DMCA
Abstract PARALLEL PROGRAM DEBUGGING BASED ON DATA-REPLAY
| Citations: | 5 - 0 self |
BibTeX
@MISC{Maruyama_abstractparallel,
author = {Masao Maruyama and Tomoaki Tsumura and Hiroshi Nakashima},
title = {Abstract PARALLEL PROGRAM DEBUGGING BASED ON DATA-REPLAY},
year = {}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Nondeterministic nature of parallel programs is the major difficulty in debugging. Order-replay, a technique to solve this problem, is widely used because of its small overhead. It has, however, several serious drawbacks: all processes of the parallel program have to participate in replay even when some of them are clearly not involved with the bug; and the programmer cannot stop the process being debugged at an arbitrary point. We adopt another method for deterministic replay, Data-replay, which logs contents of the events rather than their order, and makes it possible to run and stop each process independently. Data-replay is well able to cooperate with reverse execution mechanisms. We applied the Data-replay mechanism to MPI based parallel programs. The result of our experiment with NAS Parallel Benchmarks shows that our mechanism works at a practical cost. Logging communicated data incurs only 24 % overhead while it accelerates replayed execution by 38 %, both in average. Key Words:parallel debugger, Data-replay, reverse execution, checkpointing 1
Keyphrases
parallel program reverse execution practical cost data incurs na parallel benchmark several serious drawback major difficulty nondeterministic nature data-replay mechanism reverse execution mechanism key word programmer cannot small overhead arbitrary point replayed execution deterministic replay
