Precise Predictive Analysis for Discovering Communication Deadlocks in MPI Programs

Abstract

The Message Passing Interface (MPI) is the standard API for parallelization in high-performance and scientific computing. Communication deadlocks are a frequent problem in MPI programs, and this article addresses the problem of discovering such deadlocks. We begin by showing that if an MPI program is single path, the problem of discovering communication deadlocks is NP-complete. We then present a novel propositional encoding scheme that captures the existence of communication deadlocks. The encoding is based on modeling executions with partial orders and implemented in a tool called MOPPER . The tool executes an MPI program, collects the trace, builds a formula from the trace using the propositional encoding scheme, and checks its satisfiability. Finally, we present experimental results that quantify the benefit of the approach in comparison to other analyzers and demonstrate that it offers a scalable solution for single-path programs.

Two-Phase Dynamic Analysis of Message-Passing Go Programs Based on Vector Clocks

“…In the MPI setting, the works [10,17] use model checking ideas to predict devious schedules based on a single trace. As mentioned above, some variant of MPI's receive operation differs from Go.…”

Combining symbolic execution and model checking to verify MPI programs

“…We have already implemented the refinement in MPI-SV. The support of masterslave pattern demonstrates that MPI-SV outperforms the single path reasoning work [23,41].…”

“…Since we compute SMO(op, S) by statically matching the arguments of the recorded operations, SMO(op, S) may contain some false matchings. Calculating the precisely matched operations of op is NP-complete [23], and we suppose such an ideal method exists. We use CSP st at ic and CSP ideal to denote the generated models using SMO(op, S) and the ideal method, respectively.…”

“…Baselines. We use pure symbolic execution as the first baseline because: (1) none of the state-of-the-art symbolic execution based verification tools can analyze non-blocking MPI programs, e.g., CIVL [57,75]; (2) MPI-SPIN [74] can support input coverage and non-blocking operations, but it requires building models of the programs manually; and (3) other automated tools that support non-blocking operations, such as MOPPER [23] and ISP [83], can only verify programs under given inputs. MPI-SV aims at covering both the input space and non-determinism automatically.…”

“…MPI-SV supports both input and schedule coverages, and a larger scope of verifiable properties. MOPPER [23] encodes the deadlock detection problem under concrete inputs in a SAT equation. Similarly, Huang and Mercer [41] use an SMT formula to reason about a trace of an MPI program for deadlock detection.…”