Code

Enable Nvidia GPU Support#57

Code

Merged

NucciTheBoss merged 15 commits intocharmed-hpc:maincharmed-hpc/slurm-charms:mainfrom dsloanm:gpudsloanm/slurm-charms:gpuCopy head branch name to clipboardJan 13, 2025

+473-43

Commits

Commits on Dec 20, 2024

• chore(all): bump slurmutils to 0.11.0 and add jsonschema dependency.

  dsloanm

  committed
  a5a84c4

  Copy full SHA for a5a84c4

• feat(slurmd): add GPU detection and driver installation.

  dsloanm

  committed
  cd248fe

  Copy full SHA for cd248fe

• feat(slurmctld): add GPU and other GRES device support.

  dsloanm

  committed
  2c95313

  Copy full SHA for 2c95313

Commits on Jan 7, 2025

• tests(slurmd): add unit tests for GPU util

  dsloanm

  committed
  1bcfa0f

  Copy full SHA for 1bcfa0f

• feat(slurmd): account for Juju auto-upgrading base image

  Show description for 8fbf715
  dsloanm

  committed
  8fbf715

  Copy full SHA for 8fbf715

• refactor(slurmd): improve GPU detection comments and log messages

  dsloanm

  committed
  94c1ce3

  Copy full SHA for 94c1ce3

• chore(all): update slurm_ops to 0.12 and slurmutils requirement

  Show description for 5575440
  dsloanm

  committed
  5575440

  Copy full SHA for 5575440

Commits on Jan 8, 2025

• refactor(slurmd): improve comments and function names. Simplify

  Show description for 5ae35e1
  dsloanm

  committed
  5ae35e1

  Copy full SHA for 5ae35e1

• refactor(slurmctld): improve comments and function names

  dsloanm

  committed
  29e1169

  Copy full SHA for 29e1169

Commits on Jan 9, 2025

• refactor(slurmd): use list over set for GPU device numbers

  dsloanm

  committed
  738081e

  Copy full SHA for 738081e

• refactor(slurmd): use calculate_rs from slurmutils for ranges

  Show description for f85e6f7
  dsloanm

  committed
  f85e6f7

  Copy full SHA for f85e6f7

Commits on Jan 10, 2025

• feat(slurmd): use packaged ubuntu-drivers-common and pynvml

  dsloanm

  committed
  20b764f

  Copy full SHA for 20b764f

Commits on Jan 13, 2025

• refactor(slurmctld): improve function names and logging

  dsloanm

  committed
  d2f5a1a

  Copy full SHA for d2f5a1a

• refactor(slurmd): improve comments and logging

  dsloanm

  committed
  4cf8083

  Copy full SHA for 4cf8083

• refactor(slurmctld): remove unnecessary str conversion

  dsloanm

  committed
  fffc9d8

  Copy full SHA for fffc9d8