University of Cambridge > Talks.cam > Artificial Intelligence Research Group Talks (Computer Laboratory) > Accelerating Generative AI on Custom Hardware Accelerators - Challenges and Opportunities

Accelerating Generative AI on Custom Hardware Accelerators - Challenges and Opportunities

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Mateja Jamnik.

This presentation will center around the implementation challenges associated with generative AI on emerging spatial computing hardware. A comprehensive examination of the Tensix core architecture, integrated within the flagship Wormhole chips provided by Tenstorrent Inc., will be undertaken. Subsequently, we will assess state-of-the-art Large Language Models (LLMs) and investigate diverse approaches to effectively map these models onto the underlying hardware fabric. The presentation will be concluded by emphasizing various challenges and delineating prospects for further research in the domains of model design and optimization.

Bio: Partha Maji currently serves as Technical Director of ML Hardware Acceleration at Tenstorrent, a US based AI Hardware startup. Preceding this position, he held the position of Head of ML Research at Arm UK. With more than 18 years of industry experience, Partha’s expertise encompasses a diverse spectrum of domains that intersect AI/ML algorithms, software/compilers, and hardware architectures. Partha earned his Doctorate in Computer Science from the University of Cambridge.

You can also join us on Zoom

This talk is part of the Artificial Intelligence Research Group Talks (Computer Laboratory) series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2024 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity