评估SDC对GMRES迭代求解器的影响
Evaluating the Impact of SDC on the GMRES Iterative Solver
摘要:
随着并行性和晶体管密度的增加,以及越来越严格的能量和峰值功率约束,可能会迫使偶尔不正确的计算或存储暴露给应用程序代码。静默数据损坏(SDC)可能不常见,但一个SDC足以使迭代线性求解器等数值算法停止向正确答案前进。因此,我们专注于弹性的迭代线性求解器GMRES到一个单一的瞬态SDC。我们推导出廉价的检查来检测GMRES中的SDC的影响,该GMRES工作于比假定位翻转更一般的SDC模型。我们的实验表明,当GMRES被用作内部求解器的内部-外部迭代,它可以“贯穿”的SDC的几乎任何幅度的计算密集型正交化阶段。也就是说,它使用错误的数据得到正确的答案,而不需要任何回滚。那些它不能穿过的SDC,被我们的检测方案捕获。
英文摘要:
Increasing parallelism and transistor density, along with increasingly tighter energy and peak power constraints, may force exposure of occasionally incorrect computation or storage to application codes. Silent data corruption (SDC) will likely be infrequent, yet one SDC suffices to make numerical algorithms like iterative linear solvers cease progress towards the correct answer. Thus, we focus on resilience of the iterative linear solver GMRES to a single transient SDC. We derive inexpensive checks to detect the effects of an SDC in GMRES that work for a more general SDC model than presuming a bit flip. Our experiments show that when GMRES is used as the inner solver of an inner-outer iteration, it can "run through" SDC of almost any magnitude in the computationally intensive orthogonalization phase. That is, it gets the right answer using faulty data without any required roll back. Those SDCs which it cannot run through, get caught by our detection scheme.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。