Abstract
A description is given of a parallel computer architecture
called SIGMA and its implementation as a full custom design in
VLSI technology. The architecture is highly parallel, consisting
of many simple processing elements heavily interconnected. The processing
elements perform threshold computations on thousands of inputs. This
architecture was inspired by research under the "neural network" banner
and retains the highly interconnected nature of such systems. However,
it differs from them in some key areas. The SIGMA architecture is
digital, it provides greater functionality with respect to the type of
threshold comparison done, the connection weights remain static for the
duration of a problem, and its processing is deterministic. Communication
between units is in single bit values which are heavily multiplexed
to reduce the amount of physical interconnect and pinout.
Notes
We are currently acquiring citations for the work deposited into this collection. We recognize the distribution rights of this item may have been assigned to another entity, other than the author(s) of the work.If you can provide the citation for this work or you think you own the distribution rights to this work please contact the Institutional Repository Administrator at digitize@ucalgary.ca
