array scalars%sc_w is out of range?

Message

FengZhou · #1 Unread post by **FengZhou** » Tue Jan 29, 2008 8:48 am

Hi, all,

I just tried ROMS3 and compiled the codes successfully with Upwelling case and a realistic case. But there is an error during executing the latter. The message is :

0: Subscript out of range for array scalars%sc_w (set_scoord.f90: 49)
subscript=0, lower bound=0, upper bound=-1, dimension=1

my set-up is
5.0000E+00 theta_s S-coordinate surface control parameter.
4.0000E-01 theta_b S-coordinate bottom control parameter.
4.000 Tcline S-coordinate surface/bottom layer width (m) used
in vertical coordinate stretching.
While in ROMS2.2, there is no such error.

Do you have any clue?

Thanks in advance!

zhou

kate · #2 Unread post by **kate** » Tue Jan 29, 2008 7:51 pm

ROMS 3 needs you to specify the grid dimensions in the input file: Lm, Mm, and N. What do you have for N?

FengZhou · #3 Unread post by **FengZhou** » Wed Jan 30, 2008 12:52 am

hi,Kate,

Thanks for your reply.
My grid dimension is as following( need only to input into External/byesea.h in Roms3 ?):
Lm == 167 ! Number of I-direction INTERIOR RHO-points
Mm == 243 ! Number of J-direction INTERIOR RHO-points
N == 30 ! Number of vertical levels

while in ROMS2, they were written in Modules/mod_param.F, right?

hmin = 8
hmax= 2500

zhou

kate · #4 Unread post by **kate** » Wed Jan 30, 2008 2:21 am

This wasn't the PGI compiler, was it? I had something like that happen to me, where it forgot about the allocated memory. Anyone have a clue?

FengZhou · #5 Unread post by **FengZhou** » Wed Jan 30, 2008 6:19 am

Yes, I use PGI compiler.

zhang · #6 Unread post by **zhang** » Thu Jan 31, 2008 3:17 pm

which version of pgi are you using?

kate · #7 Unread post by **kate** » Thu Jan 31, 2008 6:57 pm

I have PGI 7.0.2 and see this with my most complex domain. Upwelling runs fine, as do all three sizes of Benchmark. Sorry I haven't pushed on it further, but we've had Pathscale and Sunstudio longer and I've got other issues to solve. In the debugger, I can see the memory for scalars%sc_w where it gets allocated, then it's not there later, when trying to write to it the first time.

FengZhou · #8 Unread post by **FengZhou** » Fri Feb 01, 2008 9:52 am

zhang wrote:which version of pgi are you using?

my pgi is 7.0.3.

other information:
Model Input Parameters: ROMS/TOMS version 3.0
Wednesday - January 30, 2008 - 3:19:38 PM
-----------------------------------------------------------------------------

Bohai-YellowSea-EastChinaSea, 1/12 degree

Operating system : Linux
CPU/hardware : i686
Compiler system : pgi
Compiler command : mpif90
Compiler flags : -g -C -Mfree

FengZhou · #9 Unread post by **FengZhou** » Sat Feb 02, 2008 11:35 am

Hi, all,

problem is probably fixed. I forgot to define the nudging block for the realistic case. Its running just by undefining the nudging items. Now I am tracking on...

Thank you all.