We describe a non-blocking concurrent hash trie based on shared-memory single-word compare-and-swap instructions. The hash trie supports standard mutable lock-free operations such as insertion, removal, lookup and their conditional variants. To ensure space-efficiency, removal operations compress the trie when necessary. We show how to implement an efficient lock-free snapshot operation for concurrent hash tries. The snapshot operation uses a single-word compare-and-swap and avoids copying the data structure eagerly. Snapshots are used to implement consistent iterators and a linearizable size retrieval. We compare concurrent hash trie performance with other concurrent data structures and evaluate the performance of the snapshot operation.

No abstract

As the number of cores increases in modern multiprocessors, it is becoming increasingly dicult to write general purpose applications that eciently utilize this computing power. This remains an open research problem.Most applications manipulate structured data. Modern languages and platforms provide collection frameworks with basic data structures like lists, hashtables and trees. These data structures have a range of predened operations which include mapping, ltering or nding elements. Such bulk operations traverse the collection and process the elements sequentially. Their implementation relies on iterators, which are not applicable to parallel operations due to their sequential nature.We present an approach to parallelizing collection operations in a generic way, used to factor out common parallel operations in collection libraries. Our framework is easy to use and straightforward to extend to new collections. We show how to implement concrete parallel collections such as parallel arrays and parallel hash maps, proposing an ecient solution to parallel hash map construction. Finally, we give benchmarks showing the performance of parallel collection operations.

Programmers who need high performance currently rely on low-level, architecture-specific programming models (e.g. OpenMP for CMPs, CUDA for GPUs, MPI for clusters). Performance optimization with these frameworks usually requires expertise in the specific programming model and a deep understanding of the target architecture. Domain-specific languages (DSLs) are a promising alternative, allowing compilers to map problem-specific abstractions directly to low-level architecture-specific programming models. However, developing DSLs is difficult, and using multiple DSLs together in a single application is even harder because existing compiled solutions do not compose together. In this paper, we present four new performance-oriented DSLs developed with Delite, an extensible DSL compilation framework. We demonstrate new techniques to compose compiled DSLs embedded in a common backend together in a single program and show that generic optimizations can be applied across the different DSL sections. Our new DSLs are implemented with a small number of reusable components (less than 9 parallel operators total) and still achieve performance up to 125x better than library implementations and at worst within 30% of optimized stand-alone DSLs. The DSLs retain good performance when composed together, and applying cross-DSL optimizations results in up to an additional 1.82x improvement.

Abstract. Implementing correct and deterministic parallel programs is challenging. Even though concurrency constructs exist in popular programming languages to facilitate the task of deterministic parallel programming, they are often too low level, or do not compose well due to underlying blocking mechanisms. In this paper, we present the design and implementation of a fundamental data structure for composable deterministic parallel dataflow computation through the use of functional programming abstractions. Additionally, we provide a correctness proof, showing that the implementation is linearizable, lock-free, and deterministic. Finally, we show experimental results which compare our FlowPool against corresponding operations on other concurrent data structures, and show that in addition to offering new capabilities, FlowPools reduce insertion time by 49 − 54% on a 4-core i7 machine with respect to comparable concurrent queue data structures in the Java standard library.