Verifying Collective MPI Calls

Träff, Jesper Larsson; Worringen, Joachim

doi:10.1007/978-3-540-30218-6_11

Cited by 17 publications

(7 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Validation can also be done inside MPI libraries such as in an extension of MPICH, allowing collective verification for the full MPI-2 standard [3,10]. The detection of runtime deadlock causes is however limited to the information available to the MPI routines.…”

Section: Online Dynamic Toolsmentioning

confidence: 99%

Combining static and dynamic validation of MPI collective communications

Saillard

Carribault

Barthou

2013

Proceedings of the 20th European MPI Users' Group Meeting

View full text Add to dashboard Cite

Collective MPI communications have to be executed in the same order by all processes in their communicator and the same number of times, otherwise a deadlock occurs. As soon as the controlflow involving these collective operations becomes more complex, in particular including conditionals on process ranks, ensuring the correction of such code is error-prone. We propose in this paper a static analysis to detect when such situation occurs, combined with a code transformation that prevents from deadlocking. We show on several benchmarks the small impact on performance and the ease of integration of our techniques in the development process. General TermsVerification

show abstract

Section: Online Dynamic Toolsmentioning

confidence: 99%

Combining static and dynamic validation of MPI collective communications

Saillard

Carribault

Barthou

2013

Proceedings of the 20th European MPI Users' Group Meeting

View full text Add to dashboard Cite

show abstract

“…Parallel debuggers operate similarly as sequential debuggers [36,31,30,28,4], but can be difficult to use effectively when there are hundreds of processes. Automatic checking tools address the drawbacks of manual checking in parallel debuggers [38,12,19,7,37]. These tools use similar rules as static analysis, but they verify the rules at runtime rather than compile time.…”

Section: Related Workmentioning

confidence: 99%

MPIWiz

Xue

Liu

et al. 2009

Proceedings of the 14th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

View full text Add to dashboard Cite

Message Passing Interface (MPI) is a widely used standard for managing coarse-grained concurrency on distributed computers. Debugging parallel MPI applications, however, has always been a particularly challenging task due to their high degree of concurrent execution and non-deterministic behavior. Deterministic replay is a potentially powerful technique for addressing these challenges, with existing MPI replay tools adopting either data-replay or order-replay approaches. Unfortunately, each approach has its tradeoffs. Data-replay generates substantial log sizes by recording every communication message. Order-replay generates small logs, but requires all processes to be replayed together. We believe that these drawbacks are the primary reasons that inhibit the wide adoption of deterministic replay as the critical enabler of cyclic debugging of MPI applications.This paper describes subgroup reproducible replay (SRR), a hybrid deterministic replay method that provides the benefits of both data-replay and order-replay while balancing their trade-offs. SRR divides all processes into disjoint groups. It records the contents of messages crossing group boundaries as in data-replay, but records just message orderings for communication within a group as in order-replay. In this way, SRR can exploit the communication locality of traffic patterns in MPI applications. During replay, developers can then replay each group individually. SRR reduces recording overhead by not recording intra-group communication, and reduces replay overhead by limiting the size of each replay group. Exposing these tradeoffs gives the user the necessary control for making deterministic replay practical for MPI applications.We have implemented a prototype, MPIWIZ, to demonstrate and evaluate SRR. MPIWIZ employs a replay framework that allows transparent binary instrumentation of both library and system calls. As a result, MPIWIZ replays MPI applications with no source code modification and relinking, and handles nondeterminism in both MPI and OS system calls. Our preliminary results show that MPIWIZ can reduce recording overhead by over a factor of four relative to data-replay, yet without requiring the entire application to be replayed as in order-replay. Recording increases execution time by 27% while the application can be replayed in just 53% of its base execution time.

show abstract

“…To help application programmers, MPI implementations must provide better support for tools that help with debugging and verification. A similar tool is described in [31]. Auxiliary tools that help in debugging are also useful.…”

Section: Support For Debugging and Verificationmentioning

confidence: 99%

Advances in Computer Systems Architecture

2007

Lecture Notes in Computer Science

View full text Add to dashboard Cite

PrefaceOn behalf of the program and organizing committee members of this conference, we are pleased to present you with the proceedings of the 12 th Asia-Pacific Computer Systems Architecture Conference (ACSAC 2007), which was hosted in Seoul, Korea on August [23][24][25] 2007. This conference has traditionally been a forum for leading researchers in the Asian, American and Oceanian regions to share recent progress and the latest results in both architectural and system issues. In the past few years the conference has become more international in the sense that the geographic origin of participants has become broader to include researchers from all around the world, including Europe and the Middle East.This year, we received 92 paper submissions. Each submission was reviewed by at least three primary reviewers along with up to three secondary reviewers. The total number of completed reviews reached 333, giving each submission 3.6 reviews on average. All the reviews were carefully examined during the paper selection process, and finally 26 papers were accepted, resulting in an acceptance rate of about 28%. The selected papers encompass a wide range of topics, with much emphasis on hardware and software techniques for state-of-the-art multicore and multithreaded architectures. In addition to the regular papers, the technical program of the conference included eight invited papers from world-class renowned researchers and featured two keynotes by Pen-Chung Yew (University of Minnesota) and Kunio Uchiyama (Hitachi), addressing a compiler framework for speculative multithreading and power-efficient heterogeneous multicore chip development, respectively. We sincerely hope that the proceedings will serve as a valuable reference for researchers and developers alike.Putting together ACSAC 2007 was a team effort. First of all, we would like to express our special gratitude to the authors and speakers for providing the contents of the program. We would also like to thank the program committee members and external reviewers for diligently reviewing the papers and providing suggestions for their improvements. We believe that you will find the outcome of their efforts in this book. In addition, we extend our thanks to the organizing committee members and student volunteers, who contributed enormously to various aspects of conference administration. Finally, we would like to express special thanks to Chris Jesshope and Jinling Xue for sharing their experience and offering fruitful feedback in the early stages of preparing the conference.

show abstract

Verifying Collective MPI Calls

Cited by 17 publications

References 5 publications

Combining static and dynamic validation of MPI collective communications

Combining static and dynamic validation of MPI collective communications

MPIWiz

Advances in Computer Systems Architecture

Contact Info

Product

Resources

About