Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures最新文献

Commitment and Slack for Online Load Maximization 在线负荷最大化的承诺和松弛

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400271

Samin Jamalabadi, Chris Schwiegelshohn, U. Schwiegelshohn

引用次数: 4

Functional Faults 功能性故障

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400261

Gali Sheffi, E. Petrank

{"title":"Functional Faults","authors":"Gali Sheffi, E. Petrank","doi":"10.1145/3350755.3400261","DOIUrl":"https://doi.org/10.1145/3350755.3400261","url":null,"abstract":"Hardware and software faults increasingly surface in today's computing environment and vast theoretical and practical research efforts are devoted to ameliorate the effects of malfunctionality in the computing process. Most research to date, however, has focused on how to discover and handle faulty data. In this paper we formalize and study faulty functionality in a modern multicore shared-memory environment. Functional faults have been previously studied in the architecture community. However, they have never been formally defined and lower/upper bounds were not previously proven. We present a model of functional faults, and study avenues that allow tolerating functional faults while maintaining the correctness of the entire computation. We exemplify this model by constructing a robust consensus protocol from functionally-faulty compare-and-swap objects. We then show a (tight) impossibility result for the same construction, when the number of faults exceeds a certain threshold. Interestingly, for some fault types, more functional faults can be tolerated than the analogue data faults, beating an impossibility result for data faults and demonstrating the difference between the two models.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127145903","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Provable Neuromorphic Advantages for Computing Shortest Paths 计算最短路径的可证明神经形态优势

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400258

J. Aimone, Yang Ho, Ojas D. Parekh, C. Phillips, A. Pınar, William M. Severa, Yipu Wang

{"title":"Provable Neuromorphic Advantages for Computing Shortest Paths","authors":"J. Aimone, Yang Ho, Ojas D. Parekh, C. Phillips, A. Pınar, William M. Severa, Yipu Wang","doi":"10.1145/3350755.3400258","DOIUrl":"https://doi.org/10.1145/3350755.3400258","url":null,"abstract":"Neuromorphic computing offers the potential of an unprecedented level of parallelism at a local scale. Although in their infancy, current first-generation neuromorphic processing units (NPUs) deliver as many as 128K artificial neurons in a package smaller than current laptop CPUs and demanding significantly less energy. Neuromorphic systems consisting of such NPUs and offering a total of 100 million neurons are anticipated in 2020. NPUs were envisioned to accelerate machine learning, and designing neuromorphic algorithms to leverage the benefits of NPUs in other domains remains an open challenge. We design and analyze neuromorphic graph algorithms, focusing on shortest path problems. Our neuromorphic algorithms are packet-passing algorithms relying on data movement for computation, and we develop data-movement lower bounds for conventional algorithms. A fair and rigorous comparison with conventional algorithms and architectures is paramount, and we prove a polynomial-factor advantage even when we assume an NPU with a simple grid-like network of neurons. To the best of our knowledge, this is one of the first examples of a provable asymptotic computational advantage for neuromorphic computing.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"219 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127196818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 10

Constant-Length Labelling Schemes for Faster Deterministic Radio Broadcast 更快确定性无线电广播的定长标记方案

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400238

Faith Ellen, Seth Gilbert

{"title":"Constant-Length Labelling Schemes for Faster Deterministic Radio Broadcast","authors":"Faith Ellen, Seth Gilbert","doi":"10.1145/3350755.3400238","DOIUrl":"https://doi.org/10.1145/3350755.3400238","url":null,"abstract":"In this paper, we consider the problem of broadcast from a specified source node in a known synchronous radio network. In 2019, Ellen, Gorain, Miller and Plc showed that this is possible if each node in the network only stores 2 (carefully chosen) bits of information. They proved that in an n-node network, their algorithm ensures that the broadcast completes within 2n-3 rounds. We show that storing only a small constant number of additional bits, it is possible to broadcast significantly faster when the source eccentricity, D, of the network is o(n). We begin by presenting a modification of Ellen, Gorain, Miller and Pelc's algorithm that stores 4 bits per node and completes within O(√Dn) rounds. Then we define a class of broadcast algorithms that includes both these algorithms and prove that any algorithm in this class requires Ω(√nD) rounds. Next, we show (non-constructively) that there exists a broadcast algorithm which stores only 3 bits of information per node, but completes within O(D log n+log2 n) rounds. Finally, using ideas from Chlamtac and Weinstein (1991), we show how to construct the relevant information and do almost as well, giving an algorithm using 3 bits per node that runs in O(D log^2 n) rounds.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"12 4","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114003398","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 7

