[mpich-discuss] Hydra issues

[Pavan Balaji]

Scott,

Ok, I did some poking around with this. It looks like the PMI 
initialization already sets the stdout buffering to NULL. So, for all 
MPI programs there shouldn't be any buffering problem. I verified it 
with a simple MPI program. This, obviously, will not help non-MPI 
programs, but I don't think that's your concern here.

Are you seeing the buffering problem for MPI programs as well?

  -- Pavan

On 08/26/2009 04:00 PM, Ralph Butler wrote:
> Keep in mind that the LD_* magic vars have restrictions for pgms that 
> use setuid stuff.
> This was an issue for mpd, but may not be for you.
> 
> On WedAug 26, at Wed Aug 26 2:39PM, Pavan Balaji wrote:
> 
>>
>>>> However you could use one of the various workarounds for this, such 
>>>> as an LD_PRELOADed setvbuf call: 
>>>> http://lists.gnu.org/archive/html/bug-coreutils/2008-11/msg00164.html
>>> This does not change the behavior.
>>> I am still stumped as to why there is no delay when using persistent 
>>> (launch-mode=2) versus a delay with no proxies (launch-mode=1).
>>
>> This works for me. We need to figure out how to make this portable now.
>>
>> -- Pavan
>>
>> -- 
>> Pavan Balaji
>> http://www.mcs.anl.gov/~balaji
> 

-- 
Pavan Balaji
http://www.mcs.anl.gov/~balaji