An Enhanced Profiling Framework for the Analysis and Development of Parallel Primitives for GPUs

Parallelizing software applications through the use of existing optimized primitives is a common trend that mediates the complexity of manual parallelization and the use of less efficient directive-based programming models. Parallel primitive libraries allow software engineers to map any sequential code to a target many-core architecture by identifying the most computational intensive code sections and mapping them into one ore more existing primitives. On the other hand, the spreading of such a primitive-based programming model and the different GPU architectures have led to a large and increasing number of thirdparty libraries, which often provide different implementations of the same primitive, each one optimized for a specific architecture. From the developer point of view, this moves the actual problem of parallelizing the software application to selecting, among the several implementations, the most efficient primitives for the target platform. This paper presents a profiling framework for GPU primitives, which allows measuring the implementation quality of a given primitive by considering the target architecture characteristics. The framework collects the information provided by a standard GPU profiler and combines them into optimization criteria. The criteria evaluations are weighed to distinguish the impact of each optimization on the overall quality of the primitive implementation. The paper shows how the tuning of the different weights has been conducted through the analysis of five of the most widespread existing primitive libraries and how the framework has been eventually applied to improve the implementation performance of a standard primitive.