# wilicc/gpu-burn

Multi-GPU CUDA stress test

Repository: https://github.com/wilicc/gpu-burn
Canonical: https://ross.abutalabs.com/products/gpu-burn
Language: C++
License: BSD-2-Clause
License Family: permissive
Last push: 2026-05-31T00:17:14+00:00

## Health v2 (maintenance only)
Score: 70/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 85, release rhythm 35, longevity 100
- inputs: {"age_days": 3203, "days_push": 95, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2322, forks 418 (observed 2026-08-28T04:06:37.161635+00:00)

## What it is
GPU Burn is a multi-GPU CUDA stress test tool that pushes NVIDIA GPUs to maximum load for stability and thermal testing. It supports configurable memory usage, double precision, tensor cores, and Docker-based deployment.

## Use cases
- stress test my gpus under full load
- check gpu stability after overclocking
- burn-in test for a new multi-gpu server
- verify gpu cooling and thermals under load
- test tensor core performance on nvidia gpus
- load test all cuda devices in a system

## When to choose
- you need to stress test NVIDIA CUDA GPUs, especially multiple GPUs at once
- you want a simple CLI burn-in tool with Docker support
- you're validating GPU stability, cooling, or power delivery

## When to avoid
- you need to benchmark AMD or Intel GPUs (CUDA-only)
- you want detailed performance metrics or profiling rather than pure stress load
- you need a GUI-based testing tool

## Facets
- artifact type: cli-tool
- maturity: active
- function: benchmarking, gpu-computing, cli
- domain: developer-tools, hardware, performance
- platform: cli, cpp
- tags: cuda, stress-testing, multi-gpu, nvidia, hardware-testing, linux, docker, gpu

## Member repositories
- wilicc/gpu-burn (main) score 70

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:37.161635+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:38:38.537421+00:00, confidence not recorded.
  - readme: https://github.com/wilicc/gpu-burn (fetched 2026-08-28T04:06:37.161635+00:00, sha 3c4522de88df)
- Data as of 2026-08-30T08:39:29.467469+00:00.