CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > FLUENT

strange win64 issue

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   October 21, 2009, 11:52
Default strange win64 issue
  #1
Member
 
Akour
Join Date: May 2009
Posts: 79
Rep Power: 9
ak6g08 is on a distinguished road
hello everyone,

Ive got a UDF and case/data file that all work and compile in win32. when I then copy the case and data files on to another win64 PC (and re-compile the udf using win64 visual studio) and then start iterating...the residuals look completely different, and it seems as if I am running a completely different problem, however the source file i use is exactly the same, along with exactly identical case and data files. I am also issuing the command:

solve/set/expert in order to free temporary memory because i am doing some gradient calculations....

do I need to do something to make the problem work the same way in win64??? i find it very odd that this should happen

thanks in advance
__________________
akour
ak6g08 is offline   Reply With Quote

Old   October 22, 2009, 01:54
Default
  #2
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,145
Rep Power: 32
-mAx- will become famous soon enough
is your data file a result, or it is just an initial guess?
if it is a result, check if the results are the same as on win32.
Else: after reading case and data into win64, write case and data on the fly, and then try to iterate.
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Old   October 22, 2009, 05:35
Default
  #3
Member
 
Akour
Join Date: May 2009
Posts: 79
Rep Power: 9
ak6g08 is on a distinguished road
Thank you for your reply,

the .dat file i am copying into the win64 machine just contains initial conditions. Actually, i have tried it now without the solve/set/expert command and the problem is still there (i thought maybe it had something to do with the temp memory). I have also tried making the .cas file from scratch in the win64 machine and then iterating from there...but the problem is the same. The problem, to be more specific, is that the continuity residual jumps to 1e15 on the first iteration (in the win64 machine)...but it doesnt do this in the win32 machine...has this ever happened to anyone???

I am not getting and compilation errors coming up...I welcome any more suggestions.

Thanks again
__________________
akour
ak6g08 is offline   Reply With Quote

Old   October 22, 2009, 06:43
Default
  #4
Member
 
Akour
Join Date: May 2009
Posts: 79
Rep Power: 9
ak6g08 is on a distinguished road
just to add something, i have also tried loading a .dat file (With data), I can display all the contours succesfully on the win64 machine, but when i try to continue the iteration from that point in time (again, with same materials etc...). the residuals jump to a much higher value then they should...as if the simulation were reading in a completely different data set...
__________________
akour
ak6g08 is offline   Reply With Quote

Old   October 22, 2009, 06:43
Default
  #5
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,145
Rep Power: 32
-mAx- will become famous soon enough
try to import your mesh file into win64, and set up the case file as in win 32.
And iterate.
If you still have the issue, let run a computation on win 32 WITHOUT udf, and do the same on win64. And compare.
Just to be sure, if it comes from the udf...
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Old   October 22, 2009, 06:44
Default
  #6
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,145
Rep Power: 32
-mAx- will become famous soon enough
Quote:
Originally Posted by ak6g08 View Post
but when i try to continue the iteration from that point in time (again, with same materials etc...). the residuals jump to a much higher value then they should...as if the simulation were reading in a completely different data set...
it is normal
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Old   October 22, 2009, 07:11
Default
  #7
Member
 
Akour
Join Date: May 2009
Posts: 79
Rep Power: 9
ak6g08 is on a distinguished road
hi,

why is that normal? when i read in a .dat file which runs on a simulation without UDFs, this doesnt happen...it just continues normally. and when i read in a data file with the UDFs into a new FLUENT window on my PC, the residuals do not jump, and when i mean jump, i mean by 3 or 4 orders of magnitude, seems a bit strange to me, i would need to lower the residuals drastically, in order to run the same simulation on a different PC?? this is odd...

I have tried running a .dat file without udfs, and the problem is not present.

thanks again
__________________
akour
ak6g08 is offline   Reply With Quote

Old   October 22, 2009, 07:40
Default
  #8
Super Moderator
 
-mAx-'s Avatar
 
Maxime Perelli
Join Date: Mar 2009
Location: Switzerland
Posts: 3,145
Rep Power: 32
-mAx- will become famous soon enough
when I say, it is normal, I mean you change something in the solver or whatever (solver on win32 and win64 shouldn't be exactly the same).
You can try to run a job, and after 100 iterations, change the URF... you will also see a jump.
But no matter, it is afterall important, that the job continues to converge after the jump..
If you don't want to see the jump, re-iterate from zero.

If the issue disappears without UDF, then I assume it is linked to the UDF compiler...
__________________
In memory of my friend Hervé: CFD engineer & freerider
-mAx- is offline   Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
Strange Flamelet Generation Problem tar FLUENT 11 March 22, 2014 11:52
Strange behaviour when using LienCubicKE and NonlinearKEShih hani OpenFOAM Running, Solving & CFD 20 March 6, 2013 11:06
Strange influence of initial condition? Julien FLOW-3D 6 November 6, 2008 06:03
FLUENT Solver i strange BastiL FLUENT 1 July 14, 2008 16:27
Memory issue - Suse 10 - Opterons Andy R FLUENT 1 June 23, 2008 14:44


All times are GMT -4. The time now is 23:15.