Graduate Fellow: Caroline Trippel (2017)

Caroline Trippel, from Mishawaka, Indiana 

Studying at Princeton University

Research: Given the fundamental importance of memory consistency models, the overarching goal of Caroline's research is to facilitate their verification, particularly in heterogeneous systems. Modern computer systems employ increasing amounts of heterogeneity and specialization to achieve performance scaling at manageable power and thermal levels. Reaping the benefits of these heterogeneous and parallel systems necessitates memory consistency models which define behavior as fundamental as what value should be returned when software loads from memory. In many systems GPUs must communicate with CPUs, with accelerators for machine learning and other specialized functions, and even with on-chip micro-controllers that facilitate data movement and other aspects of GPU functionality. Each processing element may have been designed with different memory ordering expectations, and composing them together is complex and error-prone. Research Link

Bio: Caroline Trippel is a Ph.D candidate in the Computer Science department at Princeton University. She is currently in the fourth year of her Ph.D program and is advised by Professor Margaret Martonosi. Her specialization is broadly Computer Architecture and more specifically memory consistency model verification and design in heterogeneous systems. She received her B.S. in Computer Engineering from Purdue University in May 2013, her M.A in Computer Science from Princeton University in September 2015, and was a 2016 NVIDIA Graduate Fellowship Finalist.