Avoiding checkpoint contamination in parallel systems

The execution times of large-scale parallel applications on nowadays multi/many-core systems are usually longer than the mean time between failures. Therefore, parallel applications must tolerate hardware failures to ensure that not all computation done is lost on machine failures. Checkpointing and rollback recovery is one of the most popular techniques to implement fault-tolerant applications. However, checkpointing parallel applications is expensive in terms of computing time, network utilization and storage resources. Thus, current checkpoint-recovery techniques should minimize these costs in order to be useful for large scale systems. In this paper three different and complementary techniques to reduce the size of the checkpoints generated by application-level checkpointing are proposed and implemented. Detailed experimental results obtained on a multicore cluster show the effectiveness of the proposed methods to reduce checkpointing cost.

Parallel computer systems are among the most complex of man's creations, making satisfactory performance characterization difficult. Despite this complexity, there are strong, indeed, almost irresistible, incentives to quantify parallel system performance using a single metric. The fallacy lies in succumbing to such temptations. A complete performance characterization requires not only an analysis of the system's constituent levels, it also requires both static and dynamic characterizations. Static or average behavior analysis may mask transients that dramatically alter system performance. Although the human visual system is remarkedly adept at interpreting and identifying anomalies in false color data, the importance of dynamic, visual scientific data presentation has only recently been recognized Large, complex parallel system pose equally vexing performance interpretation problems. Data from hardware and software performance monitors must be presented in ways that emphasize important events while eluding irrelevant details. Design approaches and tools for performance visualization are the subject of this paper.

CFD problems numerical simulation and visualization by means of parallel computational system

High performance computers have played key roles in many scientific and engineering advances over the past 40 years, and many more may be expected in the future. However, unless practical parallel systems can be produced in this decade, a performance crisis will arise by 2000 across the spectrum of systems from workstations to supercomputers. There is widespread confusion today about how best to proceed with future parallel systems because so many different approaches have been taken and the performance results have been so spotty. A fundamental flaw in our approach to parallel computing, as a nation, is the poor understanding we have obtained about delivered performance. This paper analyzes the situation and suggests fundamental changes that are necessary to achieve practical parallelism in this decade. A great deal of money is now being spent and more is planned, to advance the field, but money is not so much the problem as shortages of qualified people and a sharp focus for their work. Our national goals for the end of this decade must be the creation of an infrastructure for understanding performance, and its natural consequence, the development of practical parallel systems.

Responsive systems are fault-tolerant, real-time systems in parallel and distributed environments. Recently, several approaches to responsive system design have been developed. By incorporating key concepts in these approaches to communication protocols, responsive protocols are defined as protocols that can revert to a normal state in bounded time, even when they enter abnormal states due to faults. Fundamentally, communication protocols can be made responsive by adding timely checkpointing and rollback recovery to them. In this paper, we discuss fault-tolerant, real-time performance (called responsiveness) of such protocols and propose an algorithm to compute responsiveness of protocols whose reachability graphs form acyclic directed graphs. Finally, we also present several simulation results to analyze dynamic behaviors of such protocols.

Interconnection networks have been deployed as the communication fabric in a wide spectrum of parallel computer systems, ranging from chip multiprocessors (CMPs) and embedded multicore systems-on-a-chip (SoCs) to clusters and server blades. Recent technology trends have permitted a rapid growth of chip resources, faster clock rates, and wider communication bandwidths, however, these trends have also led to an increase in power consumption that is becoming a key limiting factor in the design of such scalable interconnected systems. Power-aware networks, therefore, need to become inherent components of single and multi-chip parallel systems. In the hardware arena, recent interconnection network power-management research work has employed limited-scope techniques that mostly focus on reducing the power consumed by the network communication links. As these limited-scope techniques are not tailored to the applications running on the network, power savings and the corresponding impact on network latency vary significantly from one application to the next as we demonstrate in this paper; in many cases, network performance can severely suffer. In the software arena, extensive research on compile-time optimizations has produced parallelizing compilers that can efficiently map an application onto hardware for high performance. However, research into power-aware parallelizing compilers is in its infancy. In this paper, we take the first steps toward tailoring applications' communication needs at run-time for low power. We propose software techniques that extend the flow of a parallelizing compiler in order to direct run-time network power reduction. We target network links, a significant power consumer in these systems, allowing dynamic voltage scaling (DVS) instructions extracted during static compilation to orchestrate link voltage and frequency transitions for power savings during application run-time. Concurrently, an online hardware mechanism measures network congestion levels and adapts these off-line DVS settings to maximize network performance. Our simulations over three existing parallel systems, ranging from very fine-grained single-chip to coarse-grained multi-chip architectures, show that link power consumption can be reduced by up to 76.3%, with a minor increase in latency, ranging from 0.18 to 6.78% across a number of benchmark suites.

