I've checked the code and, as suggested by  Mike, I think that the
problem is related to "unsigned long  int" type.
In fact, my laptop uses 32-bit unsigned long int, while the cluster
uses 64-bits for each value. The test that causes simulations to abort
checks if the current random seed is less than a particular thresold
value: with 64-bits, the random value is bigger than the thresold m1
(declared in RNG.cc), and the simulation fails.

Consequently, I've tried to compile ns-2.34 with the option -m32 but I
still have problems since the linking phase fails for incompatible
code (it seemes that I'm trying to linking 32-bit code with 64-bit
code).

Do you have any idea about this problem?
Do you know how to compile ns-2.34 to impose 32 bits data?

Thanks a lot.
Mario

On Mon, Apr 19, 2010 at 5:27 AM, Mike Lorenz
<[email protected]> wrote:
>
> Zitat von Mario Fanelli <[email protected]>:
>
>>
>> Dear all,
>>
>> my name is Mario Fanelli, and I'm writing this mail to ask your help
>> with a problem in ns 2.34.
>> I've already tested my simulations on my laptop, and everything seems
>> ok. But when I try to run the simulations on our cluster, all my
>> simulations end with this error:
>>
>> num_nodes is set 100
>> INITIALIZE THE LIST xListHead
>> INITIALIZE THE LIST xListHead
>> Starting Simulation...
>> ****************************************
>>
>> ERROR: Seed[0] >= 4294967087, Seed is not set.
>>
>> ****************************************
>>
>> Apparently, the error comes rng.cc class due to invalid seed initiialization.
>> The principal problem is that I do not have this error on my laptop,
>> and I cannot run any debug on the cluster due to the absence of any
>> debugging tool.
>> Every cluster node has a Intel(R) Xeon(R) CPU X3220 @ 2.40GHz
>> processor with 8 GB of ram.
>>
>> Do you have any ideas about the above problem?
>>
>> Best regards,
>> Mario Fanelli
>>
>>
>>
>
> which RNG do you use? and which seed do you set ?
> maybe, you use heuristic seed for your simulation. this one takes the
> actual time as seed via a linux/posix system-function (gettimeofday).
> and maybe on the other computer there's a difference for the data
> (32bit vs. 64bit) of this function
>
>
>



-- 
_________________________________________

Mario Fanelli, Ph.D. Student
DEIS-LIA - Università degli Studi di Bologna
Viale Risorgimento, 2 - 40136 Bologna (ITALY)
Ph.: (+39) 051 209 3541 Fax: (+39) 051 209 3073.
E-mail: [email protected] or [email protected]
_________________________________________

Reply via email to