An automatic tool for high-level algorithmic complexity evaluation and optimization for system design

The increasing complexity of nowadays applications has lead to the need of more and more intensive specification and validation tasks by means of software implementations. Besides, as the complexity grows, it is becoming very hard to understand the specific processing needs and to take appropriate architectural choices at the beginning of the design cycle, so as to enable the fast design of effective solutions. This paper presents a new automatic tool for high-level algorithmic complexity analysis for the design of complex and heterogeneous systems, the Software Instrumentation Tool (SIT). SIT allows to analyze the complexity of an application directly from its verification model, without the need of writing new descriptions, re-writing the existing source-code or manually inserting instrumentation commands. SIT analysis results are an accurate and detailed estimate of the complexity and provide the designer with a meaningful and reliable basis for the preliminary analysis phase of the design-flow and for studying local- and system-wide optimizations.