diff --git a/docs.it4i/general/capacity-computing.md b/docs.it4i/general/capacity-computing.md index b4f4759b9e6ebce5ee7115f33d67727d1029d677..e65ea43f34ebf3a01f0f7b13201741873b9de19e 100644 --- a/docs.it4i/general/capacity-computing.md +++ b/docs.it4i/general/capacity-computing.md @@ -8,15 +8,16 @@ achieving the best runtime, throughput, and computer utilization. This is called However, executing a huge number of jobs via the Slurm queue may strain the system. This strain may result in slow response to commands, inefficient scheduling, and overall degradation of performance -and user experience for all users. We **recommend** using [**Job arrays**][1] or [**HyperQueue**][2] to execute many jobs. +and user experience for all users. +We **recommend** using [**Job arrays**][1] or [**HyperQueue**][2] to execute many jobs. There are two primary scenarios: -1. Numeber of jobs < 1500, **and** the jobs are able to utilize one or more full nodes: +1. Numeber of jobs < 1500, **and** the jobs are able to utilize one or more **full** nodes: Use [**Job arrays**][2]. The Job array allows to sumbmit and control many jobs (tasks) in one packet. Several job arrays may be sumitted. -2. Number of jobs >> 1500, **or** the jobs only utilze a few cores each: +2. Number of jobs >> 1500, **or** the jobs only utilze a **few cores/accelerators** each: Use [**HyperQueue**][1]. HyperQueue can help efficiently load balance a very large number of (small) jobs amongst available computing nodes. HyperQueue may be also used if you have dependenices among the jobs.