Implementing Efficient Dynamic Formal Verification Methods for MPI Programs

Vakkalanka, Sarvani; DeLisi, Michael; Gopalakrishnan, Ganesh; Kirby, Robert M.; Thakur, Rajeev; Gropp, William

doi:10.1007/978-3-540-87475-1_34

Cited by 14 publications

(12 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [27], we wrote a description for many MPI functions in TLA+ [50]; this was expanded into a description of nearly 150 MPI functions in [28]. The ISP tool [33,32] bypasses the need to model MPI functions by relying on the native semantics of an actual MPI library implementation. This of course runs the risk of relying on the semantics of an actual MPI library implementation.…”

Section: Discussionmentioning

confidence: 99%

“…We have built a support library in Promela that models these MPI primitives. In Section 6, we discuss many related issues, including the growing success of our direct dynamic verification approach for MPI programs [32,33,34,35].…”

Section: Formal Modeling and Verificationmentioning

confidence: 99%

“…Techniques to enhance the efficiency of this algorithm were reported in [31]. In recent work, we report on an approach to model check MPI programs directly using dynamic verification methods [32,33,34,35], introducing our algorithm 'ISP' in the process. We also employed our ISP algorithm for detecting the presence of functionally irrelevant barriers in MPI programs [36].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Formal methods applied to high‐performance computing software design: a case study of MPI one‐sided communication‐based locking

et al. 2009

Self Cite

View full text Add to dashboard Cite

SUMMARYThere is growing need to address the complexity of verifying the numerous concurrent protocols employed in high performance computing software. Today's approaches for verification consist of testing detailed implementations of these protocols. Unfortunately, this approach can seldom show the absence of bugs, and often results in serious bugs escaping into the deployed software. An approach called Model Checking has been demonstrated to be eminently helpful in debugging these protocols early in the software life cycle by offering the ability to represent and exhaustively analyze simplified formal protocol models. The effectiveness of model checking has yet to be adequately demonstrated in high performance computing. This paper presents a case study of a concurrent protocol that was thought to be sufficiently well tested, but proved to contain two very non-obvious deadlocks in them. These bugs were automatically detected through model checking. The protocol models in which these bugs were detected were also easy to create. Recent work in our group demonstrates that even this tedium of model creation can be eliminated by employing dynamic source-code level analysis methods. Our case study comes from the important domain of Message Passing Interface (MPI) based programming which is universally employed for simulating and predicting anything from the structural integrity of combustion chambers to the path of hurricanes. We argue that model checking must be taught as well as used widely within HPC, given this and similar success stories.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Formal Modeling and Verificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Formal methods applied to high‐performance computing software design: a case study of MPI one‐sided communication‐based locking

et al. 2009

Self Cite

View full text Add to dashboard Cite

show abstract

“…These three goals are even more important to meet for concurrent program debuggers, because concurrent program executions are often far less intuitive than sequential program executions. This paper is on a family of new graphical user interfaces (GUI) that we have equipped our previously reported In-Situ Partial order (ISP, [1][2][3][4]) tool with, as our first attempt to meet the above goals. Information is presented by the ISP tool through a portable Java based GUI and an optional Microsoft Visual Studio GUI, and an Eclipse GUI (development in progress) that will integrate and extend these views.…”

Section: Introductionmentioning

confidence: 99%

How Formal Dynamic Verification Tools Facilitate Novel Concurrency Visualizations

Aananthakrishnan

DeLisi

Vakkalanka

et al. 2009

Recent Advances in Parallel Virtual Machine and Message Passing Interface

Self Cite

View full text Add to dashboard Cite

Abstract. With the exploding scale of concurrency, presenting valuable pieces of information collected by formal methods tools intuitively and graphically can greatly enhance concurrent system debugging. Traditional MPI program debuggers present trace views of MPI program executions. Such views are redundant, often containing equivalent traces that permute independent MPI calls. In our ISP formal dynamic verifier for MPI programs, we present a collection of alternate views made possible by the use of formal dynamic verification. Some of ISP's views help pinpoint errors, some facilitate discerning errors by eliminating redundancy, while others help understand the program better by displaying concurrent even orderings that must be respected by all MPI implementations, in the form of completes-before graphs. In this paper, we describe ISP's GUI capabilities in all these areas which are currently supported by a portable Java based GUI, a Microsoft Visual Studio GUI, and an Eclipse based GUI whose development is in progress.

show abstract

“…• Since Fib is implemented as an extension to the dynamic formal verification methodology employed in our tool ISP( [9,13,4]), it is capable of detecting deadlocks, and then aborting its analysis. Here are some example deadlock scenarios that ISP can detect: (i) deadlocks due to a collective barrier being incorrectly placed, (ii) those introduced when the user forgets to issue the (supposed) collective call from within some of the processes, (iii) the user employing the wrong communicator for one of the barrier calls, or (iv) MPI messages not matching.…”

Section: Introductionmentioning

confidence: 99%

A Formal Approach to Detect Functionally Irrelevant Barriers in MPI Programs

Sharma

Vakkalanka

Gopalakrishnan

et al. 2008

Recent Advances in Parallel Virtual Machine and Message Passing Interface

Self Cite

View full text Add to dashboard Cite

Abstract. We examine the unsolved problem of automatically and efficiently detecting functionally irrelevant barriers in MPI programs. A functionally irrelevant barrier is a set of MPI_Barrier calls, one per MPI process, such that their removal does not alter the overall MPI communication structure of the program. Static analysis methods are incapable of solving this problem, as MPI programs can compute many quantities at runtime, including send targets, receive sources, tags, and communicators, and also can have data-dependent control flows. We offer an algorithm called Fib to solve this problem based on dynamic (runtime) analysis. Fib applies to MPI programs that employ 24 widely used twosided MPI operations. We show that it is sufficient to detect barrier calls whose removal causes a wildcard receive statement placed before or after a barrier to now begin matching a send statement with which it did not match before. Fib determines whether a barrier becomes relevant in any interleaving of the MPI processes of a given MPI program. Since the number of interleavings can grow exponentially with the number of processes, Fib employs a sound method to drastically reduce this number, by computing only the relevant interleavings. We show that many MPI programs do not have data dependent control flows, thus making the results of Fib applicable to all the input data the program can accept.

show abstract

Implementing Efficient Dynamic Formal Verification Methods for MPI Programs

Cited by 14 publications

References 9 publications

Formal methods applied to high‐performance computing software design: a case study of MPI one‐sided communication‐based locking

Formal methods applied to high‐performance computing software design: a case study of MPI one‐sided communication‐based locking

How Formal Dynamic Verification Tools Facilitate Novel Concurrency Visualizations

A Formal Approach to Detect Functionally Irrelevant Barriers in MPI Programs

Contact Info

Product

Resources

About