Presence monitoring or intrusion detection in a location/area are examples of decision-support applications. Decision-support applications are applications where monitoring is used to collect (heterogeneous) data and create situational awareness, which further requires decisions and/or actions. As such, decision-support software consists of different interconnected components with very diverse roles, whose communication and synchronization are essential for the application functionality and performance. Despite this complexity, software design for decision-support is often driven by short-term functional requirements and only supported by designers’ previous experience. In the current non-systematic approach, mistakes can be easily made, and can be very difficult to repair. In this work, we describe our systematic method for efficient and effective decision-support software design, based on application design-space exploration (DSE). To this end, we describe how to build a design space and present structured methods to traverse the design space towards software solutions that meet user requirements for both functionality and performance.

Many situations in the security domain require decision-making based on complex data, i.e., many variables which need to be taken into account before adequate decisions can be made. For example, in a surveillance scenario, the size and complexity of the area of interest, the mix of objects, and the unexpected behavior of suspects are just a few examples of complex variables to be analyzed in the process. Existing decision support systems provide some analysis, but are typically limited in the complexity they can handle. Therefore, users end up with simplified models which often suffer in the accuracy of their decisions and, ultimately, may lead to incorrect decisions. In this work, we present a framework that can scale to cope with the complexity and time requirements of real-world scenarios, while remaining flexible to handle the ad-hoc adaptation to the situation. We discuss the challenges and solutions for such a scalable and flexible system, and validate it using a target tracking scenario in urban environments of different sizes.

This paper examines the impact of heavy churn on the robustness of decentralized social network-based Sybil defense (SNSD) schemes. Our analysis reveals that (i) heavy churn disintegrates the social overlay network that is fundamental to these schemes into multiple disconnected components, resulting in poor network connectivity, and (ii) a naive solution that adds links from each node to all its 2-hop neighbors improves network connectivity but comes at a significant cost of poor attack resilience of these schemes. We propose a new design point in the trade-off between network connectivity and attack resilience of SNSD schemes, where each node adds links to only a selective few of all its 2-hop neighbors based on a minimum expansion contribution (MinEC) heuristic. Extensive evaluation through simulations shows that our approach fares as good as the naive 2-hop solution in terms of network connectivity, while making little compromise on the attack resilience. Moreover, our approach preserves the fast-mixing property that is fundamental to many SNSD schemes even at high levels of churn. This result suggests that existing and potential future SNSD schemes relying on this property can incorporate our approach into their designs with minimal changes.

Heterogeneous platforms integrating different types of processing units (such as multi-core CPUs and GPUs) are in high demand in high performance computing. Existing studies have shown that using heterogeneous platforms can improve application performance and hardware utilization. However, systematic methods to design, implement, and map applications to efficiently use heterogeneous computing resources are only very few. The goal of my PhD research is therefore to study such heterogeneous systems and propose systematic methods to allow many (classes of) applications to efficiently use them. After 3.5 years of PhD study, my contributions are (1) a thorough evaluation of a suitable programming model for heterogeneous computing, (2) a workload partitioning framework to accelerate parallel applications on heterogeneous platforms, (3) a modelling-based prediction method to determine the optimal workload partitioning, (4) a systematic approach to decide the best mapping between the application and the platform by choosing the best performing hardware configuration (Only-CPU, Only-GPU, or CPU+GPU with the workload partitioning). In the near future, I plan to apply my approach to large-scale applications and platforms to expand its usability and applicability.

Heterogeneous platforms are mixes of different processing units. The key factor to their efficient usage is workload partitioning. Both static and dynamic partitioning strategies have been defined in previous work, but their applicability and performance differ significantly depending on the application to execute. In this paper, we propose an application-driven method to select the best partitioning strategy for a given workload. To this end, we define an application classification based on the application kernel structure-i.e., the number of kernels in the application and their execution flow. We also enable five different partitioning strategies, which mix the best features of both static and dynamic approaches. We further define the performance-driven ranking of all suitable strategies for each application class. Finally, we match the best partitioning to a given application by simply determining its class and selecting the best ranked strategy for that class. We test the matchmaking on six representative applications, and demonstrate that the defined performance ranking is correct. Moreover, by choosing the best performing partitioning strategy, we can significantly improve application performance, leading to average speedup of 3.0x/5.3x over the Only-GPU/Only-CPU execution, respectively.

