🚀 go-pugleaf

I need to select a textbook for a new course this spring at Cornell,
"Parallel Scientific Computing."  We are planning to use some of the
machines here including:

  (1) SIMD: Connection Machine CM-200
  (2) MIMD, tightly coupled: Intel iPSC
  (3) MIMD, loosely coupled: network of IBM RS6000's
  (4) shared memory: Kendall Square KSR-1

I am hoping to cover these architecture paradigms.  I am also hoping
to cover associated programming paradigms, for instance, FORTRAN-90
for the CM-200.

Then I am hoping to cover scientific algorithms for parallel computing
(and specifically for the machines above) including linear algebra,
systems of particles, boundary value problems, and perhaps some global
climate modeling.  In the process we will hit the paradigms for
parallel programming, such as load balancing and minimizing
communication.

I would be interested in suggestions for textbooks that cover this
material, especially from people who have taught a course like this
before.

Thanks,
Steve Vavasis (vavasis@cs.cornell.edu)