Brian Demsky

Automatic detection and repair of errors in data structures

We present a system that accepts a specification of key data structure consistency constraints, then dynamically detects and repairs violations of these constraints, enabling the program to continue to execute productively even in the face of otherwise crippling errors. Our experience using our system indicates that the specifications are relatively easy to develop once one understands the data structures. Furthermore, for our set of benchmark applications, our system can effectively repair inconsistent data structures and enable the program to continue to operate successfully.

Data structure repair using goal-directed reasoning

Data structure repair is a promising technique for enabling programs to execute successfully in the presence of otherwise fatal data structure corruption errors. Previous research in this field relied on the developer to write a specification to explicitly translate model repairs into concrete data structure repairs, raising the possibility of 1) incorrect translations causing the supposedly repaired concrete data structures to be inconsistent, and 2) repaired models with no corresponding concrete data structure representation. We present a new repair algorithm that uses goal-directed reasoning to automatically translate model repairs into concrete data structure repairs. This new repair algorithm eliminates the possibility of incorrect translations and repaired models with no corresponding representation as concrete data structures.

Role-based exploration of object-oriented programs

We present a new technique for helping developers understand heap properties of object-oriented programs and how the actions of the program affect these properties. Our dynamic analysis uses the aliasing properties of objects to synthesize a set of roles; each role represents an abstract object state intended to be of interest to the developer. We allow the developer to customize the analysis to explore the object states and behavior of the program at multiple different and potentially complementary levels of abstraction.The analysis uses roles as the basis for three abstractions: role transition diagrams, which present the observed transitions between roles and the methods responsible for the transitions; role relationship diagrams, which present the observed referencing relationships between objects playing different roles; and enhanced method interfaces, which present the observed roles of method parameters.Together, these abstractions provide useful information about important object and data structure properties and how the actions of the program affect these properties. We have used our implemented role analysis to explore the behavior of several Java programs. Our experience indicates that, when combined with a powerful graphical user interface, roles are a useful abstraction for helping developers explore and understand the behavior of object-oriented programs.

Goal-Directed Reasoning for Specification-Based Data Structure Repair

Software errors and hardware failures can cause data structures in running programs to violate key data structure consistency properties. As a result of this violation, the program may produce unacceptable results or even fail. We present a new data structure repair system. This system accepts a specification of data structure consistency properties stated in terms of an abstract set-and relation-based model of the data structures in the running program. It then automatically generates a repair algorithm that, during the execution of the program, detects and repairs any violations of these constraints. The goal is to enable the program to continue to execute acceptably in the face of otherwise crippling data structure corruption errors. We have applied our system to repair inconsistent data structures in five applications: CTAS (an air traffic control system), AbiWord (an open source word processing program), Freeciv (an interactive multiplayer game), a parallel x86 emulator, and a simplified Linux file system. Our results indicate that the generated repair algorithms can effectively repair inconsistent data structures in these applications to enable the applications to continue to operate successfully in cases where the original application would have failed. Without repair, all of the applications fail

CDSchecker: checking concurrent data structures written with C/C++ atomics

Writing low-level concurrent software has traditionally required intimate knowledge of the entire toolchain and often has involved coding in assembly. New language standards have extended C and C++ with support for low-level atomic operations and a weak memory model, enabling developers to write portable and efficient multithreaded code. Developing correct low-level concurrent code is well-known to be especially difficult under a weak memory model, where code behavior can be surprising. Building reliable concurrent software using C/C++ low-level atomic operations will likely require tools that help developers discover unexpected program behaviors. In this paper we present CDSChecker, a tool for exhaustively exploring the behaviors of concurrent code under the C/C++ memory model. We develop several novel techniques for modeling the relaxed behaviors allowed by the memory model and for minimizing the number of execution behaviors that CDSChecker must explore. We have used CDSChecker to exhaustively unit test several concurrent data structure implementations on specific inputs and have discovered errors in both a recently published C11 implementation of a work-stealing queue and a single producer, single consumer queue implementation.

OoOJava: software out-of-order execution

Developing parallel software using current tools can be challenging. Even experts find it difficult to reason about the use of locks and often accidentally introduce race conditions and deadlocks into parallel software. OoOJava is a compiler-assisted approach that leverages developer annotations along with static analysis to provide an easy-to-use deterministic parallel programming model. OoOJava extends Java with a task annotation that instructs the compiler to consider a code block for out-of-order execution. OoOJava executes tasks as soon as their data dependences are resolved and guarantees that the execution of an annotated program preserves the exact semantics of the original sequential program. We have implemented OoOJava and achieved an average speedup of 16.6x on our ten benchmarks.

