DMTCP: Transparent checkpointing for cluster computations and the desktop
23 May 2009-pp 1-12
TL;DR: DMTCP as mentioned in this paper is a transparent user-level checkpointing package for distributed applications, which is used for the runCMS experiment of the Large Hadron Collider at CERN, and it can be incorporated and distributed as a checkpoint-restart module within some larger package.
Abstract: DMTCP (Distributed MultiThreaded CheckPointing) is a transparent user-level checkpointing package for distributed applications. Checkpointing and restart is demonstrated for a wide range of over 20 well known applications, including MATLAB, Python, TightVNC, MPICH2, OpenMPI, and runCMS. RunCMS runs as a 680 MB image in memory that includes 540 dynamic libraries, and is used for the CMS experiment of the Large Hadron Collider at CERN. DMTCP transparently checkpoints general cluster computations consisting of many nodes, processes, and threads; as well as typical desktop applications. On 128 distributed cores (32 nodes), checkpoint and restart times are typically 2 seconds, with negligible run-time overhead. Typical checkpoint times are reduced to 0.2 seconds when using forked checkpointing. Experimental results show that checkpoint time remains nearly constant as the number of nodes increases on a medium-size cluster. DMTCP automatically accounts for fork, exec, ssh, mutexes/ semaphores, TCP/IP sockets, UNIX domain sockets, pipes, ptys (pseudo-terminals), terminal modes, ownership of controlling terminals, signal handlers, open file descriptors, shared open file descriptors, I/O (including the readline library), shared memory (via mmap), parent-child process relationships, pid virtualization, and other operating system artifacts. By emphasizing an unprivileged, user-space approach, compatibility is maintained across Linux kernels from 2.6.9 through the current 2.6.28. Since DMTCP is unprivileged and does not require special kernel modules or kernel patches, DMTCP can be incorporated and distributed as a checkpoint-restart module within some larger package.
Citations
More filters
23 Jun 2014
TL;DR: Snapify is introduced, a set of extensions to MPSS that provides three novel features for Xeon Phi offload applications: checkpoint and restart, process swapping, and process migration that reduce the PCI latency of storing and retrieving process snapshots.
Abstract: Intel Xeon Phi coprocessors provide excellent performance acceleration for highly parallel applications and have been deployed in several top-ranking supercomputers. One popular approach of programming the Xeon Phi is the offload model, where parallel code is executed on the Xeon Phi, while the host system executes the sequential code. However, Xeon Phi's Many Integrated Core Platform Software Stack (MPSS) lacks fault-tolerance support for offload applications. This paper introduces Snapify, a set of extensions to MPSS that provides three novel features for Xeon Phi offload applications: checkpoint and restart, process swapping, and process migration. The core technique of Snapify is to take consistent process snapshots of the communicating offload processes and their host processes. To reduce the PCI latency of storing and retrieving process snapshots, Snapify uses a novel data transfer mechanism based on remote direct memory access (RDMA). Snapify can be used transparently by single-node and MPI applications, or be triggered directly by job schedulers through Snapify's API. Experimental results on OpenMP and MPI offload applications show that Snapify adds a runtime overhead of at most 5%, and this overhead is low enough for most use cases in practice.
18 citations
TL;DR: A mathematical model is developed to estimate the average execution time of a program in the presence of failures, without and with application level checkpointing, and it is used to predict the optimum interval number of instructions which should be executed between the placement of successive checkpoints.
17 citations
Posted Content•
TL;DR: In this article, a non-invasive, cloud-agnostic approach is demonstrated for extending existing cloud platforms to include checkpoint-restart capability, which allows traditional HPC applications to take advantage of an existing cloud infrastructure.
Abstract: A non-invasive, cloud-agnostic approach is demonstrated for extending existing cloud platforms to include checkpoint-restart capability. Most cloud platforms currently rely on each application to provide its own fault tolerance. A uniform mechanism within the cloud itself serves two purposes: (a) direct support for long-running jobs, which would otherwise require a custom fault-tolerant mechanism for each application; and (b) the administrative capability to manage an over-subscribed cloud by temporarily swapping out jobs when higher priority jobs arrive. An advantage of this uniform approach is that it also supports parallel and distributed computations, over both TCP and InfiniBand, thus allowing traditional HPC applications to take advantage of an existing cloud infrastructure. Additionally, an integrated health-monitoring mechanism detects when long-running jobs either fail or incur exceptionally low performance, perhaps due to resource starvation, and proactively suspends the job. The cloud-agnostic feature is demonstrated by applying the implementation to two very different cloud platforms: Snooze and OpenStack. The use of a cloud-agnostic architecture also enables, for the first time, migration of applications from one cloud platform to another.
15 citations
05 May 2011
TL;DR: A new approach to checkpoint and an original optimization on the checkpoint structure that is implemented and evaluated to make incremental checkpointing more efficient and more appropriate, especially for CAPE.
Abstract: Checkpointing is an important method for providing fault tolerance, load balancing, process migration, periodic backup, and many other functions. It is also the basic tool used in CAPE, a paradigm which aims at distributing the execution of a program on a distributed-memory environment. This paper presents a new approach to checkpoint and an original optimization on the checkpoint structure that we have implemented and evaluated to make incremental checkpointing more efficient and more appropriate, especially for CAPE.
15 citations
Dissertation•
01 Jan 2014
TL;DR: It is shown that the PetaBricks autotuner is often able to find non-intuitive poly-algorithms that outperform more traditional hand written solutions and is shown to perform well on complex search spaces up to 103000 possible configurations in size.
