FPGA Prototyping of Emerging Manycore Architectures for Parallel Programming Research using Formic Boards

Performance evaluation of parallel software and architectural exploration of innovative hardware support face a common challenge with emerging manycore platforms: they are limited by the slow running time and the low accuracy of software simulators. Manycore FPGA prototypes are difficult to build, but they offer great rewards. Software running on such prototypes runs orders of magnitude faster than current simulators. Moreover, researchers gain significant architectural insight during the modeling process. We use the Formic FPGA prototyping board [1], which specifically targets scalable and cost-efficient multi-board prototyping, to build and test a 64-board model of a 512-core, MicroBlaze-based, non-coherent hardware prototype with a full network-on-chip in a 3D-mesh topology. We expand the hardware architecture to include the ARM Versatile Express platforms and build a 520-core heterogeneous prototype of 8 Cortex-A9 cores and 512 MicroBlaze cores. We then develop an MPI library for the prototype and evaluate it extensively using several bare-metal and MPI benchmarks. We find that our processor prototype is highly scalable, models faithfully single-chip multicore architectures, and is a very efficient platform for parallel programming research, being 50,000 times faster than software simulation.

An accelerated method of developing DBT projects using CDIO partner institutions as parallel processors

The cycle of the academic year impacts on efforts to refine and improve major group design-build-test (DBT) projects since the time to run and evaluate projects is generally a full calendar year. By definition these major projects have a high degree of complexity since they act as the vehicle for the application of a range of technical knowledge and skills. There is also often an extensive list of desired learning outcomes which extends to include professional skills and attributes such as communication and team working. It is contended that student project definition and operation, like any other designed product, requires a number of iterations to achieve optimisation. The problem however is that if this cycle takes four or more years then by the time a project’s operational structure is fine tuned it is quite possible that the project theme is no longer relevant. The majority of the students will also inevitably experience a sub-optimal project experience over the 5 year development period. It would be much better if the ratio were flipped so that in 1 year an optimised project definition could be achieved which had sufficient longevity that it could run in the same efficient manner for 4 further years. An increased number of parallel investigators would also enable more varied and adventurous project concepts to be examined than a single institution could undertake alone in the same time frame.
This work-in-progress paper describes a parallel processing methodology for the accelerated definition of new student DBT project concepts. This methodology has been devised and implemented by a number of CDIO partner institutions in the UK & Ireland region. An agreed project theme was operated in parallel in one academic year with the objective of replacing a multi-year iterative cycle. Additionally the close collaboration and peer learning derived from the interaction between the coordinating academics facilitated the development of faculty teaching skills in line with CDIO standard 10.

TProf: An energy profiler for task-parallel programs

We present TProf, an energy profiling tool for OpenMP-like task-parallel programs. To compute the energy consumed by each task in a parallel application, TProf dynamically traces the parallel execution and uses a novel technique to estimate the per-task energy consumption. To achieve this estimation, TProf apportions the total processor energy among cores and overcomes the limitation of current works which would otherwise make parallel accounting impossible to achieve. We demonstrate the value of TProf by characterizing a set of task parallel programs, where we find that data locality, memory access patterns and task working sets are responsible for significant variance in energy consumption between seemingly homogeneous tasks. In addition, we identify opportunities for fine-grain energy optimization by applying per-task Dynamic Voltage and Frequency Scaling (DVFS).

Self-Feedback Motion Control for Cable-Driven Parallel Manipulators

This article proposes a closed-loop control scheme based on joint-angle feedback for cable-driven parallel manipulators (CDPMs), which is able to overcome various difficulties resulting from the flexible nature of the driven cables to achieve higher control accuracy. By introducing a unique structure design that accommodates built-in encoders in passive joints, the seven degrees of freedom (7-DOF) CDPM can obtain joint angle values without external sensing devices, and it is used for feedback control together with a proper closed-loop control algorithm. The control algorithm has been derived from the time differential of the kinematic formulation, which relates the joint angular velocities to the time derivative of cable lengths. In addition, the Lyapunov stability theory and Monte Carlo method have been used to mathematically verify the self-feedback control law that has tolerance