An Infrastructure for Scalable Parallel Programs for Computational Chemistry

Parallel implementation of complex computational software is a difficult task. This paper describes the Super Instruction Architecture (SIA) and its application to the implementation of algorithms for electronic structure computational chemistry calculations. The methods are programmed in a domain specific programming language called super instruction assembly language (SIAL), which is based on the abstractions of super instructions and super numbers. Compiled SIAL programs are executed by a parallel virtual machine known as the Super Instruction Processor (SIP). All compute intensive operations, such as tensor contractions, diagonalizations, etc, all communication operations and all input-output operations are handled by the SIP. By separating the algorithmic complexity in the SIAL from the complexities of parallel execution on computer hardware in the SIP, a software system has been created that allows for very effective optimization and tuning on different hardware architectures with quite manageable effort.