In response to my earlier post, I'd like to thank Chris Jackson and Andrew Thomas for helping (with much patience) to solve the issue I was having using car.normal in OpenBUGS, on Linux and my 64-bit Windows PC (I was not encountering the problem using WinBUGS on 64-bit windows). Andrew Thomas has created a patch to fix the bug affecting CAR models with islands in the map, but only for BRugs (on Linux and 64-bit Windows).   Unfortunately, after installing the patch, I still encountered difficulties using BRugs, possibly due to the monitors that it uses (see comment below). Therefore, I switched to R2OpenBUGS, with Chris's suggestion, and this solved the problem.

Chris Jackson advises "to avoid R and use OpenBUGS directly, on Windows or Linux.  Scripts will make this easier.  This should also be much faster than using BRugs for such a massive model (8715 areas) with so many variables.  (in particular, using sample monitors, rather than summary monitors, slows BRugs down)."

He adds:

. . . "the alternative of R2OpenBUGS, if you still want to call OpenBUGS from R.  This shouldn't suffer from this problem. 

The syntax is a bit different, but it's essentially the same functionality as  BRugsFit which you were using for running whole analyses in a single function call.

 

Also I mentioned sample monitors because I don't know if you actually need to save entire posterior samples of those variables in your model with over 8000 elements, which takes up large amounts of memory. Summary monitors will get you the running posterior means and standard deviations, without that memory requirement."


Many thanks,

Holly




Date: Sun, 9 Sep 2012 20:35:21 -0400
From: [log in to unmask]
Subject: [BUGS] car.normal in OpenBUGS
To: [log in to unmask]

Hello list,

I am having trouble running a car.normal in OpenBUGS. Not sure if this has something to do with accessing geoBUGS?

 

On my PC, the car.normal models run in WinBUGS 1.4.3.

However, using car.normal models in both OpenBUGS 3.2.1 and 3.2.2: on my Windows PC, I get a trap, and on x86_64 GNU/Linux models I get the following errors:

 

Using BRugs:

#model is syntactically correct

#data loaded

#Error in handleRes(res) :

#  Internal "trap" error in OpenBUGS, or non-existent module or procedure called.

#Calls: BRugsFit ... .OpenBUGS.platform -> .OpenBUGS.helper -> handleRes

#Execution halted

 

Using R2WinBUGS (directing to OpenBUGS):

#after 22 hours:

#Carriage returns added to model file

#model is syntactically correct

#data loaded

#model is initialized

#1000 updates took 655 s

#can not calculate DIC for this model

#monitor set for variable …

#9000 updates took 1571 s

#Error in handleRes(res) :

#  Internal "trap" error in OpenBUGS, or non-existent module or procedure called.

#Calls: openbugs ... .OpenBUGS.platform -> .OpenBUGS.helper -> handleRes

#Error in handleRes(res) :

#  Internal "trap" error in OpenBUGS, or non-existent module or procedure called.

#Calls: openbugs ... .OpenBUGS.platform -> .OpenBUGS.helper -> handleRes

#Execution halted

 

Having referred to two earlier posts, I believe this is a problem with the car.normal in OpenBugs, not in the linux platform:

“Re: Issues with BRugs vs. OpenBugs on Linux” from 25 Aug 2012: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1208&L=BUGS&D=0&P=12673

“Re: problem with car.normal in openBUGS” from 15 May 2009: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind0905&L=BUGS&D=0&1=BUGS&9=A&J=on&d=No+Match%3BMatch%3BMatches&z=4&P=5408

 

Thank you,

Holly

------------------------------------------------------------------- This list is for discussion of modelling issues and the BUGS software. For help with crashes and error messages, first mail [log in to unmask] To mail the BUGS list, mail to [log in to unmask] Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html Please do not mail attachments to the list. To leave the BUGS list, send LEAVE BUGS to [log in to unmask] If this fails, mail [log in to unmask], NOT the whole list
------------------------------------------------------------------- This list is for discussion of modelling issues and the BUGS software. For help with crashes and error messages, first mail [log in to unmask] To mail the BUGS list, mail to [log in to unmask] Before mailing, please check the archive at www.jiscmail.ac.uk/lists/bugs.html Please do not mail attachments to the list. To leave the BUGS list, send LEAVE BUGS to [log in to unmask] If this fails, mail [log in to unmask], NOT the whole list