🚀 go-pugleaf

          The Rewrite Rule Machine and the Maude Language


                        Jose Meseguer
                  Computer Science Laboratory
                      SRI International


The Rewrite Rule Machine (RRM) is a massively parallel computer that we are
currently developing at SRI.  Because of its very fine grain SIMD parallelism
at the chip level combined with its flexible coarser grain MIMD parallelism at
the network level that allows different chips to work on different subtasks of
the same problem at once, the RRM can exploit a problem's parallelism at
several levels.  This makes the RRM very well suited for solving complex,
locally homogeneous but globally inhomogeneous, problems in many areas,
including discrete event simulation, expert systems, rapid prototyping,
vision, computational geometry, automated deduction, finite element methods,
hardware simulation, and neural networks.

Finding simple high level programming languages well suited for programming
MIMD machines is an important research problem; in the case of the RRM a
further requirement is that the language should also be well suited for data
parallelism.  Because of its flexible abstract model of computation called
concurrent rewriting, the RRM is programmable in a wide variety of declarative
high level languages that permit massive exploitation of implicit parallelism.
In particular, the RRM is programmable in Maude, a declarative language that
unifies the paradigms of functional and concurrent object-oriented
programming.

The talk will give a brief overview of Maude and the concurrent rewriting
model of computation, will summarize the architectural design of the RRM, and
will discuss performance estimates for the RRM based on detailed simulations
at the chip level together with additional simulations and modelling for
higher levels.

=-=-=-=-=-=-=-=-=-=-=
Location:	Electrical Engineering Building room 248, USC Campus
Time:		3:30 PM, Thursday, 9 April 1992
Host:		Prof. Alvin M. Despain, Advanced Computer Architecture Lab
Information:	Ms. Susan Moore, 213-740-4462, susanm@usc.edu