SATCheck: SAT-directed stateless model checking for SC and TSO

Writing low-level concurrent code is well known to be challenging and error prone. The widespread deployment of multi-core hardware and the shift towards using low-level concurrent data structures has moved the problem into the mainstream. Finding bugs in such code may require finding a specific bug-revealing thread interleaving out of a huge space of parallel executions. Model-checking is a powerful technique for exhaustively testing code. However, scaling model checking presents a significant challenge. In this paper we present a new and more scalable technique for model checking concurrent code, based on concrete execution. Our technique observes concrete behaviors, builds a model of these behaviors, encodes the model in SAT, and leverages SAT solver technology to find executions that reveal new behaviors. It then runs the new execution, incorporates the newly observed behavior, and repeats the process until it has explored all reachable behaviors. We have implemented a prototype of our approach in the SATCheck tool. Our tool supports both the Total Store Ordering (TSO) and Sequentially Consistent (SC) memory models. We evaulate SATCheck by testing several concurrent data structure implementations and comparing its performance to the original DPOR stateless model checking algorithm implemented in CDSChecker, the source DPOR algorithm implemented in Nidhugg, and CheckFence. Our experiments show that SATCheck scales better than previous approaches while at the same time operating on concrete executions.

A model counter for constraints over unbounded strings

Model counting is the problem of determining the number of solutions that satisfy a given set of constraints. Model counting has numerous applications in the quantitative analyses of program execution time, information flow, combinatorial circuit designs as well as probabilistic reasoning. We present a new approach to model counting for structured data types, specifically strings in this work. The key ingredient is a new technique that leverages generating functions as a basic primitive for combinatorial counting. Our tool SMC which embodies this approach can model count for constraints specified in an expressive string language efficiently and precisely, thereby outperforming previous finite-size analysis tools. SMC is expressive enough to model constraints arising in real-world JavaScript applications and UNIX C utilities. We demonstrate the practical feasibility of performing quantitative analyses arising in security applications, such as determining the comparative strengths of password strength meters and determining the information leakage via side channels.

Interruptible tasks: treating memory pressure as interrupts for highly scalable data-parallel programs

Real-world data-parallel programs commonly suffer from great memory pressure, especially when they are executed to process large datasets. Memory problems lead to excessive GC effort and out-of-memory errors, significantly hurting system performance and scalability. This paper proposes a systematic approach that can help data-parallel tasks survive memory pressure, improving their performance and scalability without needing any manual effort to tune system parameters. Our approach advocates interruptible task (ITask), a new type of data-parallel tasks that can be interrupted upon memory pressure---with part or all of their used memory reclaimed---and resumed when the pressure goes away. To support ITasks, we propose a novel programming model and a runtime system, and have instantiated them on two state-of-the-art platforms Hadoop and Hyracks. A thorough evaluation demonstrates the effectiveness of ITask: it has helped real-world Hadoop programs survive 13 out-of-memory problems reported on StackOverflow; a second set of experiments with 5 already well-tuned programs in Hyracks on datasets of different sizes shows that the ITask-based versions are 1.5--3x faster and scale to 3--24x larger datasets than their regular counterparts.

Bamboo: a data-centric, object-oriented approach to many-core software

Traditional data-oriented programming languages such as dataflow languages and stream languages provide a natural abstraction for parallel programming. In these languages, a developer focuses on the flow of data through the computation and these systems free the developer from the complexities of low-level, thread-oriented concurrency primitives. This simplification comes at a cost --- traditional data-oriented approaches restrict the mutation of state and, in practice, the types of data structures a program can effectively use. Bamboo borrows from work in typestate and software transactions to relax the traditional restrictions of data-oriented programming models to support mutation of arbitrary data structures. We have implemented a compiler for Bamboo which generates code for the TILEPro64 many-core processor. We have evaluated this implementation on six benchmarks: Tracking, a feature tracking algorithm from computer vision; KMeans, a K-means clustering algorithm; MonteCarlo, a Monte Carlo simulation; FilterBank, a multi-channel filter bank; Fractal, a Mandelbrot set computation; and Series, a Fourier series computation. We found that our compiler generated implementations that obtained speedups ranging from 26.2x to 61.6x when executed on 62 cores.