StarPU: Task-Based Runtime System for Heterogeneous HPC Clusters

StarPU runtime system
Team STORM’s StarPU runtime system provides a framework for task scheduling on heterogeneous, accelerated platforms, together with an API for implementing various classes of scheduling algorithms. This scheduling framework jointly works with a distributed shared-memory manager to optimize task mappings and data transfers, and to overlap communications with computations.

Link to StarPU’s web-page

Comments are closed.