Algorithms for Parallel-Search Memories

Abstract

The underlying logical structure of parailel-search memories is described; the characteristic operation of three major types is displayed in the execution of searches based on equality; and algorithms are presented for searches based on other specifications including maximum, minimum, greater than, less than, nearest to, between limits, and ordering (sorting). It is shown that there is a hierarchy of dependency among these algorithms, that they appear in pairs with each member of a pair belonging to one or the other of two distinct classes, and that every type of search can be executed within each class.

SWAMP: Smith-Waterman using associative massive parallelism

“…The IS is able to broadcast data to all of its active PEs, flip the responders and the non-responders, as well as select a single PE from the set of responder PEs. An IS may compute the global AND, OR, MAX or MIN of data in all active PEs in constant time [13] [14]. The MAX reduction is particularly useful for this local alignment algorithm since the maximum C value is returned and used for the traceback.…”

Associative Processor Architecture—a Survey

“…An associative memory has been also called catalog memory [11], contentaddressed memory [12], data-addressed memory [13], parallel search memory [14], search memory [15][16][17], search associative memory [18], content-add~'essable memory [19], distributed logic memory [20], associative pushdown memory [21], and multi-access associative memory [22].…”

Extreme-Value Search and General Selection Algorithms for Fully Parallel Associative Memories

“…Although such operations can be programmed using the basic 'masked exact match' search capability of simple AMs, provision of other types of hardware primitives can have a significant effect on performance. Starting with the pioneering works of Falkoff [15] and Estrin and Fuller [16], research in associative memory algorithms has continued up to the present [17][18][19][20][21][22][23].…”