Distributed convergence detection based on global residual error under asynchronous iterations

Abstract

Convergence of classical parallel iterations is detected by performing a reduction operation at each iteration in order to compute a residual error relative to a potential solution vector. To efficiently run asynchronous iterations, blocking communication requests are avoided, which makes it hard to isolate and handle any global vector. While some termination protocols were proposed for asynchronous iterations, only very few of them are based on global residual computation and guarantee effective convergence. But the most effective and efficient existing solutions feature two reduction operations, which constitutes an important factor of termination delay. In this paper, we present new, non-intrusive, protocols to compute a residual error under asynchronous iterations, requiring only one reduction operation. Various communication models show that some heuristics can even be introduced and formally evaluated. Extensive experiments with up to 5600 processor cores confirm the practical effectiveness and efficiency of our approach.

Cited by in corpus (9)

• A New Cyclic Gradient Method Adapted to Large-Scale Linear Systems
• Accurate Coarse Residual for Two-Level Asynchronous Domain Decomposition Methods
• Recent Developments in Iterative Methods for Reducing Synchronization
• On Extensions of Limited Memory Steepest Descent Method
• Using asynchronous simulation approach for interactive simulation
• Asynchronous scalable version of the Global-Local non-invasive coupling
• Distributed asynchronous convergence detection without detection protocol
• Convergence Detection of Asynchronous Iterations based on Modified Recursive Doubling
• Asynchronous parareal time discretization for partial differential equations