Paper on new benchmarking platform SpikeSim for spiking neural networks wins IEEE TCAD award
The 2025 IEEE Transactions on Computer-Aided Design (TCAD) Donald O. Pederson Best Paper Award has been conferred on a cross-institutional team comprising Louis John Schnell Professor Yu Cao, Professor Priyadarshini Panda of Yale University and their research teams. The team’s paper is titled “SpikeSim: An End-to-End Compute-in-Memory Hardware Evaluation Tool for Benchmarking Spiking Neural Networks.” Panda was the principal investigator on the project and the research presented in the paper was completed during Cao’s tenure at Arizona State University.
Spiking neural networks (SNNs) are neuromorphic models inspired by the human brain. Unlike conventional machine‐learning approaches, SNNs represent and process information using discrete spikes offering significant gains in energy efficiency and intelligent behavior. In their award-winning paper, the team introduces SpikeSim, a comprehensive compute‐in‐memory benchmarking platform for SNNs that enables researchers to explore the architectural design space and optimize neuromorphic systems.
To develop a simulator for SNN benchmarking, the team addressed several key challenges such as ensuring hardware fidelity by accurately modeling all components of the SNN system, efficiently managing data movement and memory resources, and effectively mapping SNN workloads onto the simulation framework. The SpikeSim platform provides critical insights into spiking system design such as the mapping of unary activations to compute-in-memory macros, the energy and area overhead associated with neuron implementations, and the communication costs within SNN architectures.
Cao joined the University of Minnesota Twin Cities in 2023. He leads the Microelectronics Co-design Research Group where they work on applications of new understandings in computer engineering and semiconductor technology, focusing on hardware-algorithm co-design for energy-efficient computing. The goal is to advance the frontiers of computing hardware through physical analysis and integrated circuit design, and pave the path toward future intelligent systems. Currently, the group’s key areas of focus are: heterogeneous integration, AI acceleration for intelligent systems, neural-inspired computing, and reconfigurable systems.
The research presented in the award-winning paper was completed as part of Cao’s collaboration with Panda’s group at the SRC/DARPA JUMP Center for Brain-Inspired Computing (C-BRIC). They are continuing their collaboration in JUMP 2.0, under the Center for the Co-Design of Cognitive Systems (CoCoSys). They are working on next-generation intelligent hardware platforms through a hardware-algorithm co-design approach.
JUMP is the Joint University Microelectronics Program, Semiconductor Research Corporation (SRC)-led “public-private partnership in cooperation with DARPA, co-sponsored by SRC, DARPA, the commercial semiconductor industry, and the defense industrial base.” The mission is to look beyond today’s technology horizon and create new general purpose architectures and system designs that relax device constraints and provide opportunities for new device types and novel, heterogeneous integration solutions (from the JUMP website). Learn more about JUMP
TCAD is the IEEE’s flagship journal in the field of electronic design automation (EDA). The Donald O. Pederson award recognizes the single best paper published in the journal each year. The paper authors were officially recognized at the Design Automation Conference in San Francisco in June. Details about other award recipients at the IEEE Council on Electronic Design Automation