Code generation and autotuning

Annotation-based tuning workflow when using Orio. In the area of performance tuning, we are investigating annotation-based approaches to empirical performance tuning. The goal of this project is to provide a lightweight, portable source transformation tool that enables application developers to easily incorporate advanced performance-enhancing transformations into their application development process. For more information, see the Orio GitHub website.

Some of the input languages are restricted versions of general-purpose languages such as C, while others are more domain-specific in nature, e.g., targeting linear algebra or stencil computations. At present, code optimization and generation is supported for multicore CPU, GPGPU and manycore (Intel MIC) targets.

Support

DOE SciDAC grant DESC0006723: SUPER: Institute for Sustained Performance, Energy, and Resilience