Customized computing and machine learning

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

• Atefeh Sohrabizadeh (UCLA)
• Monday 25 March 2024, 13:00-14:00
• Department of Electrical and Electronic Engineering, Room 909B.

If you have a question about this talk, please contact George A Constantinides.

Nowadays, data centers are filled with enormous amounts of data that require high-performance computing. While we used to be able to answer this need by scaling the frequency, the breakdown of Dennard’s scaling has rendered this approach obsolete. On the other hand, DSAs have gained a growing interest since they can offer high performance while being energy efficient. Unfortunately, despite the huge speedups that DSAs can deliver compared to general-purpose processors, their programmability has not caught up. In the past few decades, HLS tools were introduced to raise the abstraction level and free designers from delving into architecture details at the circuit level. While HLS can significantly reduce the efforts involved in the hardware architecture design, not every HLS code yields optimal performance, requiring designers to articulate the most suitable microarchitecture for the target application. This can affect the design turn-around times as there are more choices to explore at a higher level. Moreover, this limitation has confined the DSA community primarily to hardware designers, impeding widespread adoption. Our approach addresses this issue by synergizing customized computing and machine learning. Specifically, our effort consists of two core parts: 1) Customized computing for machine learning, exemplified by FlexCNN (CNN accelerator) and StreamGCN (GCN accelerator). 2) Machine learning facilitates the optimization of customized computing through AutoDSE (bottleneck-based optimizer), GNN -DSE, and HARP (HLS tool surrogates). This talk provides an overview of the frameworks developed to tackle this challenge.

This talk is part of the CAS Talks series.

This talk is included in these lists:

• Andrea Picciau's list
• CAS Talks
• Circuits and Systems Group: Internal Seminars
• Department of Electrical and Electronic Engineering, Room 909B

Note that ex-directory lists are not shown.