|
[Sponsors] |
December 16, 2009, 10:03 |
Memory error in ILURES.
|
#1 |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Hi,
I got the following error after a couple of iterations: +--------------------------------------------------------------------+ | ERROR #001100279 has occurred in subroutine ErrAction. | Message: | Memory error in ILURES. Please increase the value of the expert | parameter 'ilures memory factor' (default value is 1.0) and try a- | gain. | | | +--------------------------------------------------------------------+ I couldnt find 'ilures memory factor' under expert parameters (I guess its hidden somewhere?), but I ran exactly the same def file again and then the simulation worked without any errors.... strange to say the least. Any idea why this happened? Could it be hardware related? Lance |
|
August 10, 2010, 03:24 |
|
#2 |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Ok, I am getting this error again, and it dont seem to be hardware related as it happens on different compute nodes each time.
And I cant find the expert parameter - at least not in the GUI. Any ideas? |
|
August 10, 2010, 08:44 |
|
#3 |
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,872
Rep Power: 144 |
Almost certainly this is caused by some physical model you have defined which is not working properly. Can you describe what physical models your model has, or upload the CCL?
|
|
August 10, 2010, 08:57 |
|
#4 |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Im running internal pipe flow, LES (WALE) with heat transfer, specified mass flow at the inlet, pressure at the outlets. The thing is that sometimes the case works, sometimes it doesnt...
Previously I ran the case on 20 compute nodes (=160 cores) which didnt work. Now I have tried 10 and 30 nodes (i.e. 80 and 240 cores) and the case works perfectly. There has always been different compute nodes in use when I got the error, but I guess it could be some sort of error on several nodes at the same time. |
|
August 10, 2010, 23:01 |
|
#5 |
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,872
Rep Power: 144 |
I see. This is just a guess but it could be the linear solver is having problems over so many processes. What version of CFX are you running?
|
|
August 11, 2010, 02:59 |
|
#6 |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Im using 12.0. Maybe should have mentioned that the solver always fails after the first coefficient loop in the first time step.
Are there "hidden" expert parameters somewhere? Since the error says "Please increase the value of the expert parameter 'ilures memory factor' (default value is 1.0)", but it is nowhere to be found in the GUI. |
|
August 11, 2010, 07:23 |
|
#7 |
Super Moderator
Glenn Horrocks
Join Date: Mar 2009
Location: Sydney, Australia
Posts: 17,872
Rep Power: 144 |
Definitely looks like a linear solver problem. Not much you can do about this, except adjusting the parameter as it recommends. If that does not work you will have to contact support.
Yes, there are lots of undocumented options in CFX. That s why you need to talk to support about things like this as there is no other way to sort it out. |
|
August 11, 2010, 07:47 |
|
#8 |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Thanks Glenn,
if I get something useful from the support I'll post it here if someone else gets the same problem. |
|
November 27, 2010, 13:09 |
ilures memory factor
|
#9 |
New Member
Richard
Join Date: Oct 2010
Posts: 5
Rep Power: 16 |
Hi Lance
Did you ever manage to solve the problem with the memory factor. I have been trying to run a simulation on a cluster and it keeps coming up with the same message that you were getting. I have adjust various memory factors but none of them seem to make any difference. Thanks Richard |
|
November 29, 2010, 03:17 |
|
#10 |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Hi Richard,
The reply from support was that I should increase the memory factor by trial and error, starting at 1.1 or 1.2: "The value of 1.2 is fairly arbitrary and will need to be determined by 'trial and error', e.g. 1.1 might work but you probably shouldn't need to make it as high as 2.0 or you might cause other memory problems. It's possible that ILU smoother is close to running out of memory in your case although this is an unusual error [...]" |
|
October 26, 2011, 00:02 |
|
#11 |
Member
Lingdeer
Join Date: Jul 2011
Posts: 49
Rep Power: 15 |
Hi, I have similar problem but could not find a solution.
So I started off by adding ILURES=1.1 and also ilures memory factor=1.1 in CFX-pre and encountered an error. Is it the right way to solve it? |
|
October 27, 2011, 03:31 |
|
#13 | |
Senior Member
Lance
Join Date: Mar 2009
Posts: 669
Rep Power: 22 |
Quote:
FLOW: EXPERT PARAMETERS: ilures memory factor = 1.2 END END |
||
October 27, 2011, 14:45 |
|
#14 |
Member
Lingdeer
Join Date: Jul 2011
Posts: 49
Rep Power: 15 |
Yeah I contacted the ANSYS technical staff and this is what I tried and worked:
FLOW: Flow Analysis 1 EXPERT PARAMETERS: ilures memory factor = 1.1 END END and if it doesnt then I need to increase the factor to 1.2 -2.0 Just post here to help others in case. Thanks you all so much ! |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Memory/ swap area | J-man | STAR-CD | 1 | September 29, 2009 20:43 |
How to optimize the memory usage when using FEM | vasilis | Main CFD Forum | 11 | August 25, 2009 00:57 |
Memory error? | Young | CFX | 3 | September 30, 2008 12:33 |
CFX CPU time & real time | Nick Strantzias | CFX | 8 | July 23, 2006 18:50 |
Run-Time memory conifguration Error | Graeme | CFX | 1 | February 21, 2006 00:04 |