Problem for the upwelling case when turning OpenMP on

Discussion on computers, ROMS installation and compiling

Moderators: arango, robertson

Post Reply
Message
Author
jerry19890126
Posts: 13
Joined: Wed Jul 27, 2011 3:35 am
Location: IOCAS

Problem for the upwelling case when turning OpenMP on

#1 Unread post by jerry19890126 »

There is no problems when turned off OpenMP ( comment "export USE_OpenMP=on # shared-memory parallelism" in build.bash script): I got compiled oceanS file, and I run the model successfully afterwards. But when I wanted to turn it on, strange things happened. Although I got the compiled oceanO file, but when I typed "./oceanO<ocean_upwelling.in" , it prompted the error informations as below:

STEP Day HH:MM:SS KINETIC_ENRG POTEN_ENRG TOTAL_ENRG NET_VOLUME
C => (i,j,k) Cu Cv Cw Max Speed

0 0 00:00:00 0.000000E+00 6.585677E+02 6.585677E+02 3.884376E+11
(00,00,00) 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
DEF_HIS - creating history file, Grid 01: ocean_his.nc
WRT_HIS - wrote history fields (Index=1,1) in record = 0000001

Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.

Backtrace for this error:
#0 0x7F258AC4AE08
#1 0x7F258AC49F90
#2 0x7F258A15349F
#3 0x469188 in output_
#4 0x447B4D in main3d_
#5 0x7F258A70FCBE
#6 0x402871 in __ocean_control_mod_MOD_roms_run


I used the latest gfortran. Could anyone so kind to help me?

MROCIO
Posts: 7
Joined: Tue Aug 24, 2010 12:08 am
Location: CICATA-IPN

Re: Problem for the upwelling case when turning OpenMP on

#2 Unread post by MROCIO »

Dear jerry19890126;
I'm having a similar problem running the model :oops: :oops: , could you be so kind to tell me, how did you fix this error?
lapocean01@lapocean01-XPS-8700:~/EXAMPLEUP$ ./oceanG < ocean_upwelling.in > OUT


Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.
Backtrace for this error:
#0 0x7F64EFB4FE08
#1 0x7F64EFB4EF90
#2 0x7F64EF05849F
#3 0x5229AA in output_ at output.f90:211 (discriminator 8 )
#4 0x4C6C38 in main3d_ at main3d.f90:225
#5 0x402F97 in __ocean_control_mod_MOD_roms_run._omp_fn.1 at ocean_control.f90:161
#6 0x7F64EF614CBE
#7 0x402D6F in __ocean_control_mod_MOD_roms_run at ocean_control.f90:160
#8 0x40248C in MAIN__ at master.f90:86
Floating point exception (core dumped)


thanks in advance!!!

jerry19890126
Posts: 13
Joined: Wed Jul 27, 2011 3:35 am
Location: IOCAS

Re: Problem for the upwelling case when turning OpenMP on

#3 Unread post by jerry19890126 »

I have solved this problem :D , there are 2 ways:

Way 1:
edit the ocean_upwelling.in file,
modify the value of "NQCK" (with a default value of 0) to an integer larger than 0 @around line 221


THIS WILL OUTPUT A QUICKSAVE FILE, ALTHOUGH I DON'T KNOW WHAT IT IS

Way 2:
sudo gedit ROMS\Nolinear\output.F @line 228
Change "IF (LwrtQCK(ng)) THEN" to:
"IF (LwrtQCK(ng) .and. LdefQCK(ng)) THEN"

sudo gedit ROMS\Modules\mod_scalars.F
add a line @ line 2426:
"LdefQCK(ng)=.FALSE."


THIS WILL NOT OUTPUT A QUICKSAVE FILE, BUT THE SOURCE CODE ARE MODIFED.

I wonder whether it is related to a program bug, because in the mod_scalars.F, the parameter LdefQCK(ng) is not given an initial value.

MROCIO
Posts: 7
Joined: Tue Aug 24, 2010 12:08 am
Location: CICATA-IPN

Re: Problem for the upwelling case when turning OpenMP on

#4 Unread post by MROCIO »

Thank you very much is running.

I used the way number 1, with upwelling case. Tomorrow I will try this option with a different case.

Thanks again :D

User avatar
m.hadfield
Posts: 521
Joined: Tue Jul 01, 2003 4:12 am
Location: NIWA

Re: Problem for the upwelling case when turning OpenMP on

#5 Unread post by m.hadfield »

The quick-save output facility introduced in ticket 704 had a bug that was fixed in ticket 708:

https://www.myroms.org/projects/src/ticket/704
https://www.myroms.org/projects/src/ticket/708

User avatar
AbhiPasula
Posts: 9
Joined: Thu Dec 12, 2019 4:10 pm
Location: Indian Institute of Science

Re: Problem for the upwelling case when turning OpenMP on

#6 Unread post by AbhiPasula »

Hi,

I am running 4DVAR for the Indian Region.

While running 4DPSAS with same cpp options as IS4DVAR NLM model is running fine. But with IS4DVAR, model is terminated abnormally after zeroth time step.

TIME-STEP YYYY-MM-DD hh:mm:ss.ss KINETIC_ENRG POTEN_ENRG TOTAL_ENRG NET_VOLUME
C => (i,j,k) Cu Cv Cw Max Speed

0 2016-06-02 00:00:00.00 1.396706E-02 2.330166E+03 2.330180E+03 4.754942E+14
(188,115,50) 4.046671E-02 4.194327E-02 0.000000E+00 1.150302E+00
Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.

Backtrace for this error:
#0 0x7ff84a154cd1 in ???
#1 0x7ff84a153ea5 in ???
#2 0x7ff8494b020f in ???
#3 0x55f760d8255f in output_
at /media/quest/lion1/AS/0804/Build_roms/output.f90:257
#4 0x55f760c3cb0b in main3d_
at /media/quest/lion1/AS/0804/Build_roms/main3d.f90:267
#5 0x55f760be89d9 in __ocean_control_mod_MOD_roms_run
at /media/quest/lion1/AS/0804/Build_roms/ocean_control.f90:401
#6 0x55f760bd93bc in ocean
at /media/quest/lion1/AS/0804/Build_roms/master.f90:86
#7 0x55f760bd941a in main
at /media/quest/lion1/AS/0804/Build_roms/master.f90:50
Floating point exception (core dumped)

Please suggest where is the issue?

Post Reply