Interprocessor Communication Research Articles

High-performance data mining with intelligent SSD

An intuitive way to process the big data efficiently is to reduce the volume of data transferred over the storage interface to a host system. This is the reason that the notion of intelligent SSD (iSSD) was proposed to give processing power to SSD. There is rich literature on iSSD, however, its real implementation has not been provided to the public yet. Most prior work aims to quantify the benefits of iSSD with analytical modeling. In this paper, we first develop on iSSD simulator and present the potential of iSSD in data mining through the iSSD simulator. Our iSSD simulator performs on top of the gem 5 simulator and fully simulates all the processes of data mining algorithms running in iSSD with cycle-level accuracy. Then, we further addresse how to exploit all the computing resources for efficient processing of data mining algorithms. These days, CPU, GPU, and SSD are recently equipped together in most computing environment. If SSD is replaced with iSSD later on, we have a new computing environment where the three computing resources collaborate one another to process big data quite effectively. For this, scheduling is required to decide which computing resource is going to run for which function at which time. In our heterogeneous scheduling, types of computing resources, memory sizes in computing resources, and inter-processor communication times including IO time in SSD are considered. Our scheduling results show that processing in the collaborative environment outperforms that in the traditional one by up to about 10 times.

An Novel Architecture and Inter-process Communication Scheme to Adapt Chromium Based on Docker Container

In this paper, firstly we propose a novel architecture called DBCSUEE to adapt embedded Chromium browser with Docker container virtualization as its sandboxing mechanism to isolate processes. Using this kind of container-level virtualization in browsers guarantees web applications operate in a secure environment, and gives the browser a more efficient resource management scheme applicable to multi-application situation. Secondly, when implementing the scheme, communication among applications running in their own containers becomes a problem and needs to be solved. Chromium's existing inter-process communication mechanism cannot be used. Thus, we propose a new communication scheme based on named pipe, which allows two web applications to communicate with each other without closing their containers.

Performance analysis of Linux containers for high performance computing applications

Although cloud infrastructures can be used as High Performance Computing (HPC) platforms, many issues from virtualisation overhead have kept them almost unrelated. However, with advent of container-based virtualisation, this scenario acquires new perspectives because this technique promises to decrease the virtualisation overhead, achieving a near-native performance. In this work, we analyse the performance of a container-based virtualisation solution - Linux Container (LXC) - against a hypervisor-based virtualisation solution - KVM - under HPC activities. For our experiments, we consider CPU and (network and inter-process) communication performance. Results show that hypervisor type can impact distinctly in performance according to resource used by HPC application.

Hypergraph Partitioning for Sparse Matrix-Matrix Multiplication

We propose a fine-grained hypergraph model for sparse matrix-matrix multiplication (SpGEMM), a key computational kernel in scientific computing and data analysis whose performance is often communication bound. This model correctly describes both the interprocessor communication volume along a critical path in a parallel computation and also the volume of data moving through the memory hierarchy in a sequential computation. We show that identifying a communication-optimal algorithm for particular input matrices is equivalent to solving a hypergraph partitioning problem. Our approach is nonzero structure dependent, meaning that we seek the best algorithm for the given input matrices. In addition to our three-dimensional fine-grained model, we also propose coarse-grained one-dimensional and two-dimensional models that correspond to simpler SpGEMM algorithms. We explore the relations between our models theoretically, and we study their performance experimentally in the context of three applications that use SpGEMM as a key computation. For each application, we find that at least one coarse-grained model is as communication efficient as the fine-grained model. We also observe that different applications have affinities for different algorithms. Our results demonstrate that hypergraphs are an accurate model for reasoning about the communication costs of SpGEMM as well as a practical tool for exploring the SpGEMM algorithm design space.

