Every premium and members-only resource for $9.99/month, as a no-commitment subscription. See the plans

PalateKit

Menu

Request a resource

Describe the resource you are missing in one sentence. We will email you when it is published.

0 / 280

Unlock PalateKit

Unlock everything

Plan

Paying means accepting our terms of sale.

Secure payment by Stripe. Receipts and invoices sent by email.

Welcome back

Sign in to install, download or save this resource. A free account is enough for free resources.

Diagnosing CUDA bugs with Compute Sanitizer v1.0.0

Use this resource to diagnose CUDA memory errors, data races or barrier bugs with Compute Sanitizer

New

  • 405 downloads
  • 234 likes
  • 138 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install cuda-bug-diagnosis-compute-sanitizer

With no option, the skill goes to ~/.agents/skills/; add --agent claude-code (or codex, cursor…) to target an agent.

First time: connect the CLI with

npx --yes palatekit@latest login

The agent picks the tool from the symptom (out-of-bounds access, shared-memory race, uninitialized read, divergent barrier), reads the matching reference before running it and explains the report. It does not cover host-side bugs or global-memory data races.

Mostly picked by

Updated November 2, 2026

Card for the “Diagnosing CUDA bugs with Compute Sanitizer” skill

Details

Generates with
Your agent's model
Licence
CC-BY-4.0

Similar resources

View all

What you get

  • SKILL.md: symptom routing and procedure
  • references/memcheck.md: memory errors
  • references/racecheck.md: data races

Good fit for

  • A CUDA kernel gives wrong results or crashes on an illegal access
  • A program hangs on a __syncthreads barrier

Not for

    Tags

    • CUDA
    • Compute Sanitizer
    • debugging
    • GPU

    About

    You say

    “My CUDA kernel has an intermittent illegal access: which Compute Sanitizer tool should I run?”

    Card for the “Diagnosing CUDA bugs with Compute Sanitizer” skill

    Skill published by nvidia under the CC-BY-4.0 license, delivered in its original version.

    Show more

    How to use it

    1. Install the skill in your agent's skills folder
    2. Ask your agent to diagnose CUDA memory errors, data races or barrier bugs with Compute Sanitizer
    3. Review and adjust the result

    Content preview

    Files

    • BENCHMARK.md9.7 KB
    • evals/config.yml1.2 KB
    • evals/EVAL.md2.6 KB
    • evals/evals.json16.8 KB
    • LICENSE11.3 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/changelog.md4.9 KB
    • references/initcheck.md6 KB
    • references/memcheck.md9.4 KB
    • references/racecheck.md8 KB
    • references/synccheck.md6.4 KB
    • skill-card.md4.9 KB
    • SKILL.md14.1 KB
    • skill.oms.sig6.2 KB

    Origin and licence

    Listed unmodified

    Author
    nvidia
    Source
    github.com/nvidia/skills/tree/ca116d60beafe1b4459f0ab790b7b450912c56b0/skills/compute-sanitizer-diagnose-cuda-bugs
    Licence
    CC-BY-4.0

    Notice

    « compute-sanitizer-diagnose-cuda-bugs » (nvidia, CC-BY-4.0) livré dans sa version d'origine : fiche en français et en anglais par PalateKit.

    User reviews

    –

    No ratings yet

    Loading reviews…

    Delete your review?

    Your rating and comment will be removed from the page.