Performance-Driven Interface Contract Enforcement for Scientific Components

Abstract

For their interest in and encouragement of my research from its fledgling stages,

Gaining confidence in scientific applications through executable interface contracts

“…Research thus far focuses on identifying and developing interface contract enforcement sampling techniques driven by execution time overhead limits [12,13,14,15]. Experiments are conducted to investigate the associated impacts.…”

High-performance language interoperability for scientific computing through Babel

“…Babel supports three clauses: preconditions, postconditions, and class invariants [14]. The preconditions for vuDot require the two normal SIDL arrays, u and v, both be nonnull, one-dimensional arrays of the same size.…”

“…Support for a variety of enforcement options across programming languages also requires option management. Finally, although earlier work pursued distributed enforcement decision processes [15], [16], the goal of trying to better control enforcement overhead lead to the current release containing a centralized enforcement manager [13].…”

“…Babel supports traditional and experimental contract enforcement through runtime options combining the classification of contract clauses to be enforced with the frequency of their enforcement. This approach was considered to be the most flexible for not only enforcement but also classification purposes [14]. Table III lists the more common options for each criteria.…”

“…This raises the issue of language interoperability which involves making calls between languages, converting basic types, and bridging disparate programming models. Babel provides a featurerich, extensible, high-performance solution to the language interoperability problem currently supporting C, C++, FORTRAN 77, Fortran 90/95, Fortran 2003/2008, Python, and Java. Babel supports object-oriented programming features and interface semantics with runtime-enforcement.…”