[University home]

Research Computing

Disruptive Computing

About

Disruptive Computing is a term that describes a new technological computing innovation, product, or service that provides significantly more computational power to the scientist or engineer than is currently available.

There are a number of emerging technologies that can deliver a much greater performance than traditional processors: field-programmable gate arrays (FPGAs); floating point accelerators (FPAs); multi-core and multi-threaded processor architectures. With these new HPC Technologies, the performance expected from a single processing unit can be accelerated by a factor of 10-100 while using relatively low power. This would make the size of systems required drop significantly allowing applications to scale on fewer processors at a much reduced cost.

A rather wordy, but informative, definition of Disruptive Technology can be found on Wikipedia

Projects

Manchester Computing are involved in a number of different projects using Disruptive Technologies. Brief details of the projects, including key researchers and collaborating organisations, are detailed below. Please click on the project headings to access more detailed information.

Multi-threaded Array Processing (MTAP)

Clear Speed CSX600The aim of this project is to develop a new partial differential equation (PDE) solver for ClearSpeed's CSX600 processor architecture. ClearSpeed's programmable processors have been designed to accelerate the performance of the world's most compute-intensive applications, enabling 'supercomputing' capabilities on affordable platforms including clusters and workstations. The mathematical operations of many HPC tasks are inherently array-oriented, so the CSX600 exploits this fine-grained parallelism in its design. The coprocessor uses less than 10 watts when sustaining about 25 GFLOPS (50% of peak). This is the direct result of using 96 PEs at a low clock speed of 0.25 GHz. The Advance accelerator board, with its two CSX600 chips, uses less than 25 watts and provides 100 GFLOPS peak (or 0.1 Teraflops).

The solver will be used in a finite element analysis program, taken from !ParaFEM, to undertake Advanced Materials Modelling.

Project Members

Project Partners

Active partners in the MDAP project are:

Field Programmable Gate Arrays (FPGAS)

A field programmable gate array (FPGA) is a semiconductor device containing programmable logic components and programmable interconnects. The programmable logic components can be programmed to duplicate the functionality of basic logic gates such as AND, OR, XOR, NOT or more complex combinatorial functions such as decoders or simple math functions. In most FPGAs, these programmable logic components (or logic blocks, in FPGA parlance) also include memory elements, which may be simple flip-flops or more complete blocks of memories. A hierarchy of programmable interconnects allows the logic blocks of an FPGA to be interconnected as needed by the system designer, somewhat like a one-chip programmable breadboard. These logic blocks and interconnects can be programmed after the manufacturing process by the customer/designer (hence the term "field programmable") so that the FPGA can perform whatever logical function is needed. [source: Wikipedia]

Project Members

The IBM Cell Processor

Investigation of Data Mining Algorithms on IBM Cell Broadband Engine (BE). Cell BE, Initially designed for the PlayStation 3, Sony, Toshiba and IBM's new Cell processor promises parallel computing capabilities within a single processor. It is a single-chip multiprocessor with nine processors operating on a shared, coherent memory. Manchester University has one of the first IBM Cell based systems in the world. The aim of this project is to understand this new architecture and to implement and investigate the performance of the data mining algorithms on this new architecture.

Project Members