aura.fi.muni.cz is available to employees and FI students for longer or more demanding calculations. Access to the machine is automatically generated by FI and PhD students; to other students only for study and research purposes at the request of the bachelor or master thesis supervisor,
Way of work
We do not currently consider introducing a queuing system for long-running tasks. It is sufficient if
you run low-priority running processes (10-19, 19 is the lowest). Long-term processes are those that run for over an hour. For example, use the command
nice ./váš_program . You can also restrict the program to use only some processors with a command
taskset . Startup example with minimum priority on 11th to 15th processor:
nice -n 19 taskset -c 10-14 ./váš_program
You can use the command to change the priority of the already running process
renice , which can be applied to threads, processes, or process groups (see command page). However, you need to think that the process can run in multiple threads, and changing the priority for one process can change the priority of just one thread. For example, a listing of all the threads of my processes, including priorities, can be obtained from the following command:
ps x -Lo pgid,pid,tid,user,nice,tty,time,args
You can run short-term processes or interactive debugging of your programs at a normal priority.
In the event that your process does not meet the priority constraints and will use a large amount of computing performance, the lowest priority will be set for all your processes to prevent other users from limiting it. For a repeated or more serious violation of this policy, you may temporarily block the faculty account.
Only in the case of higher interest in the computing server services will we introduce a front-end system, where the user will be able to determine in advance how many processors and memory his job will use and estimate how long they will run.
For temporary data that should be available locally quickly, directories are available
/var/tmp on a fast RAID-10 volume. To use this space, store your data in the directory with your login. Note, however, that older data stored here is automatically deleted (u
/tmp in the age of days, u
/var/tmp in the order of months). In addition, the user can use his faculty home directory. You need to
export for server
Aura NFS protocol.
Aura now running on the
HP DL980 G7 platform in the following configuration:
- 8 eight-core 64-bit Intel Xeon X7560 2.26 GHz processors ( 64 cores in total).
- 448 GiB DDR3 RAM
- Gigabit Ethernet connection
- 8 disk capacity of 300 GB, 10k RPM, connected to HW RAID-10, ie 1.2 TB of logic capacity
The server is installed on
Linux - Red Hat Enterprise Linux 6. There is also a teaching software for 64-bit and 32-bit Linux tree