Calendar

The polyhedral model has proven to be very useful to optimize and parallelize a particular class of compute intensive application kernels. A polyhedral optimizer needs to have affine functions defining loop bounds, memory accesses and branching conditions. Unfortunately, this information is not always available at compile time. To broaden the scope of polyhedral optimization opportunities, runtime information can be considered. This talk will highlight the challenges of integrating polyhedral optimization in runtime systems:

- When and how to detect opportunities for polyhedral optimization?
- How to model the observed runtime behavior in a polyhedral fashion?
- How to deal at runtime with the complexity of polyhedral algorithm?

These challenges will be illustrated in the context of both the APOLLO framework targeting C and C++ applications and of the JavaScript engine from Apple.

The recent evolution of mobile communications and the widespread use of ``smart'' mobile devices have radically changed the behavior and the needs of the mobile user. In the developed world, people that use their mobile devices just to place some calls and to text some SMS can be characterized as endangered species. Ubiquitous access to Internet, videos/songs streaming and upload/download data flows on the fly are some of the modern demands of the mobile user. It would not be an exaggeration to say that modern users have almost the same demands regardless if they are connected through wire to the network from their personal computers or if they have established a wireless connection through their cellular devices.

On the one hand the overall network load increases on the other hand base stations have limited resources due to the fact that are able to operate only to a limited part of the electromagnetic spectrum. Some (temporal) solutions for the aforementioned problem are the expansion of frequency operational bands (mm wave) or the use of more antennas (massive MIMO) or the denser deployment of small cells. In this talk, we are interested in the promising case of denser small cells that will be tighter integrated with the macro cell. Unfortunately to deploy optimally a small cell network is not trivial. A small cell network is usually deployed in an ad-hoc style and not all at once, so a part of the network exists already and cannot be planned. Additionally, there are natural obstacles and physical limitations that do not allow to deploy base stations wherever we want. So, the small cell network's topology is quite different from the traditional (macro cell) well-structure one.

From our point of view, the future of mobile communications will approach the following: A mass of users - will cause a nondescript data traffic - that will be served from an irregular planed and heterogeneous network. This chaotic picture however makes the problem of network modeling and performance analysis extremely challenging.

To this end, the our primary focus is to build up an analytical framework in order to analyze the performance of a randomly placed network, which serves randomly placed users. To achieve this, we based our analysis on two main tools: (a) stochastic geometry, to understand the impact of topological randomness and coverage maps and (b) queueing theory, to model the competition between concurrent flows within the same BS. The second goal is to propose, based on this analysis, some general design guidelines and insights about specific communication scenarios that mainly involve LTE and WiFi networks.