Clarkson Ph.D. Student Wins Best Paper Award for Research

Lin Jiang, a Ph.D. student in Electrical and Computer Engineering at Clarkson University, recently received the Prof. Avram Bar-Cohen Best Paper Award in the Emerging Technologies and Fundamentals Track for his research work on “Chip-level Thermal Simulation for a Multicore Processor Using a Multi-Block Model Enabled by Proper Orthogonal Decomposition”. […]

Prof. Avram Bar-Cohen Best Paper Award in the Emerging Technologies and Fundamentals Track

Lin Jiang, a Ph.D. student in Electrical and Computer Engineering at Clarkson University, recently received the Prof. Avram Bar-Cohen Best Paper Award in the Emerging Technologies and Fundamentals Track for his research work on “Chip-level Thermal Simulation for a Multicore Processor Using a Multi-Block Model Enabled by Proper Orthogonal Decomposition”. He received the award at The Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (Itherm 2022) in San Diego, California. Other co-authors of this paper include Anthony Dowling, a PhD. student of Electrical and Computer Engineering, Yu Liu and Ming-Cheng Cheng. 

Lin is jointly advised by Dr. Yu Liu, Assistant Professor of Software & Computer Engineering and Dr. Ming-Cheng Cheng, Professor in Electrical Engineering at Clarkson.  IEEE Itherm is a prestigious international intersociety conference for academic and industrial communities to explore thermal, thermomechanical and emerging technology issues associated with electronic devices, packages and systems.

Lin’s research involves the development of an innovative computational methodology for full-chip thermal simulations of multi-core CPUs and general purpose GPUs enabled by a physics-based data-driven learning algorithm. The approach partitions an entire semiconductor chip into standard building blocks and the simulation model for each block is then developed and stored in a database. The work significantly minimizes the training resources needed in the typical learning algorithm and facilitates efficient parallel computing.  It has been shown that this multi-block learning approach offers a very accurate prediction of the dynamic thermal profile of the entire chip with a computational speedup over a few thousand times, compared to direct numerical methods.  Accurate full-chip thermal simulation for capturing hot spots in high-performance CPUs and GPUs has been a very challenging task due to the intensive computational effort needed in accurate direct numerical simulation.  Dr. Cheng says, “The recognition by the Itherm 2022 committee under the track of “Emerging Technologies and Fundamentalsstrongly indicates that Lin’s research offers a significant contribution to the emerging thermal management issue that is becoming critical in high-performance computing environments.”  

Next Post

Graphene synapses advance brain-like computer

Wed Aug 10 , 2022
image: A transistor that operates like synapses in the brain. view more  Credit: The University of Texas at Austin Computers that think more like human brains are inching closer to mainstream adoption. But many unanswered questions remain. Among the most pressing, what types of materials can serve as the best building […]