Predicate Detection to Solve Combinatorial Optimization Problems 谓词检测解决组合优化问题

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400235

V. Garg

{"title":"Predicate Detection to Solve Combinatorial Optimization Problems","authors":"V. Garg","doi":"10.1145/3350755.3400235","DOIUrl":"https://doi.org/10.1145/3350755.3400235","url":null,"abstract":"We present a method to design parallel algorithms for constrained combinatorial optimization problems. Our method solves and generalizes many classical combinatorial optimization problems including the stable marriage problem, the shortest path problem and the market clearing price problem. These three problems are solved in the literature using Gale-Shapley algorithm, Dijkstra's algorithm, and Demange, Gale, Sotomayor algorithm. Our method solves all these problems by casting them as searching for an element that satisfies an appropriate predicate in a distributive lattice. Moreover, it solves generalizations of all these problems --- namely finding the optimal solution satisfying additional constraints called lattice-linear predicates. For stable marriage problems, an example of such a constraint is that Peter's regret is less than that of Paul. For shortest path problems, an example of such a constraint is that cost of reaching vertex v1 is at least the cost of reaching vertex v2. For the market clearing price problem, an example of such a constraint is that item1 is priced at least as much as item2. Our algorithm, called Lattice-Linear Predicate Detection (LLP) can be implemented in parallel without any locks or compare-and-set instructions. It just assumes atomicity of reads and writes.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"96 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133171233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 24

Fast Byzantine Agreement for Permissioned Distributed Ledgers 许可分布式账本的快速拜占庭协议

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400219

Thomas Locher

{"title":"Fast Byzantine Agreement for Permissioned Distributed Ledgers","authors":"Thomas Locher","doi":"10.1145/3350755.3400219","DOIUrl":"https://doi.org/10.1145/3350755.3400219","url":null,"abstract":"A consensus algorithm lies at the core of every distributed ledger as it defines how the constituent parts of the distributed system ensure that they share identical ledger copies. For a broad range of promising applications of distributed ledger technology, the parties responsible for the application can specify and control the distributed entities maintaining the ledger. Since the faulty behavior of any such entity must not cause inconsistencies in the ledger and access to the ledger is restricted, Byzantine agreement is a viable candidate as a backbone for such so-called permissioned distributed ledgers. In this setting, a primary objective is to maximize throughput, i.e., the rate at which transactions can be processed, which requires a quick settlement at the consensus layer. To this end, a synchronous, leader-based Byzantine agreement protocol is presented that is simple, easy to implement, and particularly suitable for permissioned distributed ledgers. Subject to the constraint that less than a quarter of all nodes are Byzantine, consensus is reached in 2 rounds of communication if the leader behaves correctly, whereas 3 rounds are required in expectation over multiple executions, which sets a new record for the lowest expected number of rounds for any randomized Byzantine agreement protocol. Moreover, the protocol is malleable in the sense that it can easily be adapted to meet additional requirements or constraints. For example, slight modifications make it possible to achieve the same guarantees for the partially synchronous communication model. Finally, evaluation results of a prototype implementation, focusing on throughput and scalability, show the algorithm's potential for use in practice.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116170229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 5

Time- and Space-Optimal Discrete Clock Synchronization in the Beeping Model 蜂鸣声模型中时间和空间最优离散时钟同步

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400246

Michel Feldmann, Ardalan Khazraei, C. Scheideler

{"title":"Time- and Space-Optimal Discrete Clock Synchronization in the Beeping Model","authors":"Michel Feldmann, Ardalan Khazraei, C. Scheideler","doi":"10.1145/3350755.3400246","DOIUrl":"https://doi.org/10.1145/3350755.3400246","url":null,"abstract":"We consider the clock synchronization problem in the (discrete) beeping model: Given a network of n nodes with each node having a clock value δ(v) ∈ {0,... T-1}, the goal is to synchronize the clock values of all nodes such that they have the same value in any round. As is standard in clock synchronization, we assume arbitrary activations for all nodes, i.e., the nodes start their protocol at an arbitrary round (not limited to {0,...,T-1}). We give an asymptotically optimal algorithm that runs in 4D + ⌊ D/⌊ T/4 ⌋ ⌋ · (T mod 4) = O(D) rounds, where D is the diameter of the network. Once all nodes are in sync, they beep at the same round every T rounds. The algorithm drastically improves on the O(T D)-bound of [2] (where T is required to be at least 4n, so the bound is no better than O(nD)). Our algorithm is very simple as nodes only have to maintain 3 bits in addition to the ⌈ log T ⌉ bits needed to maintain the clock. Furthermore we investigate the complexity of self-stabilizing solutions for the clock synchronization problem: We first show lower bounds of Ω(max{T,n}) rounds on the runtime and Ω(log(max{T,n})) bits of memory required for any such protocol. Afterwards we present a protocol that runs in O(max{T,n}) rounds using at most O(log(max{T,n})) bits at each node, which is asymptotically optimal with regards to both, runtime and memory requirements.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126546715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 5

