Best Practices
This page lists important points concerning different aspects of MCMeSU supercomputer.
Computing strategy
- Test the scalability : before running your jobs on a lot of cores, make sure that the scalability is as expected.
Even if you have a program which has been proved scalable on other systems, try to test as much as possible the scalability of your program on MCMeSU : start with one CPU and increase gradually until you reach diminishing returns in execution time reduction. - Mind the walltime : do not risk to be stopped at the time limit. Most queues will indeed interrupt your job after 12 hours of computation (you can check that with squeue).
- Don’t be greedy : only request the number of CPUs you actually need to get results in a reasonable amount of time. Requiring too much resources and consuming a large share of computing resources will reduce your chance of being scheduled in a short time frame.
- Implement checkpointing : if your jobs should run for a long time (more than a week), think about implementing checkpointing. We can not guarantee a continuous availability of the servers (electricity or cooling problems, maintenance, unexpected crashes might occur). Making sure that your job frequently dumps its current state on the hard drive could eventually allow you to resume your work in case of interruption.
Files and data management
- Use SCRATCH storage for your computation, it is optimized for high-performance.
- Use STORE storage for the important results you want to keep, so that you don't leave inactive data in your SCRATCH.
- Backup your data : the SCRATCH, STORE and HOME directories are not backed up.
- Do not let your job write heavy output to your HOME directory. Not only your quota will be exceeded quickly, but you will experience bad performance in comparison of SCRATCH storage.
- Respect your quota : 100 GB per user on HOME. Do not leave inactive data in your SCRATCH (arbitrary purge of unused data may happen without notice in case of overload).
Security concerns
- Keep your data secure : MCMeSU supercomputer is secured, but often scanned on the network by would-be intruders. Although highly improbable, we can not guarantee that a data thievery won’t happen.
- Strong personal password : Your password must be strong/long/complex enough, do not share it with others (DSI policies).
- Respect Sorbonne University IT policy
- Keep the software on your laptops/PC up to date.
- Keep in mind that the Network traffic is monitored by the DSI and the MeSU administration team.