A novel datatype architecture support for programming languages

Abstract

In programmers point of view, Datatypes in programming language level have a simple description but inside hardware, huge machine codes are responsible to describe type features. Datatype architecture design is a novel approach to match programming features along with hardware design. In this paper a novel Data type-Based Code Reducer (TYPELINE) architecture is proposed and implemented according to significant data types (SDT) of programming languages. TYPELINE uses TEUs for processing various SDT operations. This architecture design leads to reducing the number of machine codes, and increases execution speed, and also improves some parallelism level. This is because this architecture supports some operation for the execution of Abstract Data Types in parallel. Also it ensures to maintain data type features and entire application level specifications using the proposed type conversion unit. This framework includes compiler level identifying execution modes and memory management unit for decreasing object read/write in heap memory by ISA support. This energy-efficient architecture is completely compatible with object oriented programming languages and in combination mode it can process complex C++ data structures with respect to parallel TYPELINE architecture support.