Search NASA⌕ Search

SEARCH · Search NASA

Results for “checkpoint and restart”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 records

Affinity-aware checkpoint restart

Current checkpointing techniques employed to overcome faults for HPC applications result in inferior application performance after restart from a checkpoint for a number of applications. This is due to a lack of page and core affinity awareness of the checkpoint/restart (C/R) mechanism, i.e., application tasks originally pinned to cores may be restarted on different cores, and in case of non-uniform memory architectures (NUMA), quite common today, memory pages associated with tasks on a NUMA node may be associated with a different NUMA node after restart. Here, this work contributes a novel design technique for C/R mechanisms to preserve task-to-core maps and NUMA node specific page affinities across restarts. Experimental results with BLCR, a C/R mechanism, enhanced with affinity awareness demonstrate significant performance benefits of 37%-73% for the NAS Parallel Benchmark codes and 6-12% for NAMD with negligible overheads instead of up to nearly four times longer an execution times without affinity-aware restarts on 16 cores.

97 MATHEMATICS AND COMPUTING↗

An unsupervised machine-learning checkpoint-restart algorithm using Gaussian mixtures for particle-in-cell simulations

We propose an unsupervised machine-learning checkpoint-restart (CR) algorithm for particle-in-cell (PIC) algorithms using Gaussian mixtures (GM). The algorithm compresses the particle population per spatial cell by constructing a velocity distribution function using GM. Particles are reconstructed at restart time by local resampling of the Gaussians. To guarantee fidelity of the CR process, we ensure the exact preservation of invariants such as charge, momentum, and energy for both compression and reconstruction stages, everywhere on the mesh. We also ensure the preservation of Gauss' law after particle reconstruction by exactly matching the density profile at restart time. As a result, the GM CR algorithm is shown to provide a clean, conservative restart capability while potentially affording orders of magnitude savings in input/output requirements. Here, we demonstrate the algorithm using a recently developed exactly energy- and charge-conserving PIC algorithm using both electrostatic and electromagnetic tests. The tests demonstrate not only a high-fidelity CR capability, but also its potential for enhancing the fidelity of the PIC solution for a given particle resolution.

97 MATHEMATICS AND COMPUTING↗

Leveraging the checkpoint-restart technique for optimizing CPU efficiency of ATLAS production applications on opportunistic platforms

Data processing applications of the ATLAS experiment, such as event simulation and reconstruction, spend considerable amount of time in the initialization phase. This phase includes loading a large number of shared libraries, reading detector geometry and condition data from external databases, building a transient representation of the detector geometry and initializing various algorithms and services. In some cases the initialization step can take as long as 10-15 minutes. Such slow initialization has a significant negative impact on overall CPU efficiency of the production job, especially when the job is executed on opportunistic, often short-lived, resources such as commercial clouds or volunteer computing. In order to improve this situation, we can take advantage of the fact that ATLAS runs large numbers of production jobs with similar configuration parameters (e.g. jobs within the same production task). This allows us to checkpoint one job at the end of its configuration step and then use the generated checkpoint image for rapid startup of thousands of production jobs. By applying this technique we can bring the initialization time of a job from tens of minutes down to just a few seconds. In addition to that we can leverage container technology for restarting checkpointed applications on the variety of computing platforms, in particular of platforms different from the one on which the checkpoint image was created. We will describe the mechanism of creating checkpoint images of Geant4 simulation jobs with AthenaMP (the multi-process version of the ATLAS data simulation, reconstruction and analysis framework Athena) and the usage of these images for running ATLAS Simulation production jobs on volunteer computing resources (ATLAS@Home) and on Supercomputers.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗

Checkpoint-Restart and Platform Design Criteria

Long computations often use checkpoint/restart to survive the inevitable failures. Optimal checkpoint/restart strategy can be derived in terms of special functions, and approximated within the regime of interest. As a further step under nominal assumptions, platforms can be optimized for maximum progress rate, computed in terms of non-dimensional parameter JMTTI/Delta, where JMTTI is the Job Mean Time To Interrupt and Delta is the time it takes to write (or read) a checkpoint. Scaling to a larger platform of the same technology requires scaling checkpoint bandwidth quadratically to maintain the same progress rate. This was observed during the design of LANL’s Roadrunner in 2007, led to burst buffers proposed in 2009 and introduced on Trinity in 2015. The need to compute reliably continues to impact the design of future exascale platforms.

