|Session I Chair: Christos Kozyrakis (Stanford)|
|9:00am||Garth Gibson (CMU) (Video)(Slides) - |
Abstract Coming Soon.
|9:30am||Neel Smith (Holy Cross) (Video)(Slides) - |
As early as the third century BCE, scholars at the Library of Alexandria wrestled with problems of how to represent and analyze literary and scientific information in textual formats. The history of how their work has been transformed and transmitted to us offers valuable perspective to humanities scholars today as we develop the next step in that transmission: to digital forms. In this presentation, I first look at examples of how ancient Greek scholars working only with papyrus or parchment attempted to create interactive apps for cartography, chronological computation, and grammatical analysis. I then illustrate digital editions interpreting these apps as geographic information systems, directed graphs, and natural language parsers. The results can have implications beyond the narrow limits of humanities scholarship. Generalized approaches using finite state transducers to parse the morphology of ancient Greek uniformly fail, for example, because of the distinctive ways that morphology interacts with other features of the language. These examples not only remind us that the past is useful for thinking about the future, but also show how serious engagement with digital technology can give us insights into the past that were never previously possible.
|10:00am||Amin Vahdat (Google) (Video) (Slides) - |
Networking ties together storage, distributed computing and security in the Cloud. However, the demands of distributed processing, along with exponentially increasing data and storage, are forcing a re-think of networking technologies. Today’s realization of Moore’s Law is also shifting the industry away from the single-server computing model. We will discuss a path where the network itself becomes the engine for performance gains, and how this will support the next wave of advances in compute infrastructure.
|Session II Chair: David Ungar (IBM)|
|11:00am||Krste Asanovic (Video) (Slides) - |
The most important interface in a computer system is the instruction set architecture (ISA) as it connects software to hardware. So, given the prevalence of open standards for almost all other important interfaces, why is the ISA still proprietary? We argue that a free ISA is a necessary precursor to future hardware innovation, and there's no good technical reason not to have free, open ISAs just as we have free, open networking standards and free, open operating systems.
|11:30am||Mark Hill (Univ. of Wisconsin) (Video) (Slides) -
Abstract: While computing gets the glory, remember that is vast memory that makes most interesting computation possible. This talk will sample some of the computer architectural challenges that arise from million-fold memory capacity growth, the introduction of general-purpose graphics processing unit computing, and non-volatile memory's fusing of memory and storage.
|12:00pm||Kim Keeton (HP) (Video) (Slides) - |
By end of the decade we expect over 30 billion intelligent devices connected to the Internet, resulting in unprecedented amounts of data. At the same time, scaling of today's foundational memory technologies will significantly slow down. We will need to transform the ways in which we collect, process, store, and analyze that data. "The Machine" is a new architecture from Hewlett Packard Labs that brings together byte-addressable non-volatile memory, photonic interconnects, and specialized SoCs for computing at multiple scales, ranging from handhelds to rack-scale to data center-scale. As part of this initiative, we're building hardware, a new OS, new data stores, new analytics platforms, and new programming models, with the plan to open source various parts of the software stack. This talk will discuss the technologies that comprise The Machine and their implications for systems software and application programs, as well as describe the work we're doing at HPE to address some of these challenges.