A5034054372- Parallel Computing and Optimization Techniques
- Distributed and Parallel Computing Systems
- Peer-to-Peer Network Technologies
- Advanced Data Storage Technologies
- Caching and Content Delivery
- Embedded Systems Design Techniques
- Interconnection Networks and Systems
- Logic, programming, and type systems
- Internet Traffic Analysis and Secure E-voting
- Cloud Computing and Resource Management
- Algorithms and Data Compression
- Distributed systems and fault tolerance
- Graph Theory and Algorithms
- Network Security and Intrusion Detection
- Ferroelectric and Negative Capacitance Devices
- Advanced Wireless Network Optimization
- Wireless Networks and Protocols
- Numerical Methods and Algorithms
- Digital Filter Design and Implementation
- Service-Oriented Architecture and Web Services
- Advanced Malware Detection Techniques
- Opportunistic and Delay-Tolerant Networks
- Anomaly Detection Techniques and Applications
- Game Theory and Applications
- Semantic Web and Ontologies
Delft University of Technology
2008-2022
University of Amsterdam
2014
Computational Physics (United States)
1994
This paper presents a comprehensive performance comparison between CUDA and OpenCL. We have selected 16 benchmarks ranging from synthetic applications to real-world ones. make an extensive analysis of the gaps taking into account programming models, ptimization strategies, architectural details, underlying compilers. Our results show that, for most applications, performs at 30% better than also that this difference is due unfair comparisons: in fact, OpenCL can achieve similar under fair...
Power consumption is the limiting factor for functionality of future wearable devices. Since interactive applications like wireless information access generate bursts activities, it important to match performance device accordingly. This paper describes a system with microprocessor whose speed can be varied (frequency scaling) as well its supply voltage. Voltage scaling reducing power very low values when operating at speeds. Measurements show that energy per instruction minimal 1/5 required...
Streaming multimedia content in real-time over a wireless link is challenging task because of the rapid fluctuations conditions that can occur due to movement, interference, and so on. The popular IEEE 802.11 standard includes low-level tuning parameters like transmission rate. Standard device drivers for today's products are based on gathering statistics, consequently, adapt rather slowly changes conditions. To meet strict latency requirements streaming applications, we designed implemented...
Based on Intel's Many Integrated Core (MIC) architecture, Intel Xeon Phi is one of the few truly many-core CPUs - featuring around 60 fairly powerful cores, two levels caches, and graphic memory, all interconnected by a very fast ring. Given its promised ease-of-use high performance, we took out for test drive. In this paper, present experience at different levels: (1) microbenchmark level, where stress "each nut bolt" in lab, (2) application study Phi's performance response real-life...
The Distributed ASCI Supercomputer (DAS) is a homogeneous wide-area distributed system consisting of four cluster computers at different locations. DAS has been used for research on communication software, parallel languages and programming systems, schedulers, applications, applications. paper gives preview the most interesting results obtained so far in project.
Article Share on Energy priority scheduling for variable voltage processors Authors: Johan Pouwelse Faculty of Information Technology and Systems, Delft University Technology, The Netherlands NetherlandsView Profile , Koen Langendoen Henk Sips Authors Info & Claims ISLPED '01: Proceedings the 2001 international symposium Low power electronics designAugust 2001Pages 28–33https://doi.org/10.1145/383082.383089Published:06 August 2001Publication History 51citation350DownloadsMetricsTotal...
Seamless video streaming over wireless links imposes strong demands on codecs and the underlying network. It is not sufficient that only codec or radio adapts to changes in link quality; efforts should be applied both layers, - if possible synchronized. Also, disturbing effect of background traffic same shared medium has taken into account. In this article we present a communication architecture for 802.11 capable adapting quality sharing channel various use scenarios. Experimental results...
A well-known problem in P2P systems is freeriding, where users do not share content if there no incentive to so. In this paper, we distinguish lazy freeriders that are merely reluctant but follow the protocol, versus die-hard employ sophisticated methods subvert protocol. Existing designs often provide theoretically attractive resistance against yet rarely deployed real networks because of practical infeasibility. Meanwhile, communities benefit greatly from prevention have only centralized...
