Ocean Modeling Discussion

azhang wrote:
I switched to ifort 8.0. ROMS3.0 was successfully compiled.
I guess there might be bugs in ifort 9.0 Fortran compiler.

If your license allows, instead of going to older versions, try to get the latest 10.0 or 10.1 compilers. They do a good job and allow the entire source to be ...

azhang wrote:
I got the following error message,

IPO link: can not find "("
ifort: error: problem during multi-file optimization compilation (code 1)

kate wrote:
Is that an error message from the compiler?

It is, and it comes from the link step. It is most likely due to the use of the "-ipo ...

[...] Oh, for the good old days when the Cray compiler would let you dump core during the actual operation that caused the NaN/Inf so you could see what was going on.

You can blame such default behavior (continuing to run way after infinities and NaNs have been produced by operations) on the IEEE ...

isopycnal wrote:
thank you very much if you can do the similar test.

I will be glad to run your test case, but I will need to have
copies of your .nc input files. Next week I will be traveling
to Europe (to visit an institution where ROMS is also used),
but the following week I should be able to ...

when I tried to use fewer cpus(originally 16 cpus, now 8 cpus), it works using oceanM. Is this the reason for the error? I really don't know too much on the linux system.

I don't understand why the error would occur with *fewer*
processors, since in that case one would expect the application
to ...

isopycnal wrote:
> [...]
> MPI: On host a3700, Program /lun/home/liuzl/latest_roms/oceanM, Rank 11, Process 22290 received signal SIGSEGV(11)
> [...]

SIGSEGV(11) occurs for one of two reasons: a bad memory access
(such as using out-of-range indices or a bad pointer), or trying to grow
the stack ...

Hernan,

Perhaps a check could be inserted somewhere in the make scripts
that terminates a build if USE_MPI and USE_OpenMP are both set?

Saludos,
--
Gerardo Cisneros

linzhenhua wrote:
> I have access to the same SGI as you.I just try ROMS on it.
> At first I'm lazy to compile the netcdf library myself and
> download the binary version from unidata,I meet with similiar
> fortcom problems as you,but with a different fortran 90 file.
> Then I compile the netcdf ...

Zhouwei,

Judging from your post, it looks like you're using an Altix 330
server. I will assume you are using Intel Fortran 9.x compilers.
I use a modified Linux-ifort.mk file. The differences are the
following (Linux-ifort8.mk is the original Linux-ifort.mk file;
Linux-ifort9.mk is mine):

nuke[104 ...

I ran the BENCHMARK1 and BENCHMARK2 test cases on 32 cores of an Altix 4700 with dual-core Itanium2 ("Montecito") processors (1.6GHz, 9MBL3 cache per core), and the BENCHMARK3 test case on 64, 128 and 192 cores of the same system.

The benchmarks ran pretty much out of the box , with minor changes ...

Thanks for your replay,Kate. It seems you are right that the problem lies in the optimizer switch -O3. I had the same error using -O2, but succeeded in compiling without using any of -O2 or -O3. However I ran into the same situation as in the first post. I wonder if the 'limit' you referred to is ...

The meaning of ROMS in Spanish is given as

Sistema Regional de Modelado Oceánico

The question is, what is it that's regional, the system or the modeling? If the latter, then a more appropriate translation would be:

Sistema de Modelado Oceánico Regional