Parallel and Distributed Computing

Research Group


Software Packages

Parallel Computing Research Group of Computer Engineering Department at Bilkent University has a long-term expertise in both theoretical and practical aspects of parallel computing. In the last decade, the group, led by Prof. Aykanat, has conducted research and development on three different distributed-memory parallel systems, namely a 32-node iPSC/2 hypercube, Parsytec CC24 with 24 processors, 32-node and 48-node Beowulf PC clusters, which are acquired by funding from Intel, EU and TUBITAK, respectively.

Research interests of the group mainly include: parallel computing (especially for irregular and unstructured applications on distributed-memory architectures), parallel scientific computing and its combinatorial aspects, parallel computer graphics applications, parallel data mining, graph and hypergraph partitioning, load balancing, high performance information retrieval and GIS systems, distributed databases, and grid computing.

The theoretical and practical research of the center led to the following application software: Data partitioning and mapping (a state-of-the-art multilevel hypergraph partitioning tool that supports optimization of various load balancing and communication metrics), iterative solvers (software utilities for development of parallel iterative methods for linear-system solutions), task assignment and allocation (models and software for minimizing the sum of task execution and inter-task communication times in distributed systems), synthetic image generation (efficient parallel progressive radiosity and ray-tracing software packages), scientific visualization (efficient parallel programs for direct volume rendering of 3D unstructured data grids), information retrieval (models, algorithms, and software for efficient query processing in parallel text retrieval systems), data mining (parallel software for discovery of frequently-occurring patterns in very large datasets), geographic information systems (high-performance storage and access schemes for transportation networks), distributed database systems (data allocation and partitioning software for efficient parallel query processing in multi-disk systems).