Action
To schedule research computing work using SLURM, follow the instructions below.
...
All jobs on the general purpose cluster request resources via SLURM. SLURM, is open source software that allocates resources to users for their computations, provides a framework for starting, executing and monitoring compute jobs, and arbitrates contention for resources by managing a queue of pending work. SLURM is widely used in the high performance computing (HPC) landscape and it is likely you will encounter it outside of our systems. For more information please see https://slurm.schedmd.com/
General Purpose Computing
...
Code Block | ||||
---|---|---|---|---|
| ||||
-bash-4.2$ sinfo -p batch -o "%n, %a, %C, %e, %O" | sort HOSTNAMES, AVAIL, CPUS(A/I/O/T), FREE_MEM, CPU_LOAD uagc19rhea-0109, up, 20/380/096/4096, 93618N/A, 0.01 uagc19-02N/A rhea-10, up, 120/280/096/4096, 54256N/A, 15.74N/A uagc19-0301, up, 2/38/0/40, 93621, 0.00 uagc19-02, up, 12/28/0/40, 54240, 15.90 uagc19-03, up, 14/26/0/40, 6793267920, 1517.8712 uagc20uagc19-0504, up, 3236/324/0/6440, 9815875889, 3216.1060 uagc20-0601, up, 320/3264/0/64, 97454189368, 320.0300 uagc20-0702, up, 20/6264/0/64, 188180189359, 0.00 uagc20-0803, up, 320/3264/0/64, 94103189367, 320.1000 uagc20-1304, up, 240/4064/0/64, 175916189461, 0.0100 uagc20-1005, up, 6432/032/0/64, 900298151, 6432.3009 uagc20-1206, up, 6432/032/0/64, 931297446, 6432.2310 uagc19uagc20-0407, up, 362/462/0/4064, 75889188191, 160.6000 uagc20-0908, up, 4032/2432/0/64, 18020894065, 4832.2111 uagc21uagc20-0609, up, 8040/024/0/8064, 339206180208, 2548.0121 uagc21uagc20-0710, up, 8064/0/0/8064, 3804458985, 1364.35 uagc21uagc20-0811, up, 800/064/0/8064, 339179189303, 250.00 uagc20-0112, up, 64/0/64/0/64, 1893699337, 064.0024 uagc20-0213, up, 024/6440/0/64, 189358176151, 0.00 uagc20-0314, up, 0/64/0/64, 189372189364, 0.00 uagc20-0415, up, 0/64/0/64, 189462189343, 0.0002 uagc20uagc21-1101, up, 80/0/64/0/6480, 189304278987, 01.0047 uagc20uagc21-1402, up, 020/6460/0/6480, 189369371334, 0.0050 uagc20uagc21-1503, up, 0/6480/0/6480, 189345381238, 0.00 uagc21-0104, up, 0/80/0/80, 278987381046, 10.4700 uagc21-0205, up, 0/80/0/80, 371335290550, 0.0100 uagc21-0306, up, 80/0/80/0/80, 381238339206, 025.0001 uagc21-0407, up, 80/0/80/0/80, 381046247070, 04.0017 uagc21-0508, up, 80/0/80/0/80, 283133338115, 425.0203 uagc21-09, up, 0/80/0/80, 380172, 0.75 uagc21-10, up, 0/80/0/80, 380682, 0.20 uagc21-11, up, 0/80/0/80, 381268381265, 0.00 uagc21-12, up, 0/80/0/80, 303282260703, 50.2700 |
Info |
---|
Note that %a reports CPUS as allocated/idle/other/available. In this example, uagc20-10 has all of it's threads allocated (64 out of 64), and is showing a CPU load of 64.30 (or that 64.30 threads are active). Whereas, many of the other nodes have lower utilization. We can use this information to make smart decisions about how many resources we request. |
...
Info |
---|
batch has some important restrictions. A job can only request 3 nodes and will run for 14 days before being automatically terminated. If you need an exception to this rule, please contact askIT@albany.edu |
Request access to more nodes, or a longer time limit
On a case by case basis, ARCC will grant users temporary access to more than the default job limitations. Please contact askIT@albany.edu if you would like to request access to more nodes, or a longer time limit.
...
Info |
---|
This job ran on rhea-09, and it's max memory size was ~52 GB. That that I requested 60000MB, so I could refine this job to request slightly less memory. It ran for 14:50:14 and used about 350 CPU hours. |
Restrict a job to a certain CPU architecture
Use the --constraint flag in #SBATCH. To few available architecture on individual nodes use scontrol show node
...