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

PalateKit

Sign in
Menu

Request a resource

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

0 / 280

A page, a video or an image online.

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.

CUDA program debugging v1.0.0

Use this resource to find memory errors, races and crashes in CUDA code

New

  • 563 downloads
  • 172 likes
  • 208 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install cuda-gpu-debugging

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 helps you build for debugging, run Compute Sanitizer to catch out-of-bounds accesses and races between threads, then step through GPU threads in cuda-gdb. It also helps interpret common CUDA runtime errors.

Mostly picked by

Updated November 2, 2026

Card for the “CUDA program debugging” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
MIT

Similar resources

View all

What you get

  • SKILL.md: debug build, Compute Sanitizer and common CUDA errors
  • SKILL.md: cuda-gdb for inspecting GPU threads

Good fit for

  • Your kernel returns an illegal memory access error
  • Your GPU results change from one run to the next

Not for

    Tags

    • cuda
    • debugging
    • compute sanitizer
    • gpu

    About

    You say

    “My kernel returns cudaErrorIllegalAddress. Help me locate the bad access with Compute Sanitizer.”

    Card for the “CUDA program debugging” skill

    Skill published by mohitmishra786 under the MIT 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 find memory errors, races and crashes in CUDA code
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • SKILL.md6.3 KB

    Origin and licence

    Listed unmodified

    Author
    mohitmishra786
    Source
    github.com/mohitmishra786/low-level-dev-skills/tree/bdc58472fa9f309ed1b3f7d985a0d8e9bd8f4608/skills/gpu/cuda-debugging
    Licence
    MIT

    Notice

    « cuda-debugging » (mohitmishra786, MIT) 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.