Accelerating Intercommunication in Highly Parallel Systems

Every HPC system consists of numerous processing nodes interconnect using a number of different inter-process communication protocols such as Messaging Passing Interface (MPI) and Global Arrays (GA). Traditionally, research has focused on optimizing these protocols and identifying the most suitable ones for each system and/or application. Recently, there has been a proposal to unify the primitive operations of the different inter-processor communication protocols through the Portals library. Portals offer a set of low-level communication routines which can be composed in order to implement the functionality of different intercommunication protocols. However, Portals modularity comes at a performance cost, since it adds one more layer in the actual protocol implementation. This work aims at closing the performance gap between a generic and reusable intercommunication layer, such as Portals, and the several monolithic and highly optimized intercommunication protocols. This is achieved through the development of a novel hardware offload engine efficiently implementing the basic Portals’ modules. Our innovative system is up to two2 orders of magnitude faster than the conventional software implementation of Portals’ while the speedup achieved over the conventional monolithic software implementations of MPI and GAs is more than an order of magnitude. The power consumption of our hardware system is less than 1/100th of what a low-power CPU consumes when executing the Portal's software while its silicon cost is less than 1/10th of that of a very simple RISC CPU. Moreover, our design process is also innovative since we have first modeled the hardware within an untimed virtual prototype which allowed for rapid design space exploration; then we applied a novel methodology to transform the untimed description into an efficient timed hardware description, which was then transformed into a hardware netlist through a High-Level Synthesis (HLS) tool.

Newton’s Law of Universal Gravitation Based Genetic Algorithm for Graph Distribution

This paper proposes a novel versatile genetic algorithm (GA) for solving the graph distribution problem. The new GA is based mainly on an inspirational idea that exploits the Newton’s universal gravitation law to introduce a novel GA fitness function. By this,the new GA preserves the workload balancing property between the different sites of the graph network and reduces the inter-processors communications overhead. Moreover, three main variants of the novel GA are developed. The two first; centralized and distributed variants, are developed to conduct a graph distribution over homogeneous architectures. The third variant is a distributed one devoted for heterogeneous architectures where the impact of the auto-adaptation features of the GA emerges. The results obtained and the comparative studies show the effectiveness of the proposed methods.

AndroDialysis: Analysis of Android Intent Effectiveness in Malware Detection

The wide popularity of Android systems has been accompanied by increase in the number of malware targeting these systems. This is largely due to the open nature of the Android framework that facilitates the incorporation of third-party applications running on top of any Android device. Inter-process communication is one of the most notable features of the Android framework as it allows the reuse of components across process boundaries. This mechanism is used as gateway to access different sensitive services in the Android framework. In the Android platform, this communication system is usually driven by a late runtime binding messaging object known as Intent. In this paper, we evaluate the effectiveness of Android Intents (explicit and implicit) as a distinguishing feature for identifying malicious applications. We show that Intents are semantically rich features that are able to encode the intentions of malware when compared to other well-studied features such as permissions. We also argue that this type of feature is not the ultimate solution. It should be used in conjunction with other known features. We conducted experiments using a dataset containing 7406 applications that comprise 1846 clean and 5560 infected applications. The results show detection rate of 91% using Android Intent against 83% using Android permission. Additionally, experiment on combination of both features results in detection rate of 95.5%.

IMPETUS – Interactive MultiPhysics Environment for Unified Simulations

We introduce IMPETUS – Interactive MultiPhysics Environment for Unified Simulations, an object oriented, easy-to-use, high performance, C++ program for three-dimensional simulations of complex physical systems that can benefit a large variety of research areas, especially in cell mechanics. The program implements cross-communication between locally interacting particles and continuum models residing in the same physical space while a network facilitates long-range particle interactions. Message Passing Interface is used for inter-processor communication for all simulations.

A diagnostic interface for the ICOsahedral Non-hydrostatic (ICON) modelling framework based on the Modular Earth Submodel System (MESSy v2.50)

Abstract. Numerical climate and weather models have advanced to finer scales, accompanied by large amounts of output data. The model systems hit the input and output (I/O) bottleneck of modern high-performance computing (HPC) systems. We aim to apply diagnostic methods online during the model simulation instead of applying them as a post-processing step to written output data, to reduce the amount of I/O. To include diagnostic tools into the model system, we implemented a standardised, easy-to-use interface based on the Modular Earth Submodel System (MESSy) into the ICOsahedral Non-hydrostatic (ICON) modelling framework. The integration of the diagnostic interface into the model system is briefly described. Furthermore, we present a prototype implementation of an advanced online diagnostic tool for the aggregation of model data onto a user-defined regular coarse grid. This diagnostic tool will be used to reduce the amount of model output in future simulations. Performance tests of the interface and of two different diagnostic tools show, that the interface itself introduces no overhead in form of additional runtime to the model system. The diagnostic tools, however, have significant impact on the model system's runtime. This overhead strongly depends on the characteristics and implementation of the diagnostic tool. A diagnostic tool with high inter-process communication introduces large overhead, whereas the additional runtime of a diagnostic tool without inter-process communication is low. We briefly describe our efforts to reduce the additional runtime from the diagnostic tools, and present a brief analysis of memory consumption. Future work will focus on optimisation of the memory footprint and the I/O operations of the diagnostic interface.

