Fault Tolerance in an Inner-Outer Solver: A GVR-Enabled Case Study
نویسندگان
چکیده
Resilience is a major challenge for large-scale systems. It is particularly important for iterative linear solvers, since they take much of the time of many scientific applications. We show that single bit flip errors in the Flexible GMRES iterative linear solver can lead to high computational overhead or even failure to converge to the right answer. Informed by these results, we design and evaluate several strategies for fault tolerance in both inner and outer solvers appropriate across a range of error rates. We implement them, extending Trilinos’ solver library with the Global View Resilience (GVR) programming model, which provides multi-stream snapshots, multi-version data structures with portable and rich error checking/recovery. Experimental results validate correct execution with low performance overhead under varied error conditions.
منابع مشابه
Return to Sender: Finding Fault with the Court’s No-fault Gvr Practice in “confession of Error” Cases
When considering the ability of this nation’s highest court to control its docket, the first procedure to leap to mind is most likely the Writ of Certiorari. By denying certiorari review, the Court promptly removes a case from its docket without ever speaking on the issues it presents. However, while it might be the most familiar, denying cert is not the only method available to the Court for p...
متن کاملEfficient Implementation of Inner-Outer Flexible GMRES for the Method of Moments Based on a Volume-Surface Integral Equation
This paper presents flexible inner-outer Krylov subspace methods, which are implemented using the fast multipole method (FMM) for solving scattering problems with mixed dielectric and conducting object. The flexible Krylov subspace methods refer to a class of methods that accept variable preconditioning. To obtain the maximum efficiency of the inner-outer methods, it is desirable to compute the...
متن کاملFrequency Aanalysis of Annular Plates Having a Small Core and Guided Edges at Both Inner and Outer Boundaries
This paper deals with frequency analysis of annular plates having a small core and guided edges at both inner and outer boundaries. Using classical plate theory the governing differential equation of motion for the annular plate having a small core is derived and solved for the case of plate being guided at inner and outer edge boundaries. The fundamental frequencies for the first six modes of ...
متن کاملA Study of the Diagnostic Amplitude of Rolling Bearing under Increasing Radial Clearance Using Modulation Signal Bispectrum
The rolling element bearing is a key part of machines. The accurate and timely diagnosis of its faults is critical for predictive maintenance. Most researches have focused on the fault location identification. To estimate the fault severity accurately, this paper focuses on the study of roller bearing vibration amplitude under increasing radial clearances due to inevitable wear using the modula...
متن کاملExact Elasticity Solutions for Thick-Walled FG Spherical Pressure Vessels with Linearly and Exponentially Varying Properties
In this paper, exact closed-form solutions for displacement and stress components of thick-walled functionally graded (FG) spherical pressure vessels are presented. To this aim, linear variation of properties, as an important case of the known power-law function model is used to describe the FG material distribution in thickness direction. Unlike the pervious studies, the vessels can have arbit...
متن کامل