Network-based detection of botnet Command and Control communication is a difficult task if the traffic has a relatively low volume and if popular protocols, such as HTTP, are used to resemble normal traffic. We present a new network-based detection approach that is capable of detecting this type of Command and Control traffic in an enterprise network by estimating the trustworthiness of the traffic destinations. If the destination identifier of a traffic flow origins directly from: human input, prior traffic from a trusted destination, or a defined set of legitimate applications, the destination is trusted and its associated traffic is classified as normal. Advantages of this approach are: the ability of zero day malicious traffic detection, low exposure to malware by passive host-external traffic monitoring, and the applicability for real-time filtering. Experimental evaluation demonstrates successful detection of diverse types of Command and Control Traffic.

Heterogeneous platforms composed of multi-core CPUs and different types of accelerators, like GPUs and Xeon Phi, are becoming popular for data parallel applications. The heterogeneity of the hardware mix and the diversity of the applications pose significant challenges to exploiting such platforms. In this situation, an effective workload partitioning between processing units is critically important for improving application performance. This partitioning is a function of the hardware capabilities as well as the application and the dataset to be used. In this work, we present a systematic approach to solve the partitioning problem. Specifically, we use modeling, profiling, and prediction techniques to quickly and correctly predict the optimal workload partitioning and the right hardware configuration to use. Our approach effectively characterizes the platform heterogeneity, efficiently determines the accurate partitioning, and easily adapts to new platforms, different application types, and different datasets. Experimental evaluation on 13 applications shows that our approach delivers excellent performance improvement of 1.2 $\times$ –14.6 $\times$ over a single-processor execution, and accurate partitioning with in most cases below 10 percent performance gap versus an oracle-based partitioning.

SummaryOpen Computing Language (OpenCL) is an open, functionally portable programming model for a large range of highly parallel processors. To provide users with access to the underlying platforms, OpenCL has explicit support for features such as local memory and vector data types (VDTs). However, these are often low‐level, hardware‐specific features, which can be detrimental to performance on different platforms. In this paper, we focus on VDTs and investigate their usage in a systematic way. First, we propose two different approaches (inter‐vdt and intra‐vdt) to use VDTs in OpenCL kernels, and show how to translate scalar OpenCL kernels to vectorized ones. After obtaining vectorized code, we evaluate the performance effects of using VDTs with two types of benchmarks: micro‐benchmarks and macro‐benchmarks. With micro‐benchmarks, we study the execution model of VDTs and the role of the compiler‐aided vectorizer on five devices. With macro‐benchmarks, we explore the changes of memory access patterns before and after using VDTs, and the resulting performance impact. Not only our evaluation provides insights into how OpenCL's VDTs are mapped on different processors, but it also indicates that using such data types introduces changes in both computation and memory accesses. Based on the lessons learned, we discuss how to deal with performance portability in the presence of VDTs. Copyright © 2014 John Wiley & Sons, Ltd.

Due to increasingly large datasets, graph analytics - traversals, all-pairs shortest path computations, centrality measures, etc. - are becoming the focus of high-performance computing (HPC). Because HPC is currently dominated by many-core architectures (both CPUs and GPUs), new graph processing solutions have to be defined to efficiently use such computing resources. Prior work focuses on platform-specific performance studies and on platform-specific algorithm development, successfully proving that algorithms highly tuned to GPUs or multi-core CPUs can provide high performance graph analytics. However, the portability of such algorithms remains an important concern for many users, especially the many companies without the resources to invest in HPC or concerned about lock-in in single-use parallel techniques. In this work, we investigate the functional portability and performance of graph analytics algorithms. We conduct an empirical study measuring the performance of 3 graph analytics algorithms (a single code implemented in OpenCL and targeted at many-core CPUs and GPUs), on 3 different platforms, using 11 real-world and synthetic datasets. Our results show that the code is functionally portable, that is, applications can run unchanged on both CPUs and GPUs. The large variation in their observed performance indicates that portability is necessary not only for productivity, but, surprisingly, also for performance. We conjecture that the impact of datasets on performance is too high to allow platform-specific algorithms to outperform the portable algorithms by large margins, in all cases. Our conclusion is that portable parallel graph analytics is feasible without significant performance loss, and provides a productive alternative to the expensive trial-and-error selection of one algorithm for each (graph,platform) pair.

We present a novel anomaly-based detection approach capable of detecting botnet Command and Control traffic in an enterprise network by estimating the trustworthiness of the traffic destinations. A traffic flow is classified as anomalous if its destination identifier does not origin from: human input, prior traffic from a trusted destination, or a defined set of legitimate applications. This allows for real-time detection of diverse types of Command and Control traffic. The detection approach and its accuracy are evaluated by experiments in a controlled environment.

