BASS Research Group

Our lab works on Computer Architecture and Performance Engineering, focusing on building high-performance computer systems. In particular, we design fast algorithms and build efficient software/hardware systems for AI & machine learning, security & privacy, science and finance. We are looking for grad and undergrad students to work on our research projects. If you are interested in joining our lab, please fill in the recruiting form.

Our recent focus is scalable and sustainable computing. The rapid escalating complexity of emerging applications and the lagging speed of computer systems creates a huge gap. For instance, the size of large language models balloons by 240 times in two years, significantly outpacing the advancements in hardware, which only manage to improve by a factor of 3. This gap largely limits the problem scale that emerging technology, e.g. AI, can be applied to, known as the scalability wall.

We use performance engineering to provide scalable and sustainable computing power in the post-Moore’s Law era. It covers three aspects: new algorithms, software performance engineering and novel hardware architectures. More specifically, application domains like machine learning and databases continually create a need for new algorithms. We explore algorithm innovations in new problem domains, and under new machine models, e.g. accelerator offloading. Beyond algorithms, software performance engineering removes software bloat and tailors software to hardware. For instance, data tiling improves cache/memory efficiency. In addition to software, hardware architecture redesign, like processor simplification and domain specialization, is inevitable. One example is to build dedicated AI and machine learning accelerators.