Describes the motivations for and design of a massively parallel computer with many (millions? 64k?) of very simple processing nodes that serve as both memory and CPU. Shows how you can have "active data structures" where each item of data is in a different CPU (or clusters of nearby CPUs) and literally migrates between CPUs as computation proceeds. I think it gives an example of the parallel bitonic sort algorithm being executed in this way. A relatively thin book and relatively easy to read for a programmer.
https://archive.org/details/connectionmachin00hill
Describes the motivations for and design of a massively parallel computer with many (millions? 64k?) of very simple processing nodes that serve as both memory and CPU. Shows how you can have "active data structures" where each item of data is in a different CPU (or clusters of nearby CPUs) and literally migrates between CPUs as computation proceeds. I think it gives an example of the parallel bitonic sort algorithm being executed in this way. A relatively thin book and relatively easy to read for a programmer.
You can also read Hillis's thesis:
https://dspace.mit.edu/bitstream/handle/1721.1/14719/1852428...