Dr S Johnson
Parallel Processing Research Group
University of Greenwich, UK.
Email: S.Johnson@gre.ac.uk
HPC has failed to make the impact in the large sections of the computer
market that was hoped and indeed expected a few years ago. Several key
reasons for this failure can be identified from both the hardware and
software perspective. A fundamental limiting factor is the volume of
efficient application software available for such machines. Porting a
code onto a distributed architecture is a significant task and in many
cases this is insurmountable due to financial and time constraints.The only real solution to this problem is the use of sophisticated parallelisation tools to port application codes quickly and with high efficiencies. A number of such tools exist but, as yet, their success is limited. Are the problems of current tools fundamental for any tool? How significant will the user's role have to be? Will the limitations of hardware prove to be a serious obstacle in the future? The answers to these questions will have a major impact on the future of HPC.