97 MATHEMATICS AND COMPUTING↗

Asynchronous Checkpoint Migration with MRNet in the Scalable Checkpoint / Restart Library

Applications running on today's supercomputers tolerate failures by periodically saving their state in checkpoint files on stable storage, such as a parallel file system. Although this approach is simple, the overhead of writing the checkpoints can be prohibitive, especially for large-scale jobs. In this paper, we present initial results of an enhancement to our Scalable Checkpoint/Restart Library (SCR). We employ MRNet, a tree-based overlay network library, to transfer checkpoints from the compute nodes to the parallel file system asynchronously. This enhancement increases application efficiency by removing the need for an application to block while checkpoints are transferred to the parallel file system. We show that the integration of SCR with MRNet can reduce the time spent in I/O operations by as much as 15x. However, our experiments exposed new scalability issues with our initial implementation. We discuss the sources of the scalability problems and our plans to address them.

97 MATHEMATICS AND COMPUTING↗

Checkpoint-Restart in User Space

CRUISE implements a user-space file system that stores data in main memory and transparently spills over to other storage, like local flash memory or the parallel file system, as needed. CRUISE also exposes file contents fo remote direct memory access, allowing external tools to copy files to the parallel file system in the background with reduced CPU interruption.

Source record↗

Checkpoint-Restart in User Space

CRUISE implements a user-space file system that stores data in main memory and transparently spills over to other storage, like local flash memory or the parallel file system, as needed. CRUISE also exposes file contents fo remote direct memory access, allowing external tools to copy files to the parallel file system in the background with reduced CPU interruption.

Source record↗

McrEngine: A Scalable Checkpointing System Using Data-Aware Aggregation and Compression

High performance computing (HPC) systems use checkpoint-restart to tolerate failures. Typically, applications store their states in checkpoints on a parallel file system (PFS). As applications scale up, checkpoint-restart incurs high overheads due to contention for PFS resources. The high overheads force large-scale applications to reduce checkpoint frequency, which means more compute time is lost in the event of failure. We alleviate this problem through a scalable checkpoint-restart system, mcrEngine. McrEngine aggregates checkpoints from multiple application processes with knowledge of the data semantics available through widely-used I/O libraries, e.g., HDF5 and netCDF, and compresses them. Our novel scheme improves compressibility of checkpoints up to 115% over simple concatenation and compression. Our evaluation with large-scale application checkpoints show that mcrEngine reduces checkpointing overhead by up to 87% and restart overhead by up to 62% over a baseline with no aggregation or compression.

Islam, Tanzima Zerin↗

Exploring the feasibility of lossy compression for PDE simulations

Checkpoint restart plays an important role in high-performance computing (HPC) applications, allowing simulation runtime to extend beyond a single job allocation and facilitating recovery from hardware failure. Yet, as machines grow in size and in complexity, traditional approaches to checkpoint restart are becoming prohibitive. Current methods store a subset of the application’s state and exploit the memory hierarchy in the machine. However, as the energy cost of data movement continues to dominate, further reductions in checkpoint size are needed. Lossy compression, which can significantly reduce checkpoint sizes, offers a potential to reduce computational cost in checkpoint restart. This article investigates the use of numerical properties of partial differential equation (PDE) simulations, such as bounds on the truncation error, to evaluate the feasibility of using lossy compression in checkpointing PDE simulations. Restart from a checkpoint with lossy compression is considered for a fail-stop error in two time-dependent HPC application codes: PlasComCM and Nek5000. Results show that error in application variables due to a restart from a lossy compressed checkpoint can be masked by the numerical error in the discretization, leading to increased efficiency in checkpoint restart without influencing overall accuracy in the simulation.

