%%at most 350 words.


Many important computational problems, including those of computer
vision, are characterized by data-parallel, low-precision integer
operations on large volumes of data. For such highly structured
problems, this thesis develops Abacus, a high-speed reconfigurable
SIMD (single-instruction, multiple-data) architecture that outperforms
conventional microprocessors by over an order of magnitude using the
same silicon resources.

Earlier SIMD systems computed at relatively slow clock rates compared
to their uniprocessor counterparts. The thesis discusses the problems
involved in operating a large SIMD system at high clock rates,
including instruction distribution and chip-to-chip communication,
presents the solutions adopted by the Abacus design.
    
Although the chip was implemented in a 1989-era VLSI technology, it
was designed to contain 1024 processing elements (PEs), operate at 125
MHz, and deliver 2 billion 16-bit arithmetic operations per second
(GOPS).  The PE and chip architecture are described in detail, as well
as the results of testing the chip at 100 MHz.

Despite this high performance, the Abacus one-bit ALU is not the
optimal point in the design space. An analytical model is developed
for performance as a function of ALU width and off-chip memory
bandwidth. Intuition provided by the model leads to the conclusion
that an eight-bit ALU is an optimal choice for the current technology.

Finally, using the analytical model, area and time parameters from the
Abacus chip, and some lessons learned from the chip implementation, a
design is presented for a 320 GOPS low-cost single-board system.