OpenCL and OpenMP are the most commonly used programming models for multi-core processors. They also fundamentally different in their approach to parallelization. In this paper, we focus on comparing performance of OpenMP. We select three applications from Rodinia benchmark suite (which provides equivalent implementations), carry out experiments with datasets platforms. see that incorrect usage CPUs, inherent fine-grained parallelism, immature compilers main reasons lead poorer performance....
Heterogeneous platforms composed of multi-core CPUs and different types accelerators, like GPUs Xeon Phi, are becoming popular for data parallel applications. The heterogeneity the hardware mix diversity applications pose significant challenges to exploiting such platforms. In this situation, an effective workload partitioning between processing units is critically important improving application performance. This a function capabilities as well dataset be used. work, we present systematic...
Abstract Streaming multimedia content in real‐time over a wireless link is challenging task because of the rapid fluctuations conditions that can occur due to movement, interference, and so on. The popular IEEE 802.11 standard includes low‐level tuning parameters like transmission rate. Standard device drivers for today's products are based on gathering statistics consequently, adapt rather slowly changes conditions. To meet strict latency requirements streaming applications, we designed...
Most P2P systems that have some kind of incentive mechanism reward peers according to their contribution, i.e. total bandwidth offered the system. Due disparity in capacity between users on Internet, common effect such mechanisms is fastest reap highest benefits. We take a different approach and study how incentivize cooperation based effort, contribution relative capacity. make following contributions: 1) we argue contribution-based schemes unnecessarily disfavor slow decrease overall...
Distributed systems without a central authority, such as peer-to-peer (P2P) systems, employ incentives to encourage nodes follow the prescribed protocol. Game theoretic analysis is often used evaluate in systems. However, most game-theoretic analyses of distributed do not adequately model repeated interactions inherent We present popular P2P protocol, Bit-Torrent, that models protocols. also note an analytical approach for modeling infeasible given complicated nature deployed In order...
With at least 50 cores, Intel Xeon Phi is a true many-core architecture. Featuring fairly powerful two cache levels, and very fast interconnections, the can get theoretical peak of 1000 GFLOPs over 240 GB/s. These numbers, as well its flexibility - it be used both coprocessor or stand-alone processor are tempting for parallel applications looking new performance records. In this paper, we present an empirical study Phi, stressing limits relevant factors, ultimately aiming to simplified view...
A number of new algorithms to make an online CORDIC implementation are introduced. The algorithm takes n+6 clock cycles compute a function. It is estimated that the proposed six seven times as fast traditional approach for Given's rotation, and eight SVD computation.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">></ETX>
The main purpose of many current peer-to-peer (P2P) networks is off-line file sharing. However, a potentially very promising use such to share video streams (e.g., TV programs) in real time. In order do so, the peers P2P network who are interested same stream may employ Application Level Multicasting (ALM). existing networks, exhibit behavior which problematic for ALM: they not always willing donate resources (free-riding), and arrive depart at high rate (churn). this paper we propose...
The BitTorrent protocol is by far the most popular for offline peer-to-peer video distribution on Internet. has previously been extended to support streaming of recorded video, that is, Video-on-Demand (VoD). In this paper, we take a step further presenting extensions supporting live streaming, which have implemented in our client called Tribler. We tested both running simulations, and deploying implementation public trial Internet, using optimal values several parameters as found...
Peer-to-peer file-sharing networks depend on peers uploading data to each other. Some peers, called free-riders, will not upload unless there is an incentive do so. Algorithms designed prevent free-riding typically assume that connectivity a problem. However, the Internet, large fraction of resides behind firewall or NAT, making them unable accept incoming connections. In this paper, we prove it impossible when more than half are firewalled, and provide bounds sharing ratios (defined as...
With its design concept of cross-platform portability, OpenCL can be used not only on GPUs (for which it is quite popular), but also CPUs. Whether porting GPU programs to CPUs, or simply writing new code for using brings up the performance issue, usually raised in one two forms: "OpenCL portable!" "Why CPUs after all?!". We argue that both issues addressed by a thorough study factors impact This analysis focus this paper. Specifically, starting from main architectural mismatches between...