Efficient associative SIMD processing for non-tabular structured data

With recent advances in VLSI technology, it is easy to develop chips containing hundreds or thousands of processors - in effect, SIMD processing on a chip. Associative SIMD processing goes further, providing efficient parallel methods for searching tabular data. The techniques described here show that associative processing can also be extended to efficiently process non-tabular structured data (e.g., trees), as we demonstrate by supporting structure codes in our associative ASC processor.