|CF home||organization||program||venue / travel|
Recent Technological Trends and their Impact on System Design
Monday, May 7
Over the past two decades, significant advancements in VLSI technologies (e.g. Moore's Law), network bandwidth, and disk storage capacity have fueled an unprecedented integration of information technologies into the global economy. These advances have enabled the I/T (information technology) community to develop and deploy containerized and composable software stacks, while providing adequate performance to the end users. This model of driving high programmer-productivity with responsiveness in the I/T enterprises is based partly on software containerization, This model also depends significantly on obtaining high performance and robustness from the hardware and the operating systems autonomically. Because programmer-productivity and system responsiveness are key factors for the global economic growth, significant research and development efforts have been invested in continuing the innovations in hardware and operating systems. This talk will discuss the recent technological breakthroughs in VLSI technologies (i.e. high k dielectric) and in disk recording technologies (e.g. perpendicular recording). I will also discuss the impacts of the design decisions in processor designs advancements targeted for the four canonical usage segments, namely, HPC, Commercial computation, Games, and Embedded Systems. I will also speculate on future technological advances that may potentially impact system design.
Models for Parallel and Hierarchical Computation
Tuesday, May 8
Technological evolution leads to machine organizations dictated by the fundamental physical constraints on layout and signal speed. These organizations are parallel, to exploit concurrency, and hierarchical, to exploit data locality.
We will develop a theoretical perspective on machine structures that are optimized under layout and message speed constraints, surveying a number of results on memory, processor, and network organizations.
We will also consider the tradeoffs arising in the design of portable and efficient algorithms for parallel and hierarchical machines. In this context, we will discuss an approach based on adaptive algorithms and we will review some results on cache obliviousness and on network obliviousness.
The Quantum Challenge to Computer Science
Wednesday, May 9
Information is physical: the laws which govern its encoding, processing and communication are bound by those of its unavoidably physical embodiment. In today's informatics, information obeys the laws of Newton's and Maxwell's classical physics: this assertion holds all the way from commercial computers down to (up to?) their most abstract models, e.g. Turing machines and lambda-calculus. Today's computation and communication are classical.
Research in quantum information processing and communication was born some twenty five years ago, as a child of two major scientific achievements of the 20th century, namely quantum physics and information sciences. The driving force of this interdisciplinary research is that of looking for the consequences of having information encoding, computation and communication based upon the laws of quantum physics, i.e. the ultimate knowledge that we have, today, of the foreign world of elementary particles, as described by quantum mechanics. Breakthroughs in cryptography, communications, information theory, algorithmics and, more recently, in abstract computational models, programming languages and semantics frameworks, have shown that this transplantation of information from classical to quantum has far reaching consequences, both quantitative and qualitative, and opens new avenues for research within the foundations of computer science. From a computer scientist's point of view, which is my point of view, I will explain the basics, survey the main achievements, and outline the current hot topics and major challenges of this promising and stimulating research. I will not assume any prior knowledge in quantum mechanics from the audience.