97 MATHEMATICS AND COMPUTING↗

Final Report: Migration Mechanisms for Large-scale Parallel Applications

Process migration is the ability to transfer a process from one machine to another. It is a useful facility in distributed computing environments, especially as computing devices become more pervasive and Internet access becomes more ubiquitous. The potential benefits of process migration, among others, are fault resilience by migrating processes off of faulty hosts, data access locality by migrating processes closer to the data, better system response time by migrating processes closer to users, dynamic load balancing by migrating processes to less loaded hosts, and improved service availability and administration by migrating processes before host maintenance so that applications can continue to run with minimal downtime. Although process migration provides substantial potential benefits and many approaches have been considered, achieving transparent process migration functionality has been difficult in practice. To address this problem, our work has designed, implemented, and evaluated new and powerful transparent process checkpoint-restart and migration mechanisms for desktop, server, and parallel applications that operate across heterogeneous cluster and mobile computing environments. A key aspect of this work has been to introduce lightweight operating system virtualization to provide processes with private, virtual namespaces that decouple and isolate processes from dependencies on the host operating system instance. This decoupling enables processes to be transparently checkpointed and migrated without modifying, recompiling, or relinking applications or the operating system. Building on this lightweight operating system virtualization approach, we have developed novel technologies that enable (1) coordinated, consistent checkpoint-restart and migration of multiple processes, (2) fast checkpointing of process and file system state to enable restart of multiple parallel execution environments and time travel, (3) process migration across heterogeneous software environments, (4) network checkpoint-restart and migration of distributed and parallel applications, (5) a utility computing infrastructure for mobile desktop cloud computing based on process checkpoint-restart and migration functionality, (6) a process migration security architecture for protecting applications and infrastructure from denial-of-service attacks, and (7) a checkpoint-restart mobile computing system using portable storage devices.

99 GENERAL AND MISCELLANEOUS↗

rMPI

As high-performance computing (HPC) machines continue to grow in size, issues such as fault tolerance and reliability limit application scalability. Current techniques to ensure progress across faults, like checkpoint-restart, are unsuitable on their own for exascale machines due to the excessive overheads predicted to more than double an applications time to solution. An alternative mechanism to increase application reliability than just checkpoint-restart alone is redundant computation. The rMPl library enables portable and transparent redundant computation) that, at extreme scale, has significantly lower verhead then just checkpoint-restart on its own.

Source record↗

rMPI : increasing fault resiliency in a message-passing environment.

As High-End Computing machines continue to grow in size, issues such as fault tolerance and reliability limit application scalability. Current techniques to ensure progress across faults, like checkpoint-restart, are unsuitable at these scale due to excessive overheads predicted to more than double an applications time to solution. Redundant computation, long used in distributed and mission critical systems, has been suggested as an alternative to checkpoint-restart on its own. In this paper we describe the rMPI library which enables portable and transparent redundant computation for MPI applications. We detail the design of the library as well as two replica consistency protocols, outline the overheads of this library at scale on a number of real-world applications, and finally outline the significant increase in an applications time to solution at extreme scale as well as show the scenarios in which redundant computation makes sense.

99 GENERAL AND MISCELLANEOUS↗

Towards Low-Overhead Resilience for Data Parallel Deep Learning

Data parallel techniques have been widely adopted both in academia and industry as a tool to enable scalable training of deep learning models. At scale, DL training jobs can fail due to software or hardware bugs, may need to be preempted or terminated due to unexpected events, or may perform suboptimally because they were misconfigured. Under such circumstances, there is a need to recover and/or reconfigure data-parallel DL training jobs on-the-fly, while minimizing the impact on the accuracy of the DNN model and the runtime overhead. In this regard, state-of-art techniques adopted by the HPC community mostly rely on checkpoint-restart, which inevitably leads to loss of progress, thus increasing the runtime overhead. In this paper we explore alternative techniques that exploit the properties of modern deep learning frameworks (overlapping of gradient averaging and weight updates with local gradient computations through pipeline parallelism) to reduce the overhead of resilience/elasticity. To this end we introduce a failure simulation framework and two resilience strategies (immediate mini-batch rollback and lossy forward recovery), which we study compared with checkpoint-restart approaches in a variety of settings in order to understand the trade-offs between the accuracy loss of the DNN model and the runtime overhead.

data-parallel training↗

Checkpoint and restart procedures for single and multi-stage structural model analysis in NASTRAN/COSMIC on a CDC 176

The Underwater Explosions Research Division (UERD) of the David Taylor Naval Ship Research and Development Center makes extensive use of NASTRAN/COSMIC on a CDC 176 to evaluate the structural response of ship structures subjected to underwater explosion shock loadings in the time domain. As relatively new users, UERD engineers have experienced difficulties with the checkpoint/restart feature because of the vague instructions in the user manual. Working procedures for the application of the checkpoint/restart feature to the transient analysis using NASTRAN/COSMIC are illustrated.