An embedded self-adapting network service framework for networked manufacturing system

To improve the self-adapting ability and real-time performance of client/server based networked manufacturing system (NMS), this paper introduces the universal plug and play (UPnP), an intelligent network middleware, into networked manufacturing area, and proposes an embedded self-adapting network framework and related service methods. Referring to small world model and scale-free principles, a complex network model oriented to digital manufacturing is set up. Based on the model, an improved entropy vector projection algorithm is proposed to evaluate the network complexity and reveal the evolution regulars. Then, the self-adapting services for NMS are performed by UPnP service-calling and inter-process communication methods. Finally, the case studies and industrial field experiments verify the effectiveness of the proposed service framework.

Application Mapping and Scheduling for Network-on-Chip-Based Multiprocessor System-on-Chip With Fine-Grain Communication Optimization

Network-on-chip (NoC) is promising for the communication paradigm of the next-generation multiprocessor system-on-chip (MPSoC). As communication has become an integral part of on-chip computing, and even the performance bottleneck, researchers are paying much attention to its implementation and optimization. Traditional techniques that model communication inaccurately will lead to unexpected runtime performance, which is on average 90.8% worse than the predicted results based on observation, and are not suitable for the deep optimization of communication-intensive scenarios. In this paper, techniques are presented for the NoC-based MPSoCs that integrate optimization on interprocessor communications with the objective of minimizing the schedule length. A fine-grained integer-linear programming (ILP) model is proposed to properly address the communication latency with a network contention, which generates runtime scheduling with trivial performance difference from the predictions. We further propose a heuristic algorithm, unified priority-based scheduling (UPS), to effectively solve the contention problem in polynomial time by assigning priorities to messages. Evaluation results show that the solutions obtained by the ILP model outperform the state-of-the-art techniques by 31.1%, and UPS improves application performance by 34.7% and 44.4% compared with acquainted first-in-first-out (FIFO)-based and random-based methods. In addition, UPS achieves averagely 8.3% approximated results with the optimal solutions generated by ILP. A case study on H.264 high-definition television (HDTV) decoder and the digital signal processor (DSP) filter benchmarks achieves significant improvement on the performance and the results prediction accuracy, as well as the prominent reduction in the number of network contention and energy consumption.

Using NERSC High-Performance Computing (HPC) systems for high-energy nuclear physics applications with ALICE

High-Performance Computing Systems are powerful tools tailored to support large- scale applications that rely on low-latency inter-process communications to run efficiently. By design, these systems often impose constraints on application workflows, such as limited external network connectivity and whole node scheduling, that make more general-purpose computing tasks, such as those commonly found in high-energy nuclear physics applications, more difficult to carry out. In this work, we present a tool designed to simplify access to such complicated environments by handling the common tasks of job submission, software management, and local data management, in a framework that is easily adaptable to the specific requirements of various computing systems. The tool, initially constructed to process stand-alone ALICE simulations for detector and software development, was successfully deployed on the NERSC computing systems, Carver, Hopper and Edison, and is being configured to provide access to the next generation NERSC system, Cori. In this report, we describe the tool and discuss our experience running ALICE applications on NERSC HPC systems. The discussion will include our initial benchmarks of Cori compared to other systems and our attempts to leverage the new capabilities offered with Cori to support data-intensive applications, with a future goal of full integration of such systems into ALICE grid operations.

