FreeLLMNVIDIA releases open-source gds-diag for GPUDirect Storage diagnostics
The new gds-diag utility helps admins assess and optimize GPUDirect Storage on Linux systems.
NVIDIA announced gds-diag, an open-source command-line utility that checks GPUDirect Storage environments and suggests improvements.
The tool runs on Linux with Python 3.8 or newer and does not depend on GPU or CUDA Toolkit versions, though GPUDirect Storage itself requires a datacenter-class GPU and CUDA 12.2 or later.
gds-diag offers subcommands grouped into end-to-end checks, pre‑install validation, post‑install verification, and deep diagnostics for containers and specific mount points.
An optional AI agent skill adds a natural‑language interface that can interpret intent, orchestrate multi‑step workflows, and apply model‑guided reasoning to propose remediation.
The utility is distributed via the NVIDIA Magnum IO GitHub repository and can be installed by cloning https://github.com/NVIDIA/MagnumIO.git.
What it means for you
Admins can use gds-diag to quickly identify configuration issues and improve storage bandwidth for GPU workloads. The natural‑language agent makes diagnostics accessible without deep command‑line expertise, supporting both on‑prem and cloud deployments.
Sources
- How to Check and Configure NVIDIA GPUDirect Storage for Your Platform, NVIDIA Developer (official), 30 Oct 2026
Written automatically from the sources above, then checked by our code: every number and name must appear in them, and no sentence may copy a source. Spotted a mistake? Report it.
Which AI can you use for free today?
Every free AI model, ranked by blind human votes, with the memory and speed it needs on your own GPU.