high performance computing on graphics processing units: hgpu.org

We present Piko, a design framework for designing efficient programmable graphics pipelines. Piko is built around managing work granularity in a programmable and flexible manner, allowing programmers to build load-balanced parallel pipeline implementations, to exploit spatial and producer-consumer locality in the pipeline, and to explore tradeoffs between these considerations. Piko programmers describe a pipeline as a series of stages factored into AssignBin, Schedule, and Process phases, and the Piko compiler, Pikoc, analyzes and synthesizes this pipeline, outputting a set of optimized kernels that implement the pipeline and a header file containing code to launch the pipeline.