Due to the diversity of processor architectures and application memory access patterns, the performance impact of using local memory in OpenCL kernels has become unpredictable. For example, enabling the use of local memory for an OpenCL kernel can be beneficial for the execution on a GPU, but can lead to performance losses when running on a CPU. To address this unpredictability, we propose an empirical approach: by disabling the use of local memory in OpenCL kernels, we enable users to compare the kernel versions with and without local memory, and further choose the best performing version for a given platform. To this end, we have designed Grover, a method to automatically remove local memory usage from OpenCL kernels. In particular, we create a correspondence between the global and local memory spaces, which is used to replace local memory accesses by global memory accesses. We have implemented this scheme in the LLVM framework as a compiling pass, which automatically transforms an OpenCL kernel with local memory to a version without it. We have validated Grover with 11 applications, and found that it can successfully disable local memory usage for all of them. We have compared the kernels with and without local memory on three different processors, and found performance improvements for more than a third of the test cases after Grover disabled local memory usage. We conclude that such a compiler pass can be beneficial for performance, and, because it is fully automated, it can be used as an auto-tuning step for OpenCL kernels.

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 of caches, and graphic memory, all interconnected by a very fast ring. Given its promised ease-of-use and high performance, we took Xeon Phi out for a test drive. In this paper, we present this experience at two different levels: (1) the microbenchmark level, where we stress "each nut and bolt" of Phi in the lab, and (2) the application level, where we study Phi's performance response in a real-life environment. At the microbenchmarking level, we show the high performance of five components of the architecture, focusing on their maximum achieved performance and the prerequisites to achieve it. Next, we choose a medical imaging application (Leukocyte Tracking) as a case study. We observed that it is rather easy to get functional code and start benchmarking, but the first performance numbers can be far from satisfying. Our experience indicates that a simple data structure and massive parallelism are critical for Xeon Phi to perform well. When compiler-driven parallelization and/or vectorization fails, programming Xeon Phi for performance can become very challenging.

Due to the increasing complexity of multi/many-core architectures (with their mix of caches and scratch-pad memories) and applications (with different memory access patterns), the performance of many workloads becomes increasingly variable. In this work, we address one of the main causes for this performance variability: the efficiency of the memory system. Specifically, based on an empirical evaluation driven by memory access patterns, we qualify and partially quantify the performance impact of using local memory in multi/many-core processors. To do so, we systematically describe memory access patterns (MAPs) in an application-agnostic manner. Next, for each identified MAP, we use OpenCL (for portability reasons) to generate two microbenchmarks: a “naive” version (without local memory) and “an optimized” version (using local memory). We then evaluate both of them on typically used multi-core and many-core platforms, and we log their performance. What we eventually obtain is a local memory performance database, indexed by various MAPs and platforms. Further, we propose a set of composing rules for multiple MAPs. Thus, we can get an indicator of whether using local memory is beneficial in the presence of multiple memory access patterns. This indication can be used to either avoid the hassle of implementing optimizations with too little gain or, alternatively, give a rough prediction of the performance gain.

GPUs are widely used to accelerate data-parallel applications. However, while the GPU processing capability is enhanced in each generation, the CPU computing power is also increased by adding more cores and widening vector units. Compared to the rapid development of GPUs and CPUs, the bandwidth of the data transfer between GPUs and the host CPU grows much slower, resulting in a data-transfer wall for using GPUs. In this situation, choosing the right mix of hardware resources - i.e., The right hardware configuration - is critically important for improving application performance, and the right choice is a function of the available hardware resources as well as the application and the dataset to be used. In this paper, we present a systematic approach to determine the hardware configuration that leads to the best performance for a given workload. Our approach captures the variation of hardware capabilities and data-transfer overhead for different applications and datasets, and uses modeling and prediction techniques to determine the best-performing hardware configuration. We have tested our approach on 7 applications with 6 datasets per application. The results show that our approach takes the correct decision in 38 out of 42 test cases, achieving up to 12.6×/6.6× performance improvement compared to an uninformed Only-CPU/Only-GPU decision.

In recent years fully decentralized file sharing systems were developed aimed at improving anonymity among their users. These systems provide typical file sharing features such as searching for and downloading files. However, elaborate schemes originally aimed at improving anonymity cause partial keyword matching to be virtually impossible, or introduce a substantial bandwidth overhead. In this paper we introduce 4P, a system that provides users with anonymous search on top of a semantic overlay. The semantic overlay allows users to efficiently locate files using partial keyword matching, without having to resort to an expensive flooding operation. Included into 4P are a number of privacy enhancing features such as probabilistic query forwarding, path uncertainty, caching, and encrypted links. Moreover, we integrate a content retrieval channel into our protocol allowing users to start downloading a file from multiple sources immediately without requiring all intermediate nodes to cache a complete copy. Using a trace-based dataset, we mimic a real-world query workload and show the cost and performance of search using six overlay configurations, comparing random, semantic, Gnutella, RetroShare, and OneSwarm to 4P. The state-of-the-art flooding based alternatives required approximately 10,000 messages to be sent per query, in contrast 4P only required 313. Showing that while flooding can achieve a high recall (more than 85% in our experiments) it is prohibitively expensive. With 4P we achieve a recall of 76% at a considerable reduction in messages sent.

