New parallel programming language design: a bridge between brain models and multi-core/many-core computers?

TL;DR

This paper explores designing a parallel programming language, Agapia, to connect brain activity models with multi-core computer architectures, emphasizing long temporal patterns for efficient, intelligent processing.

Contribution

It introduces Agapia as a language bridging brain models and multi-core systems through long temporal structures and interactive programming.

Findings

Agapia effectively links brain activity models with multi-core architectures.

Long temporal patterns enhance intelligent processing and speculative execution.

The approach offers a unified framework for brain-inspired and computer architectures.

Abstract

The recurrent theme of this paper is that sequences of long temporal patterns as opposed to sequences of simple statements are to be fed into computation devices, being them (new proposed) models for brain activity or multi-core/many-core computers. In such models, parts of these long temporal patterns are already committed while other are predicted. This combination of matching patterns and making predictions appears as a key element in producing intelligent processing in brain models and getting efficient speculative execution on multi-core/many-core computers. A bridge between these far-apart models of computation could be provided by appropriate design of massively parallel, interactive programming languages. Agapia is a recently proposed language of this kind, where user controlled long high-level temporal structures occur at the interaction interfaces of processes. In this paper Agapia is used to link HTMs brain models with TRIPS multi-core/many-core architectures.