Exascale Computers with Petascale Data

A new RPI project will seek the balance point between hardware, data, and algorithm.

A new project at Rensselaer Polytechnic Institute (RPI) will study the complex relationship between hardware, algorithms, and data structures in high performance computing. RPI data scientist George Slota has received a NSF Faculty Early Career Development (CAREER) grant to study the problem of how to match a supercomputer that operates on exascale with a dataset intended for petascale operations.

According to Slota, the best way to understand this problem is to “...map the data to the hardware, with consideration of the algorithm itself.” Slota will develop a “graph layout,” which RPI calls “a high-quality and scalable means of partitioning, ordering, and storing data given the data type, the relevant algorithms, and the hardware platform that will be used to analyze it.

See the press release at the RPI website for more information.