Anonymous @ on
My job has been in the queue for a week.
My nanoddscat+ session has been running for over a week.
Any idea why it is taking that long. The run should finish in 2 days.
Anonymous @ on
My job has been in the queue for a week.
My nanoddscat+ session has been running for over a week.
Any idea why it is taking that long. The run should finish in 2 days.
0 Like 0 Dislike
Anonymous @ on
How many dipoles are in your structure? Usually the limit is a few hundred thousands. Otherwise long convergence time may be due to extremely sharp corners somewhere in your structure. If neither of these is a problem then let me know and I will see if there is a problem on the server side.
Reply Report abuse
Please login to answer the question.
0 Like 0 Dislike
AbderRahman N Sobh @ on
Hi Anon,
This is the same answer that I provided for another user in question 1671:
Jobs are queued to different processing clusters based on the amount of memory and processing that will be required for the simulation.
There are typically a lot of smaller size simulations that are taking up all the small size processing units in the cluster we use and some are doing things like calculating 2-D periodic simulations. These things don't technically require significantly more memory or processing, thus even though they require more time to work they go to the same queue and hold up resources for long periods of time. Resource management is currently an issue that we are trying to resolve as our user base has been increasing.
At the same time, our heavier resource using CPUs are often sitting somewhat idle- though ready for anyone who is doing something more serious and large scale (i.e. high dipole count + generating an E-field).
It is best to run small size problems in the "local" mode instead of "remote" if possible. Otherwise, if you have time on your hands and the local simulation was not able to complete your small size problem, it's best to just select for the tool to send you an email when it is done.
Reply Report abuse
Please login to answer the question.