A Discrete and Continuous Study of the Max-Chain-Formation Problem: Slow Down to Speed up 最大链形成问题的离散与连续研究:从慢到快

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400263

Jannik Castenow, Peter Kling, Till Knollmann, F. Heide

{"title":"A Discrete and Continuous Study of the Max-Chain-Formation Problem: Slow Down to Speed up","authors":"Jannik Castenow, Peter Kling, Till Knollmann, F. Heide","doi":"10.1145/3350755.3400263","DOIUrl":"https://doi.org/10.1145/3350755.3400263","url":null,"abstract":"Robot coordination problems deal with systems consisting of many autonomous, but simple, mobile agents that try to achieve a common, complex task. The agents' capabilities depend on the exact model and task but are typically very restricted. For example, there is usually no common coordinate system or sense of direction, and agents often have limited sensing capabilities. Among the most basic and well-studied type of tasks are GATHERING problems, in which initially scattered agents must gather at a single point. CHAINFORMATION problems represent another important formation primitive. Here, agents take the role of communication relays that, initially, form a winding chain connecting two base stations. The relays are to move such that the chain becomes straight, allowing for a more energy-efficient communication along the relay chain. Both GATHERING and CHAINFORMATION problems can be described as contracting: starting from an initially scattered formation, they seek to reach a smaller, more efficient structure. A natural complement are extension problems, where agents start in an initially dense formation and seek to reach an extended formation that covers as much area as possible. While there are some results about extension problems if agents move on grids or rings, results in standard discrete and continuous models for the Euclidean plane are scarce. Our work introduces the MAXFORM problem on the Euclidean plane and provides first analytical results for both the discrete and continuous case.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121676775","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 3

Efficient Concurrent Range Queries in B+-trees using RCU-HTM 使用RCU-HTM在B+-树中高效并发范围查询

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400237

Dimitrios Siakavaras, Panagiotis Billis, K. Nikas, G. Goumas, N. Koziris

引用次数: 2

Balanced Partitioning of Several Cache-Oblivious Algorithms 几种缓参无关算法的均衡分区

Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures Pub Date : 2020-07-06 DOI: 10.1145/3350755.3400214

Yuan Tang

{"title":"Balanced Partitioning of Several Cache-Oblivious Algorithms","authors":"Yuan Tang","doi":"10.1145/3350755.3400214","DOIUrl":"https://doi.org/10.1145/3350755.3400214","url":null,"abstract":"Frigo et al. proposed an ideal cache model and a recursive cache-oblivious technique to design sequential cache-efficient algorithms in an oblivious fashion. Ballard et al. pointed out that it is a fundamental open problem to extend the technique to an arbitrary architecture. Ballard et al. raised another open question on how to parallelize Strassen's algorithm exactly and efficiently on an arbitrary number of processors. We propose a novel way of partitioning a cache-oblivious algorithm to achieve perfect strong scaling on an arbitrary number, even a prime number, of processors within a certain range in a shared-memory setting. Our approach is Processor-Aware but Cache-Oblivious (PACO). We demonstrate our approach on several important cache-oblivious algorithms, including longest common sub-sequence (LCS), classic rectangular matrix multiplication (MM), Strassen's algorithm, and comparison-based sorting. By our approach, we provide an almost exact solution to the open problem on parallelizing Strassen. We discuss how to extend our approach to a distributed-memory architecture, or even a heterogeneous computing system. Hence, our work may provide a new perspective on the fundamental open problem of extending the recursive cache-oblivious technique to an arbitrary architecture. By preliminary experiments, our algorithms outperform significantly state-of-the-art Processor-Oblivious (PO) and Processor-Aware (PA) counterparts.","PeriodicalId":218208,"journal":{"name":"Proceedings of the 32nd ACM Symposium on Parallelism in Algorithms and Architectures","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131884146","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 2