Adaptive Computing Workshop Sept 2012
We are pleased to announce this workshop on Adaptive Computing in Sept 2012.
Humans are adaptive. Other organisms are Adaptive. Should Computers and Computing be Adaptive? Yes, of course !!
While there is a lot of focus on a Design’s Infant Mortality and Aging of the Design’s implementation (manifestation of faults), significant research needs to be done on what all can be done in normal mode.
The speakers at this workshop intend to provide that insight.
When and where is this workshop?
This workshop will be held at the Indian Institute of Science, Bangalore. It will take place on Sept 14th, 2012. For further details, please visit accsindia.org/accs-adaptive-computing.html
Regards
is this different from the flexible processor architecture [fpl] work at stanford? Their arhitecture too proivdes for dynamic recon guration and very efficient hybrid and coarse grain techniques. . Increasing instruction level parallelism and clock rates will not increase system performance. it has been the memory access time and bandwidth that impedes system performance. are u considering fail-safe computation issues?
The workshop is indeed meant to represent different views on the issues raised here. Among the Metrics not well understood is the “degree of reconfigurability” and its relationship to time to reconfigure and post reconfiguration performance. Memory access time issue is sometimes mitigated by large local memory, but this as pointed out is a critical issue.
We are currently documenting various Published Architecture. Failure of any kind and failure management of any kind is (by definition) beyond the scope of the Workshop. Our own effort is also related to fully funcational systems. Welcome efforts in Fault Tolerance although it is already overcrowded field!!
Dr. Kumar, I wrote to you earlier on this topic. I found the link to the early work done at Stanford on adaptive computing. ftp://arith.stanford.edu/tr/flynn_lid.pdf
I think memory management will be your key challenge.
Syed
Syed, Thanks for the link. Shall look it up. There are many TRs from Stanford’s Centre for Reliable Computing. Please look at them , sorry dont have the links anymore. We did enhance some oftheCRComputing work recently. Yet to be approved for publication.