Workshop on Parallel and Distributed Systems:
Testing, Analysis, and Debugging
(PADTAD - VI)
In conjunction with
International Symposium on Software Testing and Analysis
July 20-21 2008
The PADTAD 2008 workshop is a day and a half event at ISSTA 2008 focusing on techniques and systems that aid in the testing, analysis, and debugging of multi-threaded/parallel/distributed applications systems. The workshop has a practical and applied emphasis on systems that have been implemented in (at least) prototype form. This workshop concentrates on works whose main contributions are in the field of testing and debugging.
Although debuggers and profilers are the traditional examples of these tools on sequential machines, there are issues unique to Multi-threaded/Parallel/Distributed systems that are not commonly addressed. Examples of such significant challenges in Multi-threaded/Parallel/Distributed systems include deadlock, load imbalance, data sharing patterns, race conditions, and contention. In order to find certain classes of bugs in Multi-threaded/Parallel/Distributed applications, specific timing conditions are required. As a result, timing bugs are found very late in the testing process or by end users. Many established testing techniques and tools are insufficient for non-sequential programs.
Dr. Ramesh Peri (Intel) will give a keynote presentation on the work that Intel is doing to enable multi-core software development.
Support for students, especially paper authors, is being provided by our sponsors. For applications, please contact Shmuel Ur.
A special session will focus on teaching students and professionals how to develop Multi-threaded/Parallel/Distributed systems. In addition to dealing with the input space, developers of such systems also need to know how to deal with the huge space of possible schedules. Another inherent Difficulty is non-determinism. As a result, locally reviewing Multi-threaded/Parallel/Distributed code does not reveal all of its possible behaviors while overcoming non-determinism requires domain specific testing techniques such as replay. We are soliciting abstracts as well as regular papers on teaching how to design, test, and verify Multi-threaded/Parallel/Distributed systems. Discussion of overall approach, useful exercises and projects, experience reports, etc., are welcome.
Accepted papers, as well as education session abstracts, will be published in a CDROM proceedings and will be included in the ACM Digital Library (just like other ISSTA 2008 papers).
Topics of interest to the workshop include, but are not limited to:
- Techniques for multi-core processors
- Techniques for MPI and OpenMP or other library based applications
- Transactional memory
- Tools for testing or debugging of Multi-threaded/Parallel/Distributed applications
- Test generation algorithms for Multi-threaded/Parallel/Distributed applications
- Debugging advanced network interface technologies (e.g., Myrinet, VIA)
- Debugging and testing Multi-threaded/Parallel/Distributed applications
- Testing and Debugging of Multi-threaded/Parallel/Distributed applications written using domain-specific languages
- Using static analysis or formal verification to enhance debugging and testing of Multi-threaded/Parallel/Distributed applications
- Formal specification of concurrency libraries, and uses in compliance testing of implementations
- Detecting race conditions and deadlocks
- Replay of Multi-threaded/Parallel/Distributed applications
- Finding timing bugs early in the process
- Testing real-time Multi-threaded/Parallel/Distributed applications
- Fault injection of Multi-threaded/Parallel/Distributed applications
- Testing the fault tolerance of Multi-threaded/Parallel/Distributed applications
- Testing and debugging techniques for timing related bugs in hardware
- Pilots in applying new testing techiniques to Multi-threaded/Parallel/Distributed applications
- Multi-threaded/Parallel/Distributed review techniques and review tools
- Teaching of Multi-threaded/Parallel/Distributed system design, verification and testing