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 profiling with Nsight v1.0.0

Use this resource to find the bottleneck of a CUDA kernel with Nsight

New

  • 464 downloads
  • 228 likes
  • 123 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install nsight-cuda-profiling

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 choose between Nsight Systems and Nsight Compute, collect a profile from the command line, annotate code with NVTX ranges and read the metrics to tell whether a kernel is memory-bound or compute-bound.

Mostly picked by

Updated November 2, 2026

Card for the “CUDA profiling with Nsight” skill

Details

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

Similar resources

View all

What you get

  • SKILL.md: choosing the tool, Nsight Systems and NVTX annotations
  • SKILL.md: Nsight Compute, ncu from the command line and roofline reading

Good fit for

  • A CUDA kernel is slower than expected and you do not know why
  • You compare two tiling or block size variants

Not for

    Tags

    • nsight
    • profiling
    • cuda
    • roofline

    About

    You say

    “Profile this matrix multiplication kernel with ncu and tell me whether it is memory-bound.”

    Card for the “CUDA profiling with Nsight” 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 the bottleneck of a CUDA kernel with Nsight
    3. Review and adjust the result

    Content preview

    Files

    • LICENSE1.1 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • SKILL.md6.6 KB

    Origin and licence

    Listed unmodified

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

    Notice

    « cuda-profiling » (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.