Hybrid shared‐memory and message‐passing multiprocessor system‐on‐chip for UWB MAC layer

Ultra-wideband (UWB) is a well-known radio technology whose media access control (MAC) protocol, WiMedia MAC, has considerable potential to ensure high-speed and high-quality data communication for wireless personal area networks. However, these benefits involve a heavy computational workload, thereby posing a challenge to the conventional very-large-scale integration (VLSI) approach in terms of providing the required performance and power efficiency. Therefore, this study aims to optimise the VLSI implementation of the WiMedia MAC system by proposing a hybrid shared-memory and message-passing multiprocessor system-on-chip (MPSoC) architecture. The proposed solution combines the state-of-the-art MPSoC technology and application-specific instruction-set processor techniques to (i) accelerate the MAC protocol at task level by using parallel processing, (ii) enable the using of custom instructions to optimise the inter-processor communication by using an explicit message passing mechanism, and (iii) ease the implementation process by using a high-level software/hardware co-design methodology. The proposed platform is implemented on both system-level SystemC for architecture exploration and standard-cell technology for future chip implementation. Experimental results show that the proposed hybrid MPSoC architecture achieves 24% performance improvement and 22% power savings over the conventional shared-memory-only one.

Monte Carlo MP2 on Many Graphical Processing Units.

In the Monte Carlo second-order many-body perturbation (MC-MP2) method, the long sum-of-product matrix expression of the MP2 energy, whose literal evaluation may be poorly scalable, is recast into a single high-dimensional integral of functions of electron pair coordinates, which is evaluated by the scalable method of Monte Carlo integration. The sampling efficiency is further accelerated by the redundant-walker algorithm, which allows a maximal reuse of electron pairs. Here, a multitude of graphical processing units (GPUs) offers a uniquely ideal platform to expose multilevel parallelism: fine-grain data-parallelism for the redundant-walker algorithm in which millions of threads compute and share orbital amplitudes on each GPU; coarse-grain instruction-parallelism for near-independent Monte Carlo integrations on many GPUs with few and infrequent interprocessor communications. While the efficiency boost by the redundant-walker algorithm on central processing units (CPUs) grows linearly with the number of electron pairs and tends to saturate when the latter exceeds the number of orbitals, on a GPU it grows quadratically before it increases linearly and then eventually saturates at a much larger number of pairs. This is because the orbital constructions are nearly perfectly parallelized on a GPU and thus completed in a near-constant time regardless of the number of pairs. In consequence, an MC-MP2/cc-pVDZ calculation of a benzene dimer is 2700 times faster on 256 GPUs (using 2048 electron pairs) than on two CPUs, each with 8 cores (which can use only up to 256 pairs effectively). We also numerically determine that the cost to achieve a given relative statistical uncertainty in an MC-MP2 energy increases as O(n3) or better with system size n, which may be compared with the O(n5) scaling of the conventional implementation of deterministic MP2. We thus establish the scalability of MC-MP2 with both system and computer sizes.

Asynchronous communication in spectral-element and discontinuous Galerkin methods for atmospheric dynamics – a case study using the High-Order Methods Modeling Environment (HOMME-homme_dg_branch)

Abstract. The scalability of computational applications on current and next-generation supercomputers is increasingly limited by the cost of inter-process communication. We implement non-blocking asynchronous communication in the High-Order Methods Modeling Environment for the time integration of the hydrostatic fluid equations using both the spectral-element and discontinuous Galerkin methods. This allows the overlap of computation with communication, effectively hiding some of the costs of communication. A novel detail about our approach is that it provides some data movement to be performed during the asynchronous communication even in the absence of other computations. This method produces significant performance and scalability gains in large-scale simulations.

Analytical modeling and feasibility study of a multi-GPU cloud-based server (MGCS) framework for non-voxel-based dose calculations.

In this paper, a multi-GPU cloud-based server (MGCS) framework is presented for dose calculations, exploring the feasibility of remote computing power for parallelization and acceleration of computationally and time intensive radiotherapy tasks in moving toward online adaptive therapies. An analytical model was developed to estimate theoretical MGCS performance acceleration and intelligently determine workload distribution. Numerical studies were performed with a computing setup of 14 GPUs distributed over 4 servers interconnected by a 1 Gigabits per second (Gbps) network. Inter-process communication methods were optimized to facilitate resource distribution and minimize data transfers over the server interconnect. The analytically predicted computation time predicted matched experimentally observations within 1-5%. MGCS performance approached a theoretical limit of acceleration proportional to the number of GPUs utilized when computational tasks far outweighed memory operations. The MGCS implementation reproduced ground-truth dose computations with negligible differences, by distributing the work among several processes and implemented optimization strategies. The results showed that a cloud-based computation engine was a feasible solution for enabling clinics to make use of fast dose calculations for advanced treatment planning and adaptive radiotherapy. The cloud-based system was able to exceed the performance of a local machine even for optimized calculations, and provided significant acceleration for computationally intensive tasks. Such a framework can provide access to advanced technology and computational methods to many clinics, providing an avenue for standardization across institutions without the requirements of purchasing, maintaining, and continually updating hardware.

