siltide
Share on:
Enter Your Mastodon Instance
Copy the link below to share to Mastodon
https://terminaltrove.com/siltide/
A terminal fleet monitor for GPUs, NPUs, and 13 other accelerator kinds.
Script Preview
Review this script before executing in your terminal.
siltide is a terminal monitoring tool that monitors consumer GPU and data-center grade hardware and other computing accelerators through a TUI.
It supports accelerators from 15 different vendors and tracks device utilization, memory bandwidth, power, temperatures and health metrics where supported with visible graphs with each GPU. It also associates device processes with Kubernetes pods and Slurm jobs.
siltide keeps device history on disk which can be retrieved back through earlier readings or replaying a recorded session after a job has finished. It can also export snapshots as JSON or collect readings through its Prometheus endpoint integration to track usage in other monitoring tools.
Infrastructure engineers, machine-learning engineers and cluster administrators would benefit from using siltide to inspect accelerator usage and identify the jobs consuming resources.
