Computer and Mathematical Sciences

Reconfigurable System A19

  • Prof. Hiroki Nakahara      
KeywordsReconfigurable system, FPGA, AI, Machine learning, VLSI, Computer Architecture

Research and development of computing systems based on next-generation reconfigurable LSI

Our course focuses on the research and development of computing systems utilizing novel reconfigurable LSI technologies. Our research primarily centers on reconfigurable systems such as Field Programmable Gate Arrays (FPGAs), particularly exploring methods of implementing and applying AI on FPGAs, as well as high-performance computing techniques.
With the anticipated end of Moore’s Law and the stagnation of conventional FPGA development, our lab is investigating next-generation architectures and related advanced LSI designs. The project also encompasses the development of open EDA tools like OpenLANE.
A major focus of our work is on accelerating deep learning—one of the key techniques in AI—using FPGA-based custom accelerators. We specialize in AI model optimization (model compression and quantization) and the design of dedicated hardware, and have achieved leading results in this area.
The lab actively collaborates with multiple industry partners, offering opportunities to work with cutting-edge FPGA technologies. In addition to publishing research findings in academic venues, we pursue real-world impact through commercialization efforts in partnership with startups such as Tokyo Artisan Intelligence Inc., a university spin-off from Tohoku University.
  • Reconfigurable LSI

  • Reconfigurable Computing Architecture