We state the Problem of Knot Identification as a way to achieve consensus in dynamic networks. The network adversary is asynchronous and not oblivious. The network may be disconnected throughout the computation. We determine the necessary and sufficient conditions for the existence of a solution to the Knot Identification Problem: the knots must be observable by all processes and the first observed knot must be the same for all processes. We present an algorithm KIA that solves it. We conduct KIA performance evaluation.
We present a two-stage algorithm for the parallel reduction of a pencil to Hessenberg-triangular form. Traditionally, two-stage Hessenberg-triangular reduction algorithms achieve high performance in the first stage, but struggle to achieve high performance in the second stage. Our algorithm extends techniques described by Karlsson et al. to also achieve high performance in the second stage. Experiments in a shared memory environment demonstrate that the algorithm can outperform state-of-the-art implementations.
This paper presents a approach for measuring the time spent by HPC applications in the operating system's kernel. We use the SystemTap interface to insert timers before and after system calls, and take advantage of its stability to design a tool that can be used with multiple versions of the kernel. We evaluate its performance overhead, using an OS-intensive mini-benchmark and a raytracing mini app.
We aims to provide an interpretation of the the design background, motivation, and key innovations of FreeRider. The technique utilized by FreeRider enables tags to transform codewords present in commodity signals into another valid ones from the same codebook during reflection. As a result, the backscattered signal remains valid as commodity radios such as ZigBee or Bluetooth. By using commodity radios, FreeRider's backscatter system addresses the issue of specialized hardware requirements, thereby expanding its potential applications across various fields such as smart homes, healthcare, and industrial environments.
We propose a parallel graph-based data clustering algorithm using CUDA GPU, based on exact clustering of the minimum spanning tree in terms of a minimum isoperimetric criteria. We also provide a comparative performance analysis of our algorithm with other related ones which demonstrates the general superiority of this parallel algorithm over other competing algorithms in terms of accuracy and speed.
Carlos G. Oliver, Alessandro Ricottone, Pericles Philippopoulos
We propose a proof-of-work algorithm that rewards blockchain miners for using computational resources to solve NP-complete puzzles. The resulting blockchain will publicly store and improve solutions to problems with real world applications while maintaining a secure and fully functional transaction ledger.
This paper presents the benefits of formal modelling and verification techniques for self-stabilising distributed algorithms. An algorithm is studied, that takes a set of processes connected by a tree topology and converts it to a ring configuration. The Coloured Petri net model not only facilitates the proof that the algorithm is correct and self-stabilising but also easily shows that it enjoys new properties of termination and silentness. Further, the formal results show how the algorithm can be simplified without loss of generality.
High Performance Computing is notorious for its long and expensive software development cycle. To address this challenge, we present Bind: a "partitioned global workflow" parallel programming model for C++ applications that enables quick prototyping and agile development cycles for high performance computing software targeting heterogeneous distributed many-core architectures. We present applications of Bind to Linear Algebra and MapReduce algorithms alongside with performance results.
Today, a paradigm shift is being observed in science, where the focus is gradually shifting toward the cloud environments to obtain appropriate, robust and affordable services to deal with Big Data challenges (Sharma et al. 2014, 2015a, 2015b). Cloud computing avoids any need to locally maintain the overly scaled computing infrastructure that include not only dedicated space, but the expensive hardware and software also. In this paper, we study the evolution of as-a-Service modalities, stimulated by cloud computing, and explore the most complete inventory of new members beyond traditional cloud computing stack.
In this paper we describe an autotuning tool for optimization of OpenMP applications on highly multicore and multithreaded architectures. Our work was motivated by in-depth performance analysis of scientific applications and synthetic benchmarks on IBM Power 775 architecture. The tool provides an automatic code instrumentation of OpenMP parallel regions. Based on measurement of chosen hardware performance counters the tool decides on the number of parallel threads that should be used for execution of chosen code fragments.
Opacity of Transactional Memory is proposed to be established by incremental validation. Quiescence in terms of epoch-based memory reclamation is applied to deal with doomed transactions causing memory access violations. This method unfortunately involves increased memory consumption and does not cover reclamations outside of transactions. This paper introduces a different method which combines incremental validation with elements of sandboxing to solve these issues.
Cloud computing is getting mature, and the interoperability and standardization of the clouds is still waiting to be solved. This paper discussed the interoperability among clouds about message transmission, data transmission and virtual machine transfer. Starting from IEEE Pioneering Cloud Computing Initiative, this paper discussed about standardization of the cloud computing, especially intercloud cloud computing. This paper also discussed the standardization from the market-oriented view.
This paper presents a comparative study of distributed systems and the security issues associated with those systems. Four commonly used distributed systems were considered for detailed analysis in terms of technologies involved, security issues faced by them and solution proposed to circumvent those issues. Finally the security issues and the solutions were summarized and compared with each other.
The paper is devoted to a mathematical model of concurrency the special case of which is asynchronous system. Distributed asynchronous automata are introduced here. It is proved that the Petri nets and transition systems with independence can be considered like the distributed asynchronous automata. Time distributed asynchronous automata are defined in standard way by the map which assigns time intervals to events. It is proved that the time distributed asynchronous automata are generalized the time Petri nets and asynchronous systems.
This paper presents a realization for the reliable and fast startup of distributed systems written in Erlang. The traditional startup provided by the Erlang/OTP library is sequential, parallelization usually requires unsafe and ad-hoc solutions. The proposed method calls only for slight modifications in the Erlang/OTP stdlib by applying a system dependency graph. It makes the startup safe, quick, and it is equally easy to use in newly developed and legacy systems.
We address the problem of finding nice labellings for event structures of degree 3. We develop a minimum theory by which we prove that the labelling number of an event structure of degree 3 is bounded by a linear function of the height. The main theorem we present in this paper states that event structures of degree 3 whose causality order is a tree have a nice labelling with 3 colors. Finally, we exemplify how to use this theorem to construct upper bounds for the labelling number of other event structures of degree 3.
Entry in: Encyclopedia of Algorithms, Ming-Yang Kao, Ed., Springer, To appear. Synonyms: Wait-free registers, wait-free shared variables, asynchronous communication hardware. Problem Definition: Consider a system of asynchronous processes that communicate among themselves by only executing read and write operations on a set of shared variables (also known as shared registers). The system has no global clock or other synchronization primitives.