Parallelization of Bulk Operations for STL Dictionaries

Frias, Leonor; Singler, Johannes

doi:10.1007/978-3-540-78474-6_8

Cited by 14 publications

(26 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a first effort, we show that our adapted weight-balanced B-tree forms a viable baseline to act as a Pareto queue and in fact does outperform the implementation of Frias and Singler [6]. Following up on these results, we discuss the baseline for our comparison and present a thorough evaluation of paPaSearch for the bi-objective case.…”

Section: Experimental Evaluationmentioning

confidence: 89%

“…In [19] this result is generalized to accommodate the operations needed for paPaSearch. In principle, this approach can be extended to B-trees resulting in better cache-efficiency than [6,8,11] Here we go one step further by using weight-balanced B-trees [3]. This way we get easily predictable highly coarse grained re-balancing by partial rebuilding and we naturally integrate dynamic load balancing by work stealing.…”

Section: Parallel Search Treesmentioning

confidence: 96%

“…This way we get easily predictable highly coarse grained re-balancing by partial rebuilding and we naturally integrate dynamic load balancing by work stealing. This makes our solution more elegant and robust than [19] since [6,19] uses both static explicit load balancing for bulk updates and dynamic load balancing for finding Pareto optima.…”

Section: Parallel Search Treesmentioning

confidence: 97%

“…Their approach also suffers from a massive sequential bottleneck in the worst case, when all updates go to the same leaf. Frias and Singler show how to do parallel bulk insertion into binary search trees [6]. This algorithm is scalable by splitting the tree into pieces such that each piece receives an equal number of insertions.…”

Section: Parallel Search Treesmentioning

confidence: 98%

See 3 more Smart Citations

Parallel Bi-objective Shortest Paths Using Weight-Balanced B-trees with Bulk Updates

Erb

Kobitzsch

Sanders

2014

Experimental Algorithms

View full text Add to dashboard Cite

Abstract. We present a practical parallel algorithm for finding shortest paths in the presence of two objective functions. The algorithm builds on a recent theoretical result that on the first glance looks impractical. We address the problem of significant constant factor overheads due to numerous prefix sum computations by carefully re-engineering the algorithm for moderate parallelism. In addition, we develop a parallel weight-balanced B-tree data structure that cache efficiently supports bulk updates. This result might be of independent interest and closes the gap between the full-blown search tree data structure required by the theoretical result over the simple priority queue for the sequential algorithm. Comparing our implementation against a highly tuned sequential bi-objective search, we achieve speedups of 8 on 16 cores.

show abstract

Section: Experimental Evaluationmentioning

confidence: 89%

Section: Parallel Search Treesmentioning

confidence: 96%

Section: Parallel Search Treesmentioning

confidence: 97%

Section: Parallel Search Treesmentioning

confidence: 98%

See 2 more Smart Citations

Parallel Bi-objective Shortest Paths Using Weight-Balanced B-trees with Bulk Updates

Erb

Kobitzsch

Sanders

2014

Experimental Algorithms

View full text Add to dashboard Cite

show abstract

“…Examples that use work stealing include Cilk [48], Cilk++ [49], Java fork-join [50] (wich is used by Akka), X10 [51], Intel's Threading Building Blocks [52], NUMA-aware OpenMP schedulers [53], and parallelized implementations of the STL [54]. Further, Lifflander et al demonstrated that a hierarchical version of the algorithm scales to cluster deployments with up to 163,840 cores [55].…”

Section: Scheduling Algorithmmentioning

confidence: 99%

Revisiting actor programming in C++

Charousset

Hiesgen

Schmidt

2016

Computer Languages, Systems & Structures

View full text Add to dashboard Cite

The actor model of computation has gained significant popularity over the last decade. Its high level of abstraction makes it appealing for concurrent applications in parallel and distributed systems. However, designing a real-world actor framework that subsumes full scalability, strong reliability, and high resource efficiency requires many conceptual and algorithmic additives to the original model.In this paper, we report on designing and building CAF, the "C++ Actor Framework". CAF targets at providing a concurrent and distributed native environment for scaling up to very large, highperformance applications, and equally well down to small constrained systems. We present the key specifications and design concepts-in particular a message-transparent architecture, type-safe message interfaces, and pattern matching facilities-that make native actors a viable approach for many robust, elastic, and highly distributed developments. We demonstrate the feasibility of CAF in three scenarios: first for elastic, upscaling environments, second for including heterogeneous hardware like GPGPUs, and third for distributed runtime systems. Extensive performance evaluations indicate ideal runtime behaviour for up to 64 cores at very low memory footprint, or in the presence of GPUs. In these tests, CAF continuously outperforms the competing actor environments Erlang, Charm++, SalsaLite, Scala, ActorFoundry, and even the OpenMPI.

show abstract

Operators for Data Redistribution: Applications to the STL Library and RayTracing Algorithm

et al. 2021

View full text Add to dashboard Cite

In distributed-memory systems, data redistributions are operations that change the ownership and location of a selected subset of a data structure at runtime. They allow the improvement of the performance of parallel algorithms which operate on changing or partial domains, aiming to create a balanced workload among the active processes. To manually redistribute data is a cumbersome and error-prone task. In this paper, we present a method based on four combinable operators to redistribute partial domains selected by the programmer at runtime in an efficient and simple way. They abstract to the programmer the data-redistribution implementation details, such as the new mapping, relocation, and communication of the selected data. We also present the application of the proposed operators to a RayTracing application and to a significant part of STL (C++ Standard Template Library). Our experimental results show that our approach automatically generates a good load balance, which leads to performance improvements for generic data-distribution policies. It does not introduce significant performance overheads compared with tailored data redistributions directly programmed using MPI (Message Passing Interface), while it greatly reduces the code development effort.

show abstract

Parallelization of Bulk Operations for STL Dictionaries

Cited by 14 publications

References 7 publications

Parallel Bi-objective Shortest Paths Using Weight-Balanced B-trees with Bulk Updates

Parallel Bi-objective Shortest Paths Using Weight-Balanced B-trees with Bulk Updates

Revisiting actor programming in C++

Operators for Data Redistribution: Applications to the STL Library and RayTracing Algorithm

Contact Info

Product

Resources

About