Spaces:
No application file
No application file
File size: 1,708 Bytes
0e83daa 875cad3 0e83daa fea9424 875cad3 fea9424 4f01886 875cad3 65d7347 fea9424 a0fdb46 10a11e5 a0fdb46 10a11e5 4f01886 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 |
---
title: README
emoji: 😻
colorFrom: gray
colorTo: purple
sdk: gradio
pinned: false
---
Tenstorrent is a next-generation computing company building AI accelerator and RISC-V CPU designs. Tenstorrent is bringing to market highly programmable chips at an accessible price point as a means of driving innovation.
Tenstorrent has developed a proprietary processor core called the Tensix which it uses to make its own AI chips, but also sells its IP and other technology to customers looking to make their own AI chips.
Tenstorrent has two open source software stacks:
TT-Forge (https://github.com/tenstorrent/tt-forge), and
TT-Metalium (https://github.com/tenstorrent/tt-metal/)
To contact us, please reach to our Discord channel (https://discord.gg/tenstorrent)
-------
TT-Metalium’s bare-metal programming stack defines a low-level programming model to allow users to write custom kernels for deployment on Tenstorrent hardware.
The software stack comes with debug and performance profiling tools as well as Tenstorrent's Python-based TT-NN library, which can run AI models out of the box without needing to deal with the complexities of the hardware. This negates theneed to write custom kernels unless the application requires it.
TT-Metalium will be used to fine-tune performance for production deployments, where small efficiencies can be multiplied by the scale of the deployment, and to develop custom kernels for scientific computing, where operations are not well-supported by existing kernel libraries.
For more details on the software, please visit https://tenstorrent.com/software/tt-metalium/
and the documentation https://docs.tenstorrent.com/tt-metal/latest/tt-metalium/index.html
|