[mpich-devel] HYDRA_TOPO_DEBUG=1 needs a line break

[Kenneth Raffenetti]

Just pushed a commit to master that should support many processing 
elements. Thanks for the report.

Ken

On 11/18/2015 10:24 AM, Kenneth Raffenetti wrote:
> The newline is added after all the hwthreads are added to the string -
> line 469. You might be overrunning the max string length, as that is not
> well handled. Can you try doubling DBG_STR_LENGTH at the top of the file?
>
> Ken
>
> On 11/18/2015 09:09 AM, Jeff Hammond wrote:
>> I followed the example on
>> https://wiki.mpich.org/mpich/index.php/Using_the_Hydra_Process_Manager#Examples
>>
>> but got a huge mess rather than the output shown on the Wiki.
>>
>> Is the lack of line break here intentional?  I think it would be more
>> useful if it did.
>>
>> jhammond at nid00218:~/> HYDRA_TOPO_DEBUG=1 mpiexec -rmk slurm -n 8
>> -bind-to core true
>>
>> process 0 binding: 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>> 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 2
>> binding: 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>> 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 3 binding: 0 0 0 1
>> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
>> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 5 binding: 0 0 0 0 0 1 0 0 0 0 0 0
>> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
>> 0 0 0 0 0 0 0process 6 binding: 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
>> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>> 0process 1 binding: 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>> 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 4
>> binding: 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
>> 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0process 7 binding: 0 0 0 0
>> 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
>> 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0jhammond at nid00218:~/>
>>
>> The offending line appears to be 463 in
>> src/pm/hydra/tools/topo/hwloc/topo_hwloc.c, but I don't know if the "\n"
>> goes there or in the definition of HYDU_snprintf.
>>
>> Thanks,
>>
>> Jeff
>>
>> --
>> Jeff Hammond
>> jeff.science at gmail.com <mailto:jeff.science at gmail.com>
>> http://jeffhammond.github.io/
>>
>>
>> _______________________________________________
>> To manage subscription options or unsubscribe:
>> https://lists.mpich.org/mailman/listinfo/devel
>>
> _______________________________________________
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/devel