ScaLab is a research group focusing on designing and optimizing algorithms, software, and computer architecture to improve the scalability of future computing systems and accelerate scientific discovery. Our research revolve around large-scale parallel systems, including high-performance computing and data centers. Our main research areas include:
Heterogeneous Computing
parallel systems with novel accelerators, including GPU, PIM, TPU, DPU, RISC-V accelerators, smartNIC, etc.
Memory-centric Computing
CXL, Disaggregated memory, near-memory processing, NVM, HBM, Heterogeneous memory
Converged Computing
architecture and middleware for coordinated utilization of HPC, cloud, and edge computing