Architecting More Cooling-Efficient Datacenters By Removing the Peaks With Phase-Change Materials
Datacenters are rapidly increasing in size and computational ability. However, this growth places great stress on the power delivery and heat removal of the datacenter. The cost of power and cooling, and the computational capacity of the datacenter, are both driven by the peak demands on the power infrastructure and the cooling infrastructure. We use phase-change materials (e.g., wax) to remove the peak cooling demand with no impact on the workload or performance. We'll discuss Thermal Time Shifiting, which stores heat in place during peak load to be released only when cooling capacity is available, Virtual Melting Temperature that allows the datacenter to act as if we brought in new wax with different properties every night, and Thermal Gradient Transportation, which allows us to aggregate wax across the datacenter for increased efficiency.
About the Speaker:
Dean Tullsen is a professor and chair of the computer science and engineering department at University of California, San Diego. He received his PhD from the University of Washington in 1996, where he introduced simultaneous multithreading (hyper-threading). He has continued to work in the area of computer architecture and back-end compilation, where with various co-authors he has introduced many new ideas to the research community, including threaded multipath execution, symbiotic job scheduling for multithreaded processors, dynamic critical path prediction, speculative precomputation, heterogeneous multi-core architectures, conjoined core architectures, event-driven simultaneous code optimization, and data triggered threads. He is a Fellow of the ACM and the IEEE. He has twice won the Influential ISCA Paper Award.