On 07/29/2011 07:04 AM, [email protected] wrote:
Dear Guangping
I read that, but I am not able to solve the problem, could you please
tell in the details what should I do???
Just as Guangping have told you, you shouldn't let the NP's value larger
the the number of k points of the system. In this case, the number of k
point is 1, so the $NPROCS - I myself use lsf and not so familiar with
pbs, so just my guess based on the name of the argument's name - should
equal to 1 in your case.
For more detail, see here:
http://www.mail-archive.com/[email protected]/msg02885.html
Regards
Anjana
Quoting Guangping Zhang <[email protected]>:
Dear Anjana,
tbtrans can only parallel over k-points. That is to say, the Nodes you
used should be less than the k-points you used.
Please refer to page 99 of the manual, there is some comment on this
issue under the option "TS.TBT.Npoints" .
Best.
Guangping
2011-07-28
Guangping Zhang
发件人: [email protected]
发送时间: 2011-07-28 03:33
主 题: [SIESTA-L] problem while executing with tbtrans
收件人: [email protected]
Dear Siesta Users
I am trying to execute the examples provided on this following link.
http://www.mail-archive.com/[email protected]/msg01041.html
I have a problem getting the output using the tbtrans. Whenever I
execute it. a warning appear on the output file which is as follows.
WARNING !!!!
>> Number of Nodes greater than number of k-Points !!
>> Some nodes will be innactive in k-point parallelization !
I use a pbs script to submit the job which is as follows
#!/bin/bash
#PBS -q workq
#PBS -l nodes=1:ppn=4
#PBS -l walltime=72:00:00
#PBS -o output_TB
#PBS -j oe
#PBS -N siesta-Job
cd /work/anjana/transiesta/IV
for i in `seq -w 0.0 0.1 1.8`
do
cd $i
export NPROCS=`wc -l $PBS_NODEFILE |gawk '//{print $1}'`
mpirun -machinefile $PBS_NODEFILE -np $NPROCS
/home/anjana/siesta-3.0-rc2/Util/TBTrans/tbtrans < INPUT.fdf >
OUTPUT_tbt.fdf & #
cd ..
done
what changes do I have to make to get the output...
Anjana
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.
--
Hongyi Zhao <[email protected]>
Institute of Semiconductors, Chinese Academy of Sciences
GnuPG DSA: 0xD108493
