TT-Metal / TT-NN GitHub repository

Fields

id

sources/niche-tt-metal-repo

URL

https://github.com/tenstorrent/tt-metal

Publisher

Tenstorrent

Kind

repo

Accessed

2026-10-03

notes

tt-metal README: TT-NN is the Python & C++ neural-network operator library and TT-Metalium the low-level kernel programming model for Tenstorrent silicon. Carries Tenstorrent's own LLM throughput tables per hardware config (Galaxy/Wormhole, n300, QuietBox, n150, p150 Blackhole) with TTFT and T/S/U at batch 32 and batch 1. The README's own caveat is the important part: metrics come from the tt-metal model demos and 'Results may vary when using other runtimes such as the vLLM inference server', with a separate vLLM-TT-plugin release column. API metadata 2026-10-03: created 2023-02-13, last push 2026-10-03, v0.79.0 (2026-09-18), Apache-2.0, 24855 open issues, 1696 stars.

Notes

tt-metal README: TT-NN is the Python & C++ neural-network operator library and TT-Metalium the low-level kernel programming model for Tenstorrent silicon. Carries Tenstorrent's own LLM throughput tables per hardware config (Galaxy/Wormhole, n300, QuietBox, n150, p150 Blackhole) with TTFT and T/S/U at batch 32 and batch 1. The README's own caveat is the important part: metrics come from the tt-metal model demos and 'Results may vary when using other runtimes such as the vLLM inference server', with a separate vLLM-TT-plugin release column. API metadata 2026-10-03: created 2023-02-13, last push 2026-10-03, v0.79.0 (2026-09-18), Apache-2.0, 24855 open issues, 1696 stars.

Cross-references

This record references no other records.