Journal ArticleDOI
Scheduling multithreaded computations by work stealing
TLDR
This paper gives the first provably good work-stealing scheduler for multithreaded computations with dependencies, and shows that the expected time to execute a fully strict computation on P processors using this scheduler is 1:1.Abstract:
This paper studies the problem of efficiently schedulling fully strict (i.e., well-structured) multithreaded computations on parallel computers. A popular and practical method of scheduling this kind of dynamic MIMD-style computation is “work stealing,” in which processors needing work steal computational threads from other processors. In this paper, we give the first provably good work-stealing scheduler for multithreaded computations with dependencies.Specifically, our analysis shows that the expected time to execute a fully strict computation on P processors using our work-stealing scheduler is T1/P + O(T ∞ , where T1 is the minimum serial execution time of the multithreaded computation and (T ∞ is the minimum execution time with an infinite number of processors. Moreover, the space required by the execution is at most S1P, where S1 is the minimum serial space requirement. We also show that the expected total communication of the algorithm is at most O(PT ∞( 1 + nd)Smax), where Smax is the size of the largest activation record of any thread and nd is the maximum number of times that any thread synchronizes with its parent. This communication bound justifies the folk wisdom that work-stealing schedulers are more communication efficient than their work-sharing counterparts. All three of these bounds are existentially optimal to within a constant factor.read more
Citations
More filters
Proceedings ArticleDOI
Learning scheduling algorithms for data processing clusters
TL;DR: Decima as discussed by the authors uses reinforcement learning and neural networks to learn workload-specific scheduling algorithms without any human instruction beyond a high-level objective, such as minimizing average job completion time, and shows that RL techniques can generate highly-efficient policies automatically.
Posted Content
Learning Scheduling Algorithms for Data Processing Clusters
TL;DR: It is shown that modern machine learning techniques can generate highly-efficient policies automatically and improve average job completion time by at least 21% over hand-tuned scheduling heuristics, achieving up to 2x improvement during periods of high cluster load.
Proceedings ArticleDOI
CIEL: a universal execution engine for distributed data-flow computing
Derek G. Murray,Malte Schwarzkopf,Christopher Smowton,Steven G. Smith,Anil Madhavapeddy,Steven Hand +5 more
TL;DR: The execution engine provides transparent fault tolerance and distribution to Skywriting scripts and high-performance code written in other programming languages, and achieves scalable performance for both iterative and non-iterative algorithms.
Posted Content
Demystifying Parallel and Distributed Deep Learning: An In-Depth Concurrency Analysis
Tal Ben-Nun,Torsten Hoefler +1 more
TL;DR: The problem of parallelization in DNNs is described from a theoretical perspective, followed by approaches for its parallelization, and potential directions for parallelism in deep learning are extrapolated.
Proceedings ArticleDOI
Dynamic circular work-stealing deque
David R. Chase,Yossi Lev +1 more
TL;DR: This work presents a simple lock-free work-stealing deque, which stores the elements in a cyclic array that can grow when it overflows, and has no limit other than integer overflow on the number of elements on the deque.
References
More filters
Journal ArticleDOI
Cilk: An Efficient Multithreaded Runtime System
Robert D. Blumofe,Christopher F. Joerg,Bradley C. Kuszmaul,Charles E. Leiserson,Keith H. Randall,Yuli Zhou +5 more
TL;DR: It is shown that on real and synthetic applications, the “work” and “critical-path length” of a Cilk computation can be used to model performance accurately, and it is proved that for the class of “fully strict” (well-structured) programs, the Cilk scheduler achieves space, time, and communication bounds all within a constant factor of optimal.
Journal ArticleDOI
Bounds for certain multiprocessing anomalies
TL;DR: In this paper, precise bounds are derived for several anomalies of this type in a multiprocessing system composed of many identical processing units operating in parallel, and they show that an increase in the number of processing units can cause an increased total length of time needed to process a fixed set of tasks.
Proceedings ArticleDOI
The implementation of the Cilk-5 multithreaded language
TL;DR: Cilk-5's novel "two-clone" compilation strategy and its Dijkstra-like mutual-exclusion protocol for implementing the ready deque in the work-stealing scheduler are presented.
Journal ArticleDOI
The Parallel Evaluation of General Arithmetic Expressions
TL;DR: It is shown that arithmetic expressions with n ≥ 1 variables and constants; operations of addition, multiplication, and division; and any depth of parenthesis nesting can be evaluated in time 4 log 2 + 10(n - 1) using processors which can independently perform arithmetic operations in unit time.