<p>Thanks for your reply. It seems that interactive use is your main goal. However, wouldn't out be better to submit every task separately using qsub (instead of submitting ipengines)?</p>
<p>Thanks,<br>
Chris<br>
</p>
<div class="gmail_quote">On Jan 17, 2011 2:59 AM, "Brian Granger" <<a href="mailto:ellisonbg@gmail.com">ellisonbg@gmail.com</a>> wrote:<br type="attribution">> Hi,<br>> <br>> On Thu, Jan 13, 2011 at 7:41 AM, Chris Filo Gorgolewski<br>
> <<a href="mailto:chris.gorgolewski@gmail.com">chris.gorgolewski@gmail.com</a>> wrote:<br>>> Hi,<br>>> I have recently played with ipython on our SGE cluster. I was<br>>> surprised to discover that ipython does not use qsub to submit every<br>
>> job, but submits prespecified number of ipengines as jobs. Those I<br>>> presume run indefinitely and accept ipython tasks. This setup seem to<br>>> have two major drawbacks:<br>> <br>> Yes, this is a correct description of what happens.<br>
> <br>>> 1) my cluster have nodes with different max job time. Depending what<br>>> you specify in the qsub option the job gets send to different node.<br>>> The limit is 48h. This means that after 48h (assuming that I use a<br>
>> custom submit script with this option) all of my engines will be<br>>> killed and ipython will stop receiving jobs?<br>> <br>> Yes, that is right.<br>> <br>>> In other words I cannot<br>>> run a set of jobs that would run longer than two days using ipython?<br>
> <br>> Yep, there is no way of getting around the limitations/constraints of<br>> the queues.<br>> <br>>> Additionally if I decide to specify max job time 48h I will most<br>>> likely wait longer for the appropriate nodes to become free which is<br>
>> not really necessary when my atomic jobs run much faster.<br>> <br>> Yep, such is life on shared clusters with batch system :(<br>> <br>> What about just firing up an EC2 cluster using startcluster?<br>
> <br>>> 2) I need to specify how many engines I want to use. Assuming i want<br>>> my set of jobs to be done as quickly as possible I should specify a<br>>> number that would be bigger than the number of available nodes. This<br>
>> means that in many situations I will spawn way too many ipengines that<br>>> will just sit there doing nothing. This solution seems to lack<br>>> scalability.<br>>><br>>> Or maybe I am using ipython/SGE in a wrong way?<br>
> <br>> From what we have said, I think you are using ipython/sge in the right<br>> manner, you are just running into the fact that batch systems are not<br>> setup for truly interactive usage.<br>> <br>> Cheers,<br>
> <br>> Brian<br>> <br>>> Best regards,<br>>> Chris<br>>> _______________________________________________<br>>> IPython-User mailing list<br>>> <a href="mailto:IPython-User@scipy.org">IPython-User@scipy.org</a><br>
>> <a href="http://mail.scipy.org/mailman/listinfo/ipython-user">http://mail.scipy.org/mailman/listinfo/ipython-user</a><br>>><br>> <br>> <br>> <br>> -- <br>> Brian E. Granger, Ph.D.<br>> Assistant Professor of Physics<br>
> Cal Poly State University, San Luis Obispo<br>> <a href="mailto:bgranger@calpoly.edu">bgranger@calpoly.edu</a><br>> <a href="mailto:ellisonbg@gmail.com">ellisonbg@gmail.com</a><br></div>