skills.rest
tenstorrenttenstorrentOfficial·6 Agent Skills Included

tt-forge

Compile, optimize, and debug models on Tenstorrent hardware

Guides development of machine learning models and custom kernels on Tenstorrent hardware using TT-Lang, TTNN, and TT-Metal. Eliminates guesswork in kernel fusion, trace capture, multi-chip sharding, and performance profiling with step-by-step workflows. Includes remote device scripts, bug report filing, and C++ kernel export to speed up hardware validation.
341|51
↗
npx skills add tenstorrent/tt-forge --all -g -y
Available:

Gives your AI agent the compiler stack architecture, test and benchmark commands, and coding standards needed to work correctly across the tt-forge sub-projects.

All Skills in This Repository (6)

Pure Emerald Level Indicators
📦 In Repo
tenstorrenttenstorrent

tt-enable-tracing

Enable TTNN trace capture and replay to eliminate host dispatch overhead.

Official
Advanced
📦 In Repo
tenstorrenttenstorrent

tt-bug-report

Creates GitHub issues for bug reports with minimized reproducers across TT repos.

Official
Intermediate
📦 In Repo
tenstorrenttenstorrent

tt-lang

Write TT hardware kernels using a Python DSL with data movement primitives.

Official
Advanced
📦 In Repo
tenstorrenttenstorrent

tt-connect-remote-device

Establish remote connections to Tenstorrent hardware and run kernels, copy files, and read logs.

Official
Intermediate
📦 In Repo
tenstorrenttenstorrent

tt-lang-profile-optimize

Profile and optimize TT-Lang kernels using TT-Metal tooling.

Official
Intermediate
📦 In Repo
tenstorrenttenstorrent

ttnn

Convert PyTorch tensors and execute tensor operations on Tenstorrent MeshDevice topologies.

Official
Advanced

Frequently Asked Questions

FAQPage Schema
How to install tt-forge?▼

Run `npx skills add tenstorrent/tt-forge --all -g -y` in your terminal to install all skills in this suite globally.

How to write custom kernels for Tenstorrent hardware?▼

The tt-lang skill provides the full TT-Lang DSL reference, letting you write fused Python kernels with compute and data movement threads instead of low-level C++.

How to speed up model inference on Tenstorrent chips?▼

Use the tt-enable-tracing skill to capture and replay operation traces, removing host dispatch overhead for real-time and multi-chip inference.

Can I profile and optimize TT-Lang kernels?▼

Yes. The tt-lang-profile-optimize skill walks through baseline measurement, bottleneck analysis, and iterative optimization using built-in perf summary and auto-profiling tools.

Does tt-forge work without physical hardware access?▼

Yes. The tt-connect-remote-device skill runs kernels in a functional simulator first and connects to remote Tenstorrent machines for final hardware validation.

Related Repositories in Software Engineering

View All in Software Engineering→
openclawopenclaw
388.4k

openclaw

Run a personal AI assistant across your devices and chat apps

Official📦 120 Skills
obraobra
280.0k

superpowers

Gives coding agents a disciplined workflow from idea to merged code

Community📦 15 Skills
facebookfacebook
248.4k

react

AI agent skills for building, testing, and porting React core

Official📦 14 Skills