TripleMon: A multi-layer security framework for mediating inter-process communication on Android

As smartphones have become an indispensable part of daily life, mobile users are increasingly relying on them to process personal information with feature-rich applications. This situation requires robust security mechanisms for protecting sensitive applications and data on mobile devices. Android, as one the most popular smartphone operating systems, provides two core security mechanisms, application sandboxing and a permission system. However, recent studies show that these mech- anisms are vulnerable to be passed by a variety of attacks. In this paper, we argue for the need of designing and implementing more comprehensive security mechanisms for Android. We realize that mediating Inter-Process Communication (IPC) channels used by Android applications can mitigate prominent attacks effectively and efficiently. Based on this observation, we propose a practical multi-layer security framework called TRIPLEMON to support policy-based mediation on Android IPC. We also dis- cuss and evaluate a proof-of-concept prototype of TRIPLEMON along with the experimental results derived from real malware samples and synthetic attacks.

Segmenting large traces of inter-process communication with a focus on high performance computing systems

The understanding of the interactions among processes of a High Performance Computing (HPC) system can be made easier if trace analysis is used. Traces, however, can be quite large, making it difficult to analyze their content unless some abstraction is provided. This paper presents a novel trace abstraction approach that aims to facilitate the analysis of large execution traces generated from HPC applications. Our approach allows automatic segmentation of large traces into smaller and meaningful clusters that reflect the various execution phases of the traced scenarios. Our approach is based on the application of information theory principles to the analysis of sequences of communication patterns extracted from traces of HPC systems. This work is inspired by recent studies in the field of bioinformatics where several techniques have been proposed to segment DNA sequences into homogeneous sub-domains, where each sub-domain exhibits a certain degree of internal homogeneity. Trace segments can be used in a number of applications such as recovering high-level views of the system behavior and program understanding. We demonstrate the usefulness of our approach by applying it to different traces of hundreds of millions of events, generated from two HPC systems.

A Flexible System Architecture for Acquisition and Storage of Naturalistic Driving Data

Innovation in intelligent transportation systems relies on analysis of high-quality data. In this paper, we describe the design principles behind our data management infrastructure. The principles we adopt place an emphasis on flexibility and maintainability. This is achieved by breaking up code into a modular design that can be run on many independent processes. Message passing over a publish–subscribe network enables interprocess communication and promotes data-driven execution. By following these principles, rapid prototyping and experimentation with new sensing modalities and algorithms are possible. The communication library underpinning our proposed architecture is compared against several popular communication libraries. Features designed into the system make it decentralized, robust to failure, and amenable to scaling across multiple machines with minimal configuration. Code written using the proposed architecture is compact, transparent, and easy to maintain. Experimentation shows that our proposed architecture offers a high performance when compared against alternative communication libraries.

PSIN: A scalable, Parallel algorithm for Seismic INterferometry of large-N ambient-noise data

Seismic interferometry is a technique for extracting deterministic signals (i.e., ambient-noise Green's functions) from recordings of ambient-noise wavefields through cross-correlation and other related signal processing techniques. The extracted ambient-noise Green's functions can be used in ambient-noise tomography for constructing seismic structure models of the Earth's interior. The amount of calculations involved in the seismic interferometry procedure can be significant, especially for ambient-noise datasets collected by large seismic sensor arrays (i.e., “large-N” data). We present an efficient parallel algorithm, named pSIN (Parallel Seismic INterferometry), for solving seismic interferometry problems on conventional distributed-memory computer clusters. The design of the algorithm is based on a two-dimensional partition of the ambient-noise data recorded by a seismic sensor array. We pay special attention to the balance of the computational load, inter-process communication overhead and memory usage across all MPI processes and we minimize the total number of I/O operations. We have tested the algorithm using a real ambient-noise dataset and obtained a significant amount of savings in processing time. Scaling tests have shown excellent strong scalability from 80 cores to over 2000 cores.