Multiple Nesting and parallel computing

Discuss the nesting capability within the model itself and any problems you might have run into.

Multiple Nesting and parallel computing

Postby sebastian » Wed Jan 06, 2010 8:41 am

Dear Users,

I run a simulation with nine nests which are at the same level and have the same parent domain.
If executed in serial mode it works fine. So my namlist.input should be correct.

However, if I trie to run it in parallel mode a segmentation fault occured while executing wrf.exe.
If only one nest is included it works fine as well. The problem occurs when adding a second nest.

I've compiled wrf with icc and ifort for dmpar. I'm using WRF 3.0. I've already increased the stacksize.

Here is the extract from the rsl.error


d01 2009-01-10_12:00:00 med_initialdata_input: calling input_model_input
INITIALIZE THREE Noah LSM RELATED TABLES
STEPRA,STEPCU,STEPBL 540 90 1
INITIALIZE THREE Noah LSM RELATED TABLES
STEPRA,STEPCU,STEPBL 540 90 1
INITIALIZE THREE Noah LSM RELATED TABLES
STEPRA,STEPCU,STEPBL 180 30 1
*************************************
Nesting domain
ids,ide,jds,jde 1 40 1 25
ims,ime,jms,jme -4 45 7 30
ips,ipe,jps,jpe 1 40 13 25
INTERMEDIATE domain
ids,ide,jds,jde 14 32 29 42
ims,ime,jms,jme 9 37 29 47
ips,ipe,jps,jpe 12 34 35 44
*************************************
forrtl: severe (174): SIGSEGV, segmentation fault occurred
Image PC Routine Line Source
wrf.exe 000000000043835A Unknown Unknown Unknown
wrf.exe 000000000049386A Unknown Unknown Unknown
wrf.exe 00000000004978AC Unknown Unknown Unknown
wrf.exe 0000000000408429 Unknown Unknown Unknown
wrf.exe 000000000040A773 Unknown Unknown Unknown
wrf.exe 000000000043D1EC Unknown Unknown Unknown
wrf.exe 0000000000405883 Unknown Unknown Unknown
wrf.exe 0000000000405837 Unknown Unknown Unknown
wrf.exe 00000000004057CC Unknown Unknown Unknown
libc.so.6 0000003FBAE1D974 Unknown Unknown Unknown
wrf.exe 00000000004056D9 Unknown Unknown Unknown


Thanks,
Sebastian
sebastian
 
Posts: 1
Joined: Wed Jan 06, 2010 8:35 am

Re: Multiple Nesting and parallel computing

Postby vincentajayi » Fri Feb 26, 2010 6:05 am

I have the same problem and I am only running one nest, if you solve it please let me know
vincentajayi
 
Posts: 2
Joined: Tue Oct 06, 2009 6:30 am

Re: Multiple Nesting and parallel computing

Postby jimmyc » Fri Feb 26, 2010 12:49 pm

Could this be a memory error? Or is this one of those "too many nests" error. I remember reading somewhere that there is flag to change if you want to run more than 5 nests. I don't recall where I read that ( and if it was a previous wrf version).
The views expressed in this message do not necessarily reflect those of NOAA or the National Weather Service or the University of Oklahoma.
James Correia, Jr
jimmyc
 
Posts: 519
Joined: Tue Apr 15, 2008 1:10 am

Re: Multiple Nesting and parallel computing

Postby ziad » Sat Feb 27, 2010 1:39 pm

The most nests I've run is four telescopic, 2-way nesting. Ran into segmentation faults here and there and the best way to get rid of them was to make sure the time steps for the nests were small enough. It really came down to a CFL stability problem which tended to get worse as the solution advanced so I had to use some pretty small time steps on the innermost nests.

Probably completely unrelated to your problem but small time steps cannot hurt and might allow the code to converge.

How much memory does your setup require anyway and how much do you have available on your system? running parallel, serial, shared, distributed?
ziad
 
Posts: 9
Joined: Tue May 05, 2009 4:41 pm


Return to Nesting

Who is online

Users browsing this forum: No registered users and 3 guests