Although GPUs are considered ideal to accelerate massively data-parallel applications, there are still exceptions to this rule. For example, imbalanced applications cannot be efficiently processed by GPUs: despite the massive data parallelism, a varied computational workload per data point remains GPU-unfriendly. To efficiently process imbalanced applications, we exploit the use of heterogeneous platforms (GPUs and CPUs) by partitioning the workload to fit the usage patterns of the processors. In this work, we present our flexible and adaptive method that predicts the optimal partitioning. Our method aims to match a quantitative model of the application with the hardware capabilities of the platform, and calculates the optimal match according to a user-given criterion. We evaluate our method in terms of overall performance gain, prediction accuracy, flexibility and adaptivity. Our results, gathered from both synthetic and real-world workloads, show performance gains of up to 60%, accurate predictions for more than 90% of all the 1395 imbalanced workloads we have tested, and confirm that the method adapts correctly to application, dataset, and platform changes (both hardware and software). We conclude that model-based prediction of workload partitioning for heterogeneous platforms is feasible and useful for performance improvement.

Heterogeneous platforms integrating different processors like GPUs and multi-core CPUs become popular in high performance computing. While most applications are currently using the homogeneous parts of these platforms, we argue that there is a large class of applications that can benefit from their heterogeneity: massively parallel imbalanced applications. Such applications emerge, for example, from variable time step based numerical methods and simulations. In this paper, we present Glinda, a framework for accelerating imbalanced applications on heterogeneous computing platforms. Our framework is able to correctly detect the application workload characteristics, make choices based on the available parallel solutions and hardware configuration, and automatically obtain the optimal workload decomposition and distribution. Our experiments on parallelizing a heavily imbalanced acoustic ray tracing application show that Glinda improves application performance in multiple scenarios, achieving up to 12x speedup against manually configured parallel solutions.

Due to the increasing complexity of multi/many-core architectures (with their mix of caches and scratch-pad memories) and applications (with different memory access patterns), the performance of many workloads becomes increasingly variable. In this work, we address one of the main causes for this performance variability: the efficiency of the memory system. Specifically, based on an empirical evaluation driven by memory access patterns, we qualify and partially quantify the performance impact of using local memory in multi/many-core processors. To do so, we systematically describe memory access patterns (MAPs) in an application-agnostic manner. Next, for each identified MAP, we use OpenCL (for portability reasons) to generate two microbenchmarks: a “naive” version (without local memory) and “an optimized” version (using local memory). We further evaluate both of them on typically used multi-core and many-core platforms, and we log their performance. What we eventually obtain is a local memory performance database, indexed by various MAPs and platforms. To show how our database can be used to provide information on local memory usage, we demonstrate a scenario for performance prediction: given an application, its MAPs, and a platform, a query in the database can indicate the performance impact of using local memory for the given application. This indication can be used to either avoid the hassle of implementing optimizations with too little gain or, alternatively, give a rough prediction of the performance gain.

No part of this series may be reproduced in any form or by any means without prior written permission of the publisher. PDS Wp Wp With a minimum of 50 cores, Intel's Xeon Phi is a true many-core architecture. Featuring fairly powerful cores, two levels of caches, and a very fast interconnection, the Xeon Phi is able to achieve theoretical peak of 1000 GFLOPs and over 240 GB/s. These numbers, as well as its flexibility-it can be used as both coprocessor or a stand-alone processor-are very tempting for parallel applications looking for new performance records. In this paper, we present four hardware-centric guidelines and a machine model for Xeon Phi programmers in search for performance. Specifically, we have benchmarked the main hardware components of the processor-the cores, the memory hierarchies, and the ring interconnect. We show that, in ideal microbenchmarking conditions, the achieved performance is very close to the theoretical one as given in the official programmer's guide. Furthermore, we have identified and quantified several causes for significant performance penalties, which are not available in the official documentation. Based on this information, we synthesized four optimization guidelines and applied them to a set of kernels, aiming to systematically optimize their performance. The optimization process is guided by performance roofs, derived from the same benchmarks. Our experimental results show that, using this strategy, we can achieve impressive performance gains and, more importantly, a high utilization of the processor.