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.

Enable TMA in an NVSHMEM application v1.0.0

Use this resource to prepare or review NVSHMEM CUDA code to use TMA without changing unrelated kernels

New

  • 340 downloads
  • 248 likes
  • 217 saves

Free

Install with the CLI

In a terminal, from the folder you want:

npx --yes palatekit@latest install nvshmem-tma-smem-transfers

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 assesses whether TMA is useful and available on your target, classifies each operation, applies the smallest safe change and explains the launch-time requirements. A static assessment needs no GPU and no installed NVSHMEM runtime.

Mostly picked by

Updated November 2, 2026

Card for the “Enable TMA in an NVSHMEM application” skill

Details

Works in
Install guides
Generates with
Your agent's model
Licence
Apache-2.0

Similar resources

View all

What you get

  • SKILL.md : five-step procedure and scope
  • references/registration-and-launch.md : registration and launch
  • references/validation-and-troubleshooting.md : validation and troubleshooting

Good fit for

  • You want to speed up NVSHMEM transfers on a recent GPU
  • You review an NVSHMEM kernel before enabling shared-memory transfers

Not for

    Tags

    • nvshmem
    • tma
    • shared memory
    • cuda
    • transfers

    About

    You say

    “Check whether my NVSHMEM kernel can use TMA for its transfers.”

    Card for the “Enable TMA in an NVSHMEM application” skill

    Skill published by nvidia under the Apache-2.0 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 prepare or review NVSHMEM CUDA code to use TMA without changing unrelated kernels
    3. Review and adjust the result

    Content preview

    Files

    • BENCHMARK.md7.6 KB
    • evals/evals.json6.1 KB
    • evals/files/cuda-transpose.cu0.6 KB
    • evals/files/direct-smem-put.cu0.9 KB
    • evals/files/gmem-staging.cu0.9 KB
    • evals/files/multi-node-put.cu0.5 KB
    • LICENSE11.3 KB
    • NOTICE-PALATEKIT.md0.5 KB
    • references/direct-smem-put.md2 KB
    • references/gmem-staging.md1.2 KB
    • references/registration-and-launch.md6.9 KB
    • references/validation-and-troubleshooting.md2.5 KB
    • skill-card.md4.3 KB
    • SKILL.md9.2 KB
    • skill.oms.sig6.4 KB

    Origin and licence

    Listed unmodified

    Author
    nvidia
    Source
    github.com/nvidia/skills/tree/ca116d60beafe1b4459f0ab790b7b450912c56b0/skills/nvshmem-enable-tma
    Licence
    Apache-2.0

    Notice

    « nvshmem-enable-tma » (nvidia, Apache-2.0) 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.