Abstract: The process of optimizing programs and libraries, both for performance and quality of service, can be viewed as a search problem over the space of implementation choices. This search is traditionally manually conducted by the programmer and often must be repeated when systems, tools, or requirements change. The overriding goal of this work is to automate this search so that programs can change themselves and adapt to achieve performance portability across different environments and requirements. To achieve this, first, this work presents the PetaBricks programming language which focuses on ways for expressing program implementation search spaces at the language level. Second, this work presents OpenTuner which provides sophisticated techniques for searching these search spaces in a way that can easily be adopted by other projects. PetaBricks is a implicitly parallel language and compiler where having multiple implementations of multiple algorithms to solve a problem is the natural way of programming. Choices are provided in a way that also allows our compiler to tune at a finer granularity. The PetaBricks compiler autotunes programs by making both fine-grained as well as algorithmic choices. Choices also include different automatic parallelization techniques, data distributions, algorithmic parameters, transformations, and blocking. PetaBricks also introduces novel techniques to autotune algorithms for different convergence criteria or quality of service requirements. We show that the PetaBricks autotuner is often able to find non-intuitive poly-algorithms that outperform more traditional hand written solutions. OpenTuner is a open source framework for building domain-specific multi-objective program autotuners. OpenTuner supports fully-customizable configuration representations, an extensible technique representation to allow for domain-specific techniques, and an easy to use interface for communicating with the program to be autotuned. A key capability inside OpenTuner is the use of ensembles of disparate search techniques simultaneously; techniques that perform well will dynamically be allocated a larger proportion of tests. OpenTuner has been shown to perform well on complex search spaces up to 103000 possible configurations in size. Thesis Supervisor: Saman Amarasinghe Title: Professor
14 citations
References
More filters
01 May 2007
TL;DR: The IPython project as mentioned in this paper provides an enhanced interactive environment that includes, among other features, support for data visualization and facilities for distributed and parallel computation for interactive work and a comprehensive library on top of which more sophisticated systems can be built.
Abstract: Python offers basic facilities for interactive work and a comprehensive library on top of which more sophisticated systems can be built. The IPython project provides on enhanced interactive environment that includes, among other features, support for data visualization and facilities for distributed and parallel computation
3,355 citations
TL;DR: An algorithm by which a process in a distributed system determines a global state of the system during a computation, which helps to solve an important class of problems: stable property detection.
Abstract: This paper presents an algorithm by which a process in a distributed system determines a global state of the system during a computation. Many problems in distributed systems can be cast in terms of the problem of detecting global states. For instance, the global state detection algorithm helps to solve an important class of problems: stable property detection. A stable property is one that persists: once a stable property becomes true it remains true thereafter. Examples of stable properties are “computation has terminated,” “ the system is deadlocked” and “all tokens in a token ring have disappeared.” The stable property detection problem is that of devising algorithms to detect a given stable property. Global state detection can also be used for checkpointing.
2,738 citations
01 Jan 1996
TL;DR: The MPI Message Passing Interface (MPI) as discussed by the authors is a standard library for message passing that was defined by the MPI Forum, a broadly based group of parallel computer vendors, library writers, and applications specialists.
Abstract: MPI (Message Passing Interface) is a specification for a standard library for message passing that was defined by the MPI Forum, a broadly based group of parallel computer vendors, library writers, and applications specialists. Multiple implementations of MPI have been developed. In this paper, we describe MPICH, unique among existing implementations in its design goal of combining portability with high performance. We document its portability and performance and describe the architecture by which these features are simultaneously achieved. We also discuss the set of tools that accompany the free distribution of MPICH, which constitute the beginnings of a portable parallel programming environment. A project of this scope inevitably imparts lessons about parallel computing, the specification being followed, the current hardware and software environment for parallel computing, and project management; we describe those we have learned. Finally, we discuss future developments for MPICH, including those necessary to accommodate extensions to the MPI Standard now being contemplated by the MPI Forum.
2,065 citations
Proceedings Article•
16 Jan 1995TL;DR: In this paper, the authors describe a portable checkpointing tool for Unix that implements all applicable performance optimizations which are reported in the literature and also supports the incorporation of user directives into the creation of checkpoints.
Abstract: Checkpointing is a simple technique for rollback recovery: the state of an executing program is periodically saved to a disk file from which it can be recovered after a failure. While recent research has developed a collection of powerful techniques for minimizing the overhead of writing checkpoint files, checkpointing remains unavailable to most application developers. In this paper we describe libckpt, a portable checkpointing tool for Unix that implements all applicable performance optimizations which are reported in the literature. While libckpt can be used in a mode which is almost totally transparent to the programmer, it also supports the incorporation of user directives into the creation of checkpoints. This user-directed checkpointing is an innovation which is unique to our work.
670 citations
Proceedings Article•
10 Apr 2005TL;DR: This is the first system that can migrate unmodified applications on unmodified mainstream Intel x86-based operating system, including Microsoft Windows, Linux, Novell NetWare and others, to provide fast, transparent application migration.
Abstract: This paper describes the design and implementation of a system that uses virtual machine technology [1] to provide fast, transparent application migration. This is the first system that can migrate unmodified applications on unmodified mainstream Intel x86-based operating system, including Microsoft Windows, Linux, Novell NetWare and others. Neither the application nor any clients communicating with the application can tell that the application has been migrated. Experimental measurements show that for a variety of workloads, application downtime caused by migration is less than a second.
588 citations