Interested in a PhD in parallel and heterogeneous systems? Check out the current PhD opportunities.
Exploiting Parallelism through Type Transformations for Hybrid Manycore Systems (TyTra)
This collaborative 5-year EPSRC project aims to build compilers for heterogeneous platforms, including FPGAs, using multi-party session types for compiler-based program transformations. See the TyTra project web site for more info.
FPGA computing and high-level FPGA programming
My book High-Performance Computing Using FPGAs provides a comprehensive overview of Applications, Architectures and Tools for High-Performance Reconfigurable Computing
There is a growing interest in FPGA technology for accelerating computation in various ways. I am in particular interested in high-level programming of FPGAs, i.e. using languages at a higher abstraction level than Verilog or VHDL.
In particular, I work on "Greener Search", or FPGA-accelerated Information Retrieval. In 2009, we have developed a document filtering application in Mitrion-C which filters a collection of over a million documents in seconds rather than minutes. In the summer of 2011, in collaboration with HP (Houston), we created an improved implementation on the GiDEL PROCStar-III board, with a speed increase of 20x over our earlier work. We also work on GPU-based and manycore based accelerated information retrieval.
Novo-G FPGA Supercomputer
I am a member of the Novo-G Forum, "The Novo-G Forum is an international group of academic researchers and technology providers working collaboratively on experimental applications and tools to establish and showcase the advantages of RC at large scale, via the new Novo-G machine at CHREC, the world's most powerful reconfigurable supercomputer for research". Our aim is to demonstrate high performance Information Retrieval applications on the Novo-G machine. Read more ...
GPU and manycore programming in OpenCL
Another type of acceleration technology, General Purpose Graphics Processing Units (GPGPUs) is also becoming increasingly popular. OpenCL is an open, royalty-free standard for cross-platform, parallel programming of GPGPUs and modern multicore processors and manycore platforms such as the Intel MIC (Xeon Phi). We use OpenCL to accelerate applications in the fields of information retrieval (with Dr Leif Azzopardi), weather and climate research (with Prof. Saji Hameed at the University of Aizu, Japan and Prof. Tetsuya Takemi at Kyoto University, Japan) , particle physics (with Prof David Ireland). Read more ...
For more information, see the Research page.
I am associated with the Embedded, Networked, and Distributed Systems research group. This group systematically explores architectures, models, algorithms, measurement, and control of networked/distributed systems. The group is especially interested in large–scale systems, based on both wireless and wireline interconnection technologies, as well as high–performance system interconnects.
Teaching and Supervision
And in case you like a non–technical look at the world ...