Adventures in Parallel Dimensions
Speaker: Ravi Rajwar – Intel Corporation
Date and time: Tuesday, 15 Oct at 08:35.
Abstract:
Historically, transparent hardware improvements meant software just ran faster. However, the slowing growth in single-thread performance has meant an increasingly parallel future { involving parallelism across data, threads, cores, and nodes.
In this talk, I’ll explore some of the dimensions of parallelism and their growing interplay with software, discuss the resulting system challenges, and share some experiences.
Short bio:
Simplicial complexes – Much more than a trick
for concurrent computation lower bounds
Speaker: Nati Linial – The Hebrew University (Recipient of the 2013 Dijkstra Award )
Date and time: Wednesday, 16 Oct at 08:30.
Abstract:
As (hopefully) everyone in this audience knows simplicial complexes are very useful in distributed computing. It is not hard to understand the reason for this success. Graphs are the perfect tool in the modeling of systems that are driven by pairwise interactions. In contrast, distributed systems are all about multiway interactions involving many processors at once. This suggests the search for the high-dimensional counterparts of many basic concepts and ideas from graph theory. What plays the role of trees? What is the analog of graph connectivity? Of expansion? Is there a good theory that resembles the theory of random graphs?
In this lecture I will describe work done in recent years to address these questions. I have many excellent partners and collaborators in this research e®ort: Lior Aronshtam, Tomasz Luczak, Roy Meshulam, Ilan Newman, Tahl Nowik, Yuval Peled, Yuri Rabinovich and Mishael Rosenthal
Short bio:
