high performance computing on graphics processing units: hgpu.org

Space is a circuit oriented, spatial programming language designed to exploit the massive parallelism available in a novel formal model of computation called the Synchronic A-Ram, and physically related FPGA and reconfigurable architectures. Space expresses variable grained MIMD parallelism, is modular, strictly typed, and deterministic. Barring operations associated with memory allocation and compilation, modules cannot access global variables, and are referentially transparent. At a high level of abstraction, modules exhibit a small, sequential state transition system, aiding verification. Space deals with communication, scheduling, and resource contention issues in parallel computing, by resolving them explicitly in an incremental manner, module by module, whilst ascending the ladder of abstraction. Whilst the Synchronic A-Ram model was inspired by linguistic considerations, it is also put forward as a formal model for reconfigurable digital circuits. A programming environment has been developed, that incorporates a simulator and compiler that transform Space programs into Synchronic A-Ram machine code, consisting of only three bit-level instructions, and a marking instruction. Space and the Synchronic A-Ram point to novel routes out of the parallel computing crisis.