News

Objectives

To meet ever-increasing computing needs and overcome power density limitations, the computing industry has entered the era of parallelization, with tens to hundreds of computing cores integrated into a single processor and hundreds to thousands of computing servers connected in warehouse-scale data centers. However, such highly parallel, general-purpose computing systems still face serious challenges in terms of performance, energy, heat dissipation, space, and cost. The Center for Domain-Specific Computing (CDSC) looks beyond parallelization and focuses on domain-specific customization as the next disruptive technology to bring orders-of-magnitude power-performance efficiency improvement to important application domains.

Scope and Approach

CDSC develops a general methodology for creating novel customizable computing platforms and the associated compilation tools and runtime management environment to support domain-specific computing. The recent focus is on design and implementation of accelerator-rich architectures, from single chips to data centers. It also includes highly automated compilation tools and runtime management software systems for customizable heterogeneous platforms, including multi-core CPUs, many-core GPUs, and FPGAs, as well as a general, reusable methodology for customizable computing applicable across different domains. By combining these critical capabilities, the goal is to deliver a supercomputer-in-a-box or supercomputer-in-a-cluster that can be customized to an application domain to enable disruptive innovations in that domain. This approach has been successfully demonstrated by the researchers in the domains of deep learning, medical image processing, and precision medicine.  Some of the research projects in CDSC includes

• Energy-efficient large-language models (LLMs)
• Use deep learning (graph neural networks and large-language models) to predict program and circuit performance, and automate hardware accelerator designs.
• Design of fast reconfigurable architectures for customized accelerators
• Accelerator of sparse-linear algebra functions for deep learning and high-performance computing
• Near-storage and near-memory computing
• Compilation and architecture optimization for quantum computing (e.g. with superconducting technology and neutral atom arrays)

Team and Support

The current research team consists of a group of highly accomplished researchers with diversified backgrounds, including computer science and engineering, electrical engineering, medicine, and applied mathematics from UCLA, Cornell, and Georgia Institute of Technology. CDSC offers many research opportunities for graduate students, and also offers summer research opportunities for undergraduate students.

CDSC was originally funded by the National Science Foundation with a $10 million award from the 2009 Expeditions in Computing program, which was among the largest single investments made by the NSF Computer and Information Science and Engineering (CISE) Directorate. In July 2014, CDSC was awarded an additional $3 million by Intel Corporation with matching support from NSF under its Innovation Transition (InTrans) program. This award supports follow-on research on accelerator-rich architectures with applications to health care, in which personalized cancer treatment was added as an application domain in addition to medical imaging. Currently, the research programs in CDSC are supported by NSF, the SRC JUMP program, and a number of industrial partners worldwide.