Skip to content
Snippets Groups Projects

Excessive cpu load

Last edited by Alex Wiens

  • Description: A overly high CPU load, i.e., the number of processes and threads waiting for CPU time, can indicate one of the following:

    • wrong parallelization settings, defaults, or automatisms in applications for the number of MPI ranks or number of threads
    • unintentionally using additional parallelization layers like in underlying libraries like BLAS
    • blocking of processes due to IO or network activity

  • Criterion:

    • CPU load one-minute average is significantly higher than the number of CPU cores or hardware threads allocated by the job on the node

  • Works for shared jobs with existing metrics:

    • yes, if not the load_one is used but the CPU-core-specific cpu_load_core.

  • Possible false positives: *

  • Possible cures/workarounds:

    • checking parallelization settings
Eingabe:
* Metrik cpu_load im node scope
* Metadatum numHwthreads
* Parameter threshold für einen core

Regel:
load_mean      = cpu_load.mean('all')
load_threshold = job.numHwthreads * threshold
highload       = load_mean > load_threshold

Ausgabe: highload ist True

Comments

    Please register or sign in to add a comment.