A Schwarz domain decomposition method for solution of transient unsaturated water flow on parallel computers

The nodes in mobile ad hoc networks (MANETs) are susceptible to various types of failures and therefore the protocols designed for these networks are required to be fault tolerant. Checkpointing and rollback recovery is a well explored fault tolerance technique for the static and cellular mobile distributed systems. However, it is not trivially applicable to MANETs due to their dynamic topology and constraints of limited resources such as wireless bandwidth and stable storage. The paper combines asynchronous message logging with independent checkpointing for a successful rollback recovery procedure adapted to the ad hoc environment. Logging the messages asynchronously significantly reduces the number of accesses to stable storage. The recovery procedure involves only those processes whose state depends on the state of the failed process. The presented scheme requires a low number of control messages, reduces the size of control information carried by application messages and can also handle the movement of nodes across the network.

Introduces the design and implementation of a Remote System for Parallel Computing.Based on Multi-platform.This system support remote operating and running linux system parallel programes on Windows system and provide a kind of uniform mode for the applied customer to operation, The customer does not need to know the concrete structure of the parallel computing system,can operate the parallel system and complete the computation or simulation what they need.This system reduces the complication on developing parallel programs and improves the efficiency on using parallel computing system,and solves the problem of real-time data transmission between server and client machine.It has highly practical value.

Large mainframe computers are slowly dying off, as massively parallel computer systems are replacing them to become the way of the future which offers scalability, redundancy and system availability not possible previously. These new computers require power systems that are equally capable in scalability, redundancy and availability. This paper explores the system availability and MTBF requirements of distributed high power, paralleled, fault-tolerant power systems, which are necessary to match the needs of parallel computer systems within the next five years. >

Chapter 1 - Introduction

Several existing parallel computer systems partition their system resources or consist of systems from different geographical locations. This work is focusing on extending the original goal-oriented parallel computer job scheduling policies to cover such systems. The goal-oriented parallel computer job scheduling policies are proposed recently to handle conflicting objectives by utilizing a combinatorial search technique to find the most compromise schedule within a time limit. In this paper, some modifications to the original goal-oriented parallel computer job scheduling policy design are proposed and evaluated. The proposed policy is evaluated against basic priority backfilling techniques widely used in the field. Both homogeneous and heterogeneous parallel computer systems in terms of the computing power are evaluated. And, the design decision on the partition selection heuristic is also evaluated in this study. The experimental results show that the proposed policy produces good scheduling performances even when inaccurate runtime information is used.

For this special issue we have selected five papers that address, from several points of view, the problem of efficient utilization of resources in parallel and distributed systems. These papers were among the best papers presented at the IASTED PDCS 2001 conference. The topics covered include: efficient cache strategies for simultaneous execution of threads as well as for the distribution of video-on-demand, efficient communication and failure recovery, and run-time support for the automatic parallelization of dynamic structures

Checkpointing techniques have widely been studied in the literature as a way to recover from failures in sequential, distributed and parallel environments. However, most of the checkpointing mechanisms proposed so far focus only on the recovery of the application data. If the application performs some I/O operations to disk files, such schemes may not work correctly, as they do not provide rollback-recovery for the file contents. In this paper, we present a distributed checkpointing mechanism for a Parallel File System that can be integrated with any of the previous application checkpointing algorithms. Three different file checkpointing schemes will be presented, tested in that mechanism and discussed in detail. The distributed mechanism proposed was integrated in PIOUS - a public-domain parallel file system developed for the PVM distributed computing environment.

For better protection of distributed systems, two well-known techniques are: checkpointing and rollback recovery. While failure protection is often considered a separate issue, it is crucial for establishing better secure distributed systems. This article is dedicated to the proposal of a new checkpointing algorithm for saving a consistent-transaction state in distributed databases ensuring that database management systems are able, after a failure, to recover the state of the database. The proposed communication-induced algorithm does not hamper the normal transaction processing and saves a global consistent-transaction state that records only the fully completed transactions. Analysis and experimental results of our proposal show that the proposed scheme saves a minimum number of forced checkpoints and has some performance gains compared to the alternative approaches.