Camp, George H.↗

Report for the ASC CSSE L2 Milestone (4873) - Demonstration of Local Failure Local Recovery Resilient Programming Model.

Recovery from process loss during the execution of a distributed memory parallel application is presently achieved by restarting the program, typically from a checkpoint file. Future computer system trends indicate that the size of data to checkpoint, the lack of improvement in parallel file system performance and the increase in process failure rates will lead to situations where checkpoint restart becomes infeasible. In this report we describe and prototype the use of a new application level resilient computing model that manages persistent storage of local state for each process such that, if a process fails, recovery can be performed locally without requiring access to a global checkpoint file. LFLR provides application developers with an ability to recover locally and continue application execution when a process is lost. This report discusses what features are required from the hardware, OS and runtime layers, and what approaches application developers might use in the design of future codes, including a demonstration of LFLR-enabled MiniFE code from the Matenvo mini-application suite.

Heroux, Michael Allen↗

Data-driven Fault Tolerance for Work Stealing Computations

Checkpoint-restart approaches to fault tolerance typically roll back all the processes to the previous checkpoint in the event of a failure. Work stealing is a promising technique to dynamically tolerate variations in the execution environment, including faults, system noise, and energy constraints. In this paper, we present fault tolerance mechanisms for task parallel computations, a popular computation idiom, employing work stealing. The computation is organized as a collection of tasks with data in a global address space. The completion of data operations, rather than the actual messages, is tracked to derive an idempotent data store. This information is used to accurately identify the tasks to be re-executed, therefore to recompute only the lost data, in the presence of random work stealing. We consider three recovery schemes that present distinct trade-offs -- lazy recovery with potentially increased re-execution cost, immediate collective recovery with associated synchronization overheads, and noncollective recovery enabled by additional communication. We employ distributed work stealing to dynamically rebalance the tasks on the live processes and evaluate the three schemes using candidate application benchmarks. We demonstrate that the overheads (space and time) of the fault tolerance mechanism are low, the cost incurred due to failures are small, and the overheads decrease with per-process work at scale.

fault tolerance, task parallelism, load balancing,↗

Node failure resiliency for Uintah without checkpointing

The frequency of failures in upcoming exascale supercomputers may well be greater than at present due to many-core architectures if component failure rates remain unchanged. This potential increase in failure frequency coupled with I/O challenges at exascale may prove problematic for current resiliency approaches such as checkpoint restarting, although the use of fast intermediate memory may help. Algorithm-Based Fault Tolerance (ABFT) using Adaptive Mesh Refinement (AMR) is one resiliency approach used to address these challenges. For adaptive mesh codes, a coarse mesh version of the solu- tion may be used to restore the fine mesh solution. This paper addresses the implementation of the ABFT approach within the Uintah software framework: both at a software level within Uintah and in the data reconstruction method used for the recovery of lost data. This method has two problems: inaccuracies introduced during the reconstruction propagate forward in time, and the physical consistency of variables such as positivity or boundedness may be violated during interpolation. These challenges can be addressed by the combination of two techniques: 1. a fault-tolerant MPI implementation to recover from runtime node failures, and 2. high-order interpolation schemes to preserve the physical solution and reconstruct lost data. Here, the approach considered here uses a "Limited Essentially Non-Oscillatory" (LENO) scheme along with AMR to rebuild the lost data without checkpointing using Uintah. Experiments were carried out using a fault-tolerant MPI - ULFM to recover from runtime failure, and LENO to recover data on patches belonging to failed ranks, while the simulation was continued to the end. Results show that this ABFT approach is up to 10x faster than the traditional checkpointing method. The new interpolation approach is more accurate than linear interpolation and not subject to the overshoots found in other interpolation methods.

97 MATHEMATICS AND COMPUTING↗