# Tensor & Tiling Library

> software library to enable the efficient tiling and compute with tensors

**Wikidata**: [Q116872571](https://www.wikidata.org/wiki/Q116872571)  
**Source**: https://4ort.xyz/entity/tensor-tiling-library

## Summary
The Tensor & Tiling Library (TTL) is an open-source software library designed to optimize tensor tiling and computation, particularly for OpenCL-based applications. Released in 2023, it was developed by Mobileye and distributed under multiple licenses, including Apache 2.0, MIT, and CC-BY 4.0.

## Key Facts
- **Publication Date**: Released on February 21, 2023 (GitHub commit and Khronos Group announcement).
- **Developer**: Created by Mobileye.
- **Licenses**: Apache Software License 2.0, MIT License, and Creative Commons Attribution 4.0 International.
- **Aliases**: OpenCL-TTL, OpenCL Tensor Tiling Library.
- **Instance of**: Software library, free software.
- **Copyright Status**: Copyrighted.
- **Logo**: Available at [Wikimedia Commons](https://commons.wikimedia.org/wiki/Special:FilePath/Tensor_Tiling_Library_logo_2023.png).

## FAQs
### Q: What is the Tensor & Tiling Library used for?
A: The Tensor & Tiling Library is designed to enable efficient tiling and computation with tensors, particularly for OpenCL-based applications.

### Q: Who developed the Tensor & Tiling Library?
A: The library was developed by Mobileye and released under open-source licenses.

### Q: What licenses does the Tensor & Tiling Library use?
A: The library is distributed under the Apache Software License 2.0, MIT License, and Creative Commons Attribution 4.0 International.

### Q: When was the Tensor & Tiling Library released?
A: The library was released on February 21, 2023, as confirmed by GitHub and Khronos Group announcements.

### Q: What are the aliases for the Tensor & Tiling Library?
A: The library is also known as OpenCL-TTL and OpenCL Tensor Tiling Library.

## Why It Matters
The Tensor & Tiling Library addresses the need for efficient tensor operations in OpenCL environments, which are critical for high-performance computing, machine learning, and real-time data processing. By optimizing tensor tiling and computation, it enables faster and more efficient execution of tensor-based workloads, making it a valuable tool for developers working with OpenCL. Its open-source nature encourages collaboration and innovation, allowing the community to contribute to and improve the library. The library’s release marks a significant step forward in optimizing tensor operations for OpenCL, providing a robust solution for applications requiring high-performance tensor computations.

## Notable For
- **Open-Source Development**: Released under multiple open-source licenses, promoting community contribution and collaboration.
- **OpenCL Optimization**: Specifically designed to enhance tensor operations in OpenCL environments.
- **Mobileye Collaboration**: Developed in partnership with Mobileye, leveraging their expertise in real-time computing.
- **Multi-License Distribution**: Available under Apache 2.0, MIT, and CC-BY 4.0, catering to different use cases and compliance needs.
- **Khronos Group Announcement**: Officially released by the Khronos Group, ensuring broad industry recognition and adoption.

## Body
### Overview
The Tensor & Tiling Library (TTL) is an open-source software library developed by Mobileye to optimize tensor tiling and computation, particularly for OpenCL-based applications. It was released on February 21, 2023, and is distributed under multiple licenses, including Apache Software License 2.0, MIT License, and Creative Commons Attribution 4.0 International.

### Development and Release
- **Developer**: Mobileye, a subsidiary of Intel, contributed to the development of the library.
- **Publication Date**: The library was officially released on February 21, 2023, as confirmed by GitHub commit records and Khronos Group announcements.
- **Khronos Group**: The release was announced by the Khronos Group, a non-profit organization promoting open standards for graphics and parallel computing.

### Licensing
- **Apache Software License 2.0**: Allows users to freely run, study, change, and distribute the software and modified versions.
- **MIT License**: Permits similar usage with additional permissions for sublicensing and patent use.
- **Creative Commons Attribution 4.0 International**: Requires attribution but allows for broader use and modification.

### Technical Details
- **Aliases**: The library is also known as OpenCL-TTL and OpenCL Tensor Tiling Library.
- **Instance of**: Classified as a software library and free software.
- **Copyright Status**: The library is copyrighted, with specific terms outlined in its licenses.

### Significance
- **OpenCL Optimization**: TTL is designed to improve tensor operations in OpenCL, which is widely used in high-performance computing and machine learning.
- **Community Collaboration**: Its open-source nature encourages contributions from the developer community, fostering innovation and improvement.
- **Industry Recognition**: The Khronos Group’s announcement ensures broad industry adoption and recognition.

## References

1. [Source](https://github.com/KhronosGroup/OpenCL-TTL/blob/main/LICENSES/Apache-2.0.txt)
2. [Source](https://github.com/KhronosGroup/OpenCL-TTL/blob/main/LICENSES/MIT.txt)
3. [Source](https://github.com/KhronosGroup/OpenCL-TTL/blob/main/LICENSES/CC-BY-4.0.txt)
4. [Source](https://github.com/KhronosGroup/OpenCL-TTL/commit/9833f05ace05e4ed18c080b9909f0f14254a5a46)
5. [Source](https://www.khronos.org/news/permalink/khronos-releases-open-source-opencl-tensor-tiling-library)