Classic) has stopped running mid-simulation without any errors or warnings. The model is running smoothly until it stops.. This has happened maybe 10 times over the last 2 weeks. Looking at the .tlf it seems to happen when the simulation is writing outputs, but this may not always be the case. We have tried running the same model reading and writing to the local machine and it worked fine. The model is a 76hr run time, so its a bit frustrating when it stops midway 😬 We have contacted our cloud service provider and they have completed some testing which was apparently all fine from their end. Anyone else had this issue or got any suggestions what might be the cause/solution? Thanks Eirian
  2. Hi, thanks for the response. For info the model is using the direct rainfall method over an urban catchment. All buildings footprints have been removed from the active model however SA boundaries (rainfall v time, using Read GIS SA RF) are used around the perimeter of each building to allow for the runoff from the building roof. Cheers
  3. .....I have found the cause of the high memory usage: Each building roof was modelled with an SA flow boundary, with over 900 SA inputs in the model. As a temporary fix (until we add more RAM to the PC), when I remove a large proportion of these SA boundaries (outside the catchment and specific area of interest), the model runs much faster, using only 4GB of RAM.
  4. I am having trouble running a Tuflow model created by others. I have run similar sized/complex models before without much problem so I don’t think it is a hardware issue. I have got the simulation started but it is running extremely slowly. I left it running for 4 days and it was taking 6 hours of real time for 1 minute of simulation time, so that on the 4th day it had only run 15 mins of a 2hr simulation. Running a coarser model (increasing cell size, timestep etc.) doesn’t seem to speed up the simulation. I have checked the .tlf and it seems to be a memory allocation issue. The 2d ‘Boundary Conditions’ memory is over 8GB, (memory allocation screenshots attached) everything else seems to be ok. Can anyone let me know why this would be so big and how to reduce its size? For info I am running 2013-12-AC-iDP-w64 Tuflow version on a 64 bit PC with 8GB RAM: Intel(R) Xeon(R) CPU E3-1220 v3 @ 3.10GHz)
×
×
  • Create New...