[Wien] =?iso-2022-jp?B?UmU6W1dpZW5dIFJlOiBbV2llbl0=?=

[tom_y at livedoor.com]

Dear All,

Thanks for the prompt reply.

I made futher tests taking into account of the comments. 

>Does lapw1 run in single mode ? LOPW errors often occur when your struct
>file is wrong and some atoms are specified twice.

Yes, serial mode lapw1 runs without problem, and k-point level parallel version also runs without problem.

>If yes, is your lapw1-version a recent one ? We have constantly updated
>the parallel lapw1 and fixed small bugs showing up only for some compilers.

I downloaded the files from the wien2k web site on 9th of July, 2003.
As far as I checked the update info., my source files for lapw1 is the latest.

>PS: I've got a new parallel hns.F from R.Dohmen with better parallelization
>and will put it on the web soon.

I'd like to try it. I appreciate if you put it on the web site sooner.
As I wrote before, our PC-cluster system is very homogeneous, which has
Pentium-4 based linux machines connected with gigabit network cards.
Operating system is redhat7.3, and compiler and library are ifc7 and mkl6.0.
mpich1.2.5 is used for MPI.

According to the Kevin's comments, I modified the lapw1para files.
'debug' is changed from 0 to 3.
When I run 'x lapw -p' after above modification in lapw1para,
I got following message on the screen.

%x lapw1 -p
starting parallel lapw1 at Wed Aug 20 18:17:24 JST 2003
->  starting parallel LAPW1 jobs at Wed Aug 20 18:17:24 JST 2003
Wed Aug 20 18:17:24 JST 2003 -> Setting up case mgo3 for parallel execution
Wed Aug 20 18:17:24 JST 2003 -> of LAPW1
Wed Aug 20 18:17:24 JST 2003 ->
Wed Aug 20 18:17:24 JST 2003 -> non sp
running LAPW1 in parallel mode (using .machines)
Granularity set to 1
Extrafine unset
Wed Aug 20 18:17:25 JST 2003 -> klist:       3
Wed Aug 20 18:17:25 JST 2003 -> machines:    earth51
Wed Aug 20 18:17:25 JST 2003 -> procs:       1
Wed Aug 20 18:17:25 JST 2003 -> weigh(old):  3
Wed Aug 20 18:17:25 JST 2003 -> sumw:        3
Wed Aug 20 18:17:25 JST 2003 -> granularity: 1
Wed Aug 20 18:17:25 JST 2003 -> weigh(new):  3
Wed Aug 20 18:17:25 JST 2003 -> Splitting mgo3.klist.tmp into junks
earth51:1 earth52:1 earth53:1 earth54:1
.machinetmp222
1 number_of_parallel_jobs
prepare 1 on earth51
Wed Aug 20 18:17:25 JST 2003 -> Creating klist 1
1 : 3k (earth51, 3)
Wed Aug 20 18:17:25 JST 2003 ->
Wed Aug 20 18:17:25 JST 2003 -> creating lapw1_1.def:
[1] 12014
[1]  + Done                          ( cd $PWD; $t $ttt; rm -f .lock_$lockfile[$
p] ) >> .time1_$loop
waiting for all processes to complete
Wed Aug 20 18:17:27 JST 2003 -> all processes done.
testerror lapw1_1
**  LAPW1 crashed!
cat: No match.
0.100u 0.070s 0:03.81 4.4%      0+0k 0+0io 13500pf+0w

Here I used the following .machines file.
-------------------------------------------------------------------
lapw0:earth51:1 earth52:1 earth53:1 earth54:1
3:earth51:1 earth52:1 earth53:1 earth54:1
granularity:1
-------------------------------------------------------------------

Also, I used lapw1_1.def instead of lapw1.def for direct submission as follows.

%mpirun -np 4 -machinefile m /home/wien2k/WIEN2k_03_3/lapw1_mpi
lapw1_1.def
 Using            4  processors, My ID =            0
 Using            4  processors, My ID =            2
 Using            4  processors, My ID =            1
 Using            4  processors, My ID =            3
FORTRAN STOP LOPW - Error
FORTRAN STOP LOPW - Error
FORTRAN STOP LOPW - Error
FORTRAN STOP LOPW - Error

Same errors "LOPW  - Error" appeared as the case for lapw1.def.

Do you have any suggestions to do at the next step?
Thank you very much for your help in advance.

sincerely,

Tom Yamamoto  


フレッツ始めるなら今 ！キャンペーン実施中！
               http://www.livedoor.com/flets/