Singular
http://www.singular.uni-kl.de/forum/

Is there a cluster where one could use to run his codes ?
http://www.singular.uni-kl.de/forum/viewtopic.php?f=10&t=2513
Page 1 of 1

Author:  rambiz [ Thu Dec 10, 2015 7:31 pm ]
Post subject:  Is there a cluster where one could use to run his codes ?

Hi all,
I am trying a primary decomposition on a PC with an i7 8-cored cpu and 4000 MB RAM. I don't run out of memory but after 1 day or so, when I look at the performance of my PC under Start Task Manager---> Performance the cpu usage has gone down to a few % (Its maximum value being 12.5% because it is an 8 core cpu) and singular seems to be reading from and writing to the hard disc, because I am hearing some sounds. :) The memory used up is about 3800 MB which is again accessible under Task Manager---> Performance . Is it hopeless to leave it running like this?

Furthermore, I'd like to know if there is a cluster or really fast computer out there which is accessible for running similar calculations on?
I would be also be willing for pay. What system(s) do the guys from the university of Kaiserslautern use?

I really appreciate any comments with regards of running time of primary decomposition. I have integer coefficients but the solutions are obviously complex. Which ring of coefficients should I use? I am using the rational numbers (ring r=0,x1,x2,......,dp) right now.
I have a system of 5 equations in 8 unknown only and have already managed to get a reduced basis with respect to degree reverse lexicographical ordering (dp).

Greetings

Author:  steenpass [ Mon Dec 14, 2015 12:15 pm ]
Post subject:  Re: Is there a cluster where one could use to run his codes ?

Hi rambiz,

once Singular starts to use the swap space, it's usually hopeless to wait for the calculation to finish.

If you consider your system of equations in characteristic zero, then the ring you chose is correct.

Could you maybe send me the example (to [forumusername]@mathematik.uni-kl.de) or post it here? I could try to run it on one of the compute servers in Kaiserslautern. And, regarding the other thread you started, I also have a parallel implementation of the GTZ algorithm for primary decomposition.

Best regards,
Andreas

Author:  rambiz [ Fri Dec 18, 2015 7:03 pm ]
Post subject:  Re: Is there a cluster where one could use to run his codes ?

Hi all,
If I knew how to attach files here, I would make all the files with regards to the calculations I have done so far public.
I sent a mail to Andreas a minute ago and looking forward to hearing from him.
In case anybody else is interested, I would gleefully send him the problem description in a detailed manner, too.
Regards

Page 1 of 1 All times are UTC + 1 hour [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/