You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!

Survey paper on Micron's Automata Processor

sparsh

New member
Problems from a wide variety of application domains can be modeled as ``nondeterministic finite automaton'' (NFA) and hence, efficient execution of NFAs can improve the performance of several key applications. Since traditional architectures, such as CPU and GPU are not inherently suited for executing NFAs, special-purpose architectures are required for accelerating NFAs.
Micron's automata processor (AP) exploits massively parallel in-memory processing capability of DRAM for executing NFAs and hence, it can provide orders of magnitude performance improvement compared to traditional architectures.

Our paper surveys 50+ techniques that propose architectural optimizations to AP and use it for accelerating problems from various application domains such as bioinformatics, data-mining, machine learning, natural language processing, high-energy physics, and graph analytics. PDF is here

It is noteworthy that Micron has recently stopped developing AP. Instead, other companies such as Natural Intelligence Semiconductor (http://naturalsemi.com) and some academic research centers (such Center for Automata Processing at the University of Virginia) are leading the development and market-adoption of AP.
 
Last edited:

Arthur Hanson

Active member

sparsh

New member
Sorry, Arthur. I have not idea about the business aspect. Technically, as an academician, I can say that the effectiveness of Automata execution depends a lot on memory technology. If 3D Xpoint can provide larger fan-in/fan-out, then it will be helpful for modeling complex automata which have a high degree of interconnection. For example, Levenshtein automaton with large edit distance cannot be efficiently mapped on Micron's AP due to limited fan-in/out. This may change with 3D Xpoint.
 

Arthur Hanson

Active member
Thanks, I appreciate the input greatly. Do you think Crossbar memory will come to be a serious player in the market. I know in 2013 TSM was going to work on Crossbar and have read they are going to use it for onboard memory. I have also read the Chinese are working on Crossbar. Any thoughts, information or views on this would be appreciated. I feel AMAT is intimately involved in developing the stacking/layering process for both 3dXpoint and Crossbar. Any information or opinions on this would be appreciated.
 
Top