J. D. Roberts, R. F. DeMara, G. Ellis, R. Hughey, R. Levinson, and C. Noshpitz, "AHP: Advanced Hardware for PIERCE," in Proceedings of the Second International Workshop on PIERCE, pp. 26 - 29, Quebec, Canada, August 7, 1993. Abstract: PEIRCE is already being ported to a variety of platforms, both workstations and personal computers, by other researchers. The AHP group is exploring advanced, high performance hardware platforms. Our work focuses almost exclusively on parallel processing, including a system under development tailored for CG processing. The goal is to support PEIRCE with hardware platforms that can dramatically improve performance and in turn increase the size of CG research domains and applications. To date, we have been unsuccessful in obtaining information from application that is necessary for evaluation processing needs and characteristics. We urge those who have gathered statistics(averages or distributions on number of hierarchies,etc.)For applications, or who have profiled runtimes versus problem size and characteristics to contact us. We should also like to obtain PEIRCE DBs in order to extract such information ourselves. Finally, we welcome others who would like to join this hardware group. Research over last year has shown that the low-level routines required by PEIRCE can be efficiently parallelized on a SIMD architecture (single-instruction multiple-data stream massive parallelism) with near linear speedup(doubling the number of processors approximately halves computation time).It has also been demonstrated that PEIRCE can be ported to a MIMD(multiple-instruction multiple- data stream multiprocessing)architecture. We have analyzed how processing requirement and speed scales in problem size for CG DB queries and evaluated communication requirements and processor allocation. The most crucial work remaining includes extending these results to other operations(such as maximal join and close matches) and issues, and then to remaining PEIRCE modules.