A Framework for the Virtual Heterogeneous Associative Machine

Heterogeneous Associative Computing is a new programming paradigm that is a combination of Associative Computing [1] and Heterogeneous Computing as related to Superconcurrency (Super-C) [2]. Heterogeneous Computing comes from the realization that no one architecture is capable of performing all tasks well. Thus various architectural components are combined into one Heterogeneous System. Superconcurrency, employing code-type profiling and analytical benchmarking, aims to optimally distribute tasks to various machines such that the maximum performance is achieved for the overall problem [3]. Thus HAsC, using the combination of these two, is intended to maximize the throughput of computationally intensive tasks requiring the simultaneous power of parallel and/or vector processing of the various architectures generally found in a heterogeneous group of machines. Thus unlike other heterogeneous paradigms, it is not a primary concern of HAsC to perform equitable load balancing among the machines in the heterogeneous network of computers. HAsC instead uses data and code profiling in an attempt to match each code segment with the best processor available for that data/code type without regard for an equitable sharing of the tasks [4]. Therefore, one may consider HAsC to be the greedy heterogeneous computing paradigm. Associative Computing principles have generally been used to facilitate the execution of programs on a homogeneous group of processors within a single physical machine. HAsC broadens the reach of Associative Computing such that it now encompasses the execution of programs on distributed networks of heterogeneous machines. The original concept of Associative Computing as defined in [1] considers the Datum-PE as a computation cell; whereas HAsC considers Data-Machine as a computational cell. Data, within the context of HAsC, refers to large complex data structures such as files or groups of files rather than simple atomic or record based data structures. In the context of HAsC, Machine refers to any potential combination of complex processor(s), physical machine, or groups of physical machines; whereas Asso-ciative Computing's PE refers to a single simple processing element. HAsC also utilizes the Associative Computing broadcast mechanism to communicate instructions , parameters, and data to each cell. The associative broadcast mechanism when combined with the Associa-tive Search selection of (instruction-parameter-data type) enables polymorphism and produces a late run-time binding between VHAM node, instruction, and operation. Using Associative Search techniques each cell's VHAM Administrator decides locally whether it should execute a received instruction and whether it should accept a passed parameter. In a succinct statement, HAsC is an …