High-performance software acceleration on heterogeneous CPU/GPU/FPGA platforms
Our research activities are centered on enabling an implementation flow for software applications on GPUs and FPGAs for both data centers and edge computing, in order to both improve performance and reduce energy per operation, thus virtually extending the effects of Moore’s law for a few more years. We focus on automated resource management, memory interface optimization, and parallelization techniques. Application areas range from Artificial Intelligence to Computer Vision and financial algorithms.
Research topics
Leaders
Members