ECE faculty, Prof. Ulya Karpuzcu, Prof. Sachin Sapatnekar, and Prof. Jian-Ping Wang have been awarded an $800,000 grant by the NSF under the Scalable Parallelism in the Extreme (SPX) program. Prof. Karpuzcu will be the lead PI for the project.
Big data is regarded as a killer application that can resolve most major computing challenges of the future. But exploding data volumes and the rising costs of data transportation have made the traditional model of computing, where data is delivered to the computing engine, increasingly untenable. Prof. Karpuzcu’s project undertakes to change this paradigm by exploring the possibility of bringing computation to data by developing a novel scalable framework for processing-in-memory (PIM). Although traditional CMOS structures are unsuitable for this, emerging spintronic technologies show promise for such a framework. The proposed approach will develop the notion of computational RAM (CRAM) to build PIM solutions to solve data-intensive computing problems using spintronics technologies.
The idea of bringing computation to memory has been gaining circulation, and the most viable solutions perform near-memory processing by performing computation at the edge of a large memory array. But this approach comes with significant overhead costs. The proposed CRAM-based approach avoids the substantial overheads of such a method. Instead it proposes a method for reconfiguring the memory to write the output of a logic operation directly onto a memory cell.
The project seeks to advance the state of the art in electronics technology, in large scale memory-centric computing using post-CMOS spintronic technologies, paving the path for new ways to build energy-efficient, scalable integrated systems.