Observationally cooperative multithreading

Abstract

Despite widespread interest in multicore computing, concurrency models in mainstream languages often lead to subtle, error-prone code.Observationally Cooperative Multithreading (OCM) is a new approach to shared-memory parallelism. Programmers write code using the well-understood cooperative (i.e., nonpreemptive) multithreading model for uniprocessors. OCM then allows threads to run in parallel, so long as results remain consistent with the cooperative model.Programmers benefit because they can reason largely sequentially. Remaining interthread interactions are far less chaotic than in other models, permitting easier reasoning and debugging. Programmers can also defer the choice of concurrency-control mechanism (e.g., locks or transactions) until after they have written their programs, at which point they can compare concurrency-control strategies and choose the one that offers the best performance. Implementers and researchers also benefit from the agnostic nature of OCMit provides a level of abstraction to investigate, compare, and combine a variety of interesting concurrency-control techniques.

Panini: a concurrent programming model for solving pervasive and oblivious interference

“…Cooperability [3,5,9] and automatic mutual exclusion [10][11][12] are the inverse of atomic blocks, i.e. the code is atomic and interference-free unless explicitly specified.…”

“…Similarly, cooperative reasoning [3,5] and observationally cooperative multithreading [9] limit interleaving points to yield expressions. Task Types [38] enforce pervasive atomicity, i.e.…”

Toward a Concurrent Programming Model with Modular Reasoning

“…Interference control framework enables modular reasoning by making interference points and interference behaviors known syntactically, statically and modularly. Previous work either controls interference points [2,6,7,84,90,94,151,154,157,172,176,177,177] or interference behaviors [56,68,126,129] but not both. Actors [6,7,172], active objects [94,151], atomicity [58,66,171], cooperability [157,176,177] and automatic mutual exclusion [2,84,154] control interference points but not interference behaviors.…”

“…Cooperability [157,176,177] and automatic mutual exclusion [2,84,154] are the inverse of atomic blocks, i.e. the code is atomic and interference-free unless explicitly specified.…”

“…Automatic mutual exclusion [2,84,154] inverts the model of atomic blocks in transactional memory such that code is run atomically unless explicitly specified using yield expressions. Similarly, cooperative reasoning [176,177] and observationally cooperative multithreading [157] is annotated by an interference-free assertion which must hold locally in the presence of concurrent interfering tasks. These techniques extend Hoare logic and address the oblivious interference problem via environment assumptions, however, they still assume pervasive interference between each two instructions.…”