removed exaggeration |
m fix |
||
(One intermediate revision by the same user not shown) | |||
Line 1: | Line 1: | ||
{{Multiple issues|unreferenced=December 2009|orphan=November 2006|advert=August 2011|cleanup=March 2009|peacock=August 2011}} |
|||
{{Unreferenced|date=December 2009}} |
|||
{{Orphan|date=November 2006}} |
|||
{{Advert|date=August 2011}} |
|||
{{Cleanup|date=March 2009}} |
|||
{{Peacock|date=August 2011}} |
|||
''Reconfigurable Supercomputing'' is a new direction in [[Supercomputing]], where [[Reconfigurable Computing]] methods are adopted by supercomputer architectures, such as those based on [[FPGA]]s. From FPGA usage as accelerators, speed-up factors by up to two orders of magnitude are reported. Algorithmic cleverness is the secret of success, based on software to [[configware]] migration mechanisms, striving away from memory-cycle-hungry instruction-stream-based computing paradigms also using memory-cycle-hungry data movement methods. A new kind operating system modules as well as new compilation techniques are needed to ease usage. |
''Reconfigurable Supercomputing'' is a new direction in [[Supercomputing]], where [[Reconfigurable Computing]] methods are adopted by supercomputer architectures, such as those based on [[FPGA]]s. From FPGA usage as accelerators, speed-up factors by up to two orders of magnitude are reported. Algorithmic cleverness is the secret of success, based on software to [[configware]] migration mechanisms, striving away from memory-cycle-hungry instruction-stream-based computing paradigms also using memory-cycle-hungry data movement methods. A new kind operating system modules as well as new compilation techniques are needed to ease usage. |
||
Revision as of 16:17, 11 August 2011
Reconfigurable Supercomputing is a new direction in Supercomputing, where Reconfigurable Computing methods are adopted by supercomputer architectures, such as those based on FPGAs. From FPGA usage as accelerators, speed-up factors by up to two orders of magnitude are reported. Algorithmic cleverness is the secret of success, based on software to configware migration mechanisms, striving away from memory-cycle-hungry instruction-stream-based computing paradigms also using memory-cycle-hungry data movement methods. A new kind operating system modules as well as new compilation techniques are needed to ease usage.
A main hurdle on the way to new horizons of cheap highest performance are CS-related educational deficits causing the configware / software chasm and a methodology fragmentation between the different cultures of application domains. New directions in CS-related curricula are needed which are focused on a dual paradigm approach (instruction-stream-based and data-stream-based).