Von Neumann syndrome

From Wikipedia, the free encyclopedia

The von Neumann syndrome is a computing term coined by Prof. C. V. Ramamoorthy (after having listened to a keynote by Reiner Hartenstein, and named after John von Neumann's model of computer architecture). Ramamoorthy noted that for most applications in massively parallel computing systems with thousands or tens of thousands of processors the performance can be less than hoped. Sometimes called a "supercomputing crisis" it is believed to be due to two factors. Firstly a hardware barrier in the efficiency in moving data, called the memory wall or von Neumann bottleneck. Secondly a fall in programmer productivity when faced with systems that are massively parallel, the difficulties in developing for parallelism (or thread-level parallelism in multi-core CPUs) when previously this was not an issue.

The term also explains the Reconfigurable computing paradox, where the use of reconfigurable computing on FPGAs produces faster results despite a lower speed, where, instead of moving data around, the locality of execution is optimized by placement and routing. This is also called a data-stream-driven anti machine paradigm.

[edit] References

Satnam Singh: Reconfigurable Computing Systems as Platforms for Heterogeneous Many-Core Architectures; Proc. DATE-Conference, April 2007, Nice, France
Walid Najjar: Reconfigurable Supercomputing - reality or pipedreams; Proc. DATE-Conference, April 2007, Nice, France
Thomas Sterling, Peter Kogge, Ken Kennedy, Steve Scott, Don Becker, William Gropp (panelists): Multi-Core for HPC: Breakthrough or Breakdown?; Supercomputing Conference (SC06), November 2006, Tampa, Florida, USA
Tarek El-Ghazawi, Dave Bennett, Dan Poznanovic, Allan Cantle, Keith Underwood, Rob Pennington, Duncan Buell, Alan George, Volodymyr Kindratenko (panelists): Is High-Performance, Reconfigurable Computing the Next Supercomputing Paradigm?; Supercomputing Conference (SC06), November 2006, Tampa, Florida, USA