Description:
High-energy physics experiments have to process huge amounts of data every second in real time. A new form of pattern recognition technology is needed, capable of dealing with enormous data streams and filtering backgrounds from the physics signals in which scientists are interested.
Fermilab has developed a high-speed flexible pattern recognition system containing two vital innovations. VIPRAM (Vertically Integrated Pattern Recognition Associative Memory) uses 3-D vertical integrated chip technology. The PULSAR-II processor has terabits per second data processing capability and can host a bank of VIPRAM chips, in order to rapidly recognize user defined patterns in big data applications.
- 100-terabit per second data processing in real time
- Highly scalable and user configurable
- Detecting patterns of fraud in financial transactions
- Mapping behavioral trends in the stock market
- Instrumentation at photon science facilities
- Proton computed tomography (pCT)
- Any industry with large amount of streaming data
