Hardware Support for Massively Parallel Processing in Memory and Storage seeks to achieve “bare metal” performance and efficiency, i.e., as close as possible to the inherent bit-level parallelism of the data arrays. It explores novel hardware technologies for embedding processing in or near the data arrays to create intelligent memory and storage (IMS), and what type of mechanisms and abstractions this hardware should present to the software.
Current5 Research Tasks9 Universities62 Students15 Faculty Researchers31 Liaison Personnel
This Year16 Research Publications
Last Year5 Task Starts103 Research Publications1 Patent Applications