[MITgcm-support] Bug R_X86_64_PC32

[Nataliya Stashchuk]

Hi Martin,

Sorry for delay with answer but it seems to me that I found a clue with
R_X86_64_PC32.

First of all I will remind to you what we did to solve this problem.

1. You recommend me to put FFLAG -mcmodel=medium to my optfile. This did
not help.

2. I sent you my grid numbers with packets which I use in my code. You
tried the same code on your old machine. It worked.

3. You sent me your code. I tried to compile it on my machine. I got
R_X86_64_PC32 again.

4. Then I tried your code on my old machine. It worked!

I think that the answer is in restrictions of compiler g77 which was
used in computers with different architecture (old is with 32 bit, new
is with 64 bit). I can compile the code with g77 for 32 bit
architecture, but can not compile the same code when I use 64 bit
architecture. 

What compiler is better to use in computer with 64 bit?

Regards,
Nataliya



-----Original Message-----
From: mitgcm-support-bounces at mitgcm.org
[mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Martin Losch
Sent: 06 September 2006 12:57
To: mitgcm-support at mitgcm.org
Subject: Re: [MITgcm-support] Bug R_X86_64_PC32

MPI  will only make things more complicated to debug.
can you list exactly what you did to download and compile the code  
(as if in a script) and provide you code directory. If we cannot  
reproduce your problem, there is little we can do, is there? Right  
now, I don't know what is going wrong.

Martin
PS. I used only 1CPU
On Sep 6, 2006, at 12:20 PM, Nataliya Stashchuk wrote:

> Thanks a lot, Martin.
>
> I did use your code and tried to compile with it.
> However, I still have "R-X86_64_PC32 against symbol ..." with your  
> code.
>
>
> First of all, how many processors are on your PC, one or two?
> Do you use the multiprocessor regime? Probably here we should find the
> answer?
>
> In my case I've two-processor machine with 2Gb RAM and something is
> wrong with compatibility. Probably I should use the MPI option,  
> what do
> you think?
> If yes, could you recommend me which MPI-enable compiler is better for
> my case.
>
> Nataliya
>
>
>
>
>
> -----Original Message-----
> From: mitgcm-support-bounces at mitgcm.org
> [mailto:mitgcm-support-bounces at mitgcm.org] On Behalf Of Martin Losch
> Sent: 06 September 2006 10:31
> To: mitgcm-support at mitgcm.org
> Subject: Re: [MITgcm-support] Bug R_X86_64_PC32
>
> Nataliya,
>
> this is my complete code directory (now with non-hydrostatic code,
> pp81, and obcs and it still works). Please try it out on your machine
> with this code directory.
>
> Martin
>
>
> _______________________________________________
> MITgcm-support mailing list
> MITgcm-support at mitgcm.org
> http://mitgcm.org/mailman/listinfo/mitgcm-support

_______________________________________________
MITgcm-support mailing list
MITgcm-support at mitgcm.org
http://mitgcm.org/mailman/listinfo/mitgcm-support