Differences

This shows you the differences between two versions of the page.

Link to this comparison view

seminars:seminar_4_7_15 [2017/09/20 22:02] (current)
Line 1: Line 1:
 +====== Accelerating Data-Intensive Applications with Latency-Tolerant Distributed Shared Memory ======
 +==== Jacob Nelson (University of Washington) ====
 +== Tuesday, April 7, 4:30 PM to 5:30 PM ==
 +== CIC Panther Hollow ==
  
 +===== Abstract =====
 +
 +Conventional wisdom suggests that making large-scale distributed
 +computations fast requires minimizing the latency of individual
 +operations in the computation. In this talk I will discuss a system
 +called Grappa that takes the opposite view. Grappa tolerates latency
 +by exploiting application parallelism to achieve overall higher throughput.
 +
 +Grappa is a modern take on software distributed shared memory for
 +in-memory data-intensive applications. Grappa enables users to program
 +a cluster as if it were a single, large, non-uniform memory access
 +machine. Performance scales up even for applications that have poor
 +locality and input-dependent load distribution,​ as long as sufficient
 +parallelism is available.
 +
 +===== Bio =====
 +
 +Jacob Nelson is a Postdoctoral Research Associate in the Department of
 +Computer Science and Engineering at the University of
 +Washington. Jacob’s research explores new software and hardware
 +techniques to accelerate applications in big data and high-performance
 +computing. Jacob defended his Ph.D. at the University of Washington in
 +2014 working with Luis Ceze, Mark Oskin, and Simon Kahan.