HiPeC Lab

The computer industry is at a significant crossroads. Constrained by heat and power usage, today, all computing devices are composed of processors with increasing numbers and a variety of cores (CPU, GPU, FPGA, and DSP). These processors offer little or no increase in clock speed per core. This new computing era has brought a twofold challenge in building software: how to expose the parallelism in the software without much of a hassle and how to exploit the performance of all the processing elements fully. Our research aims to solve these complex challenges by building parallel programming models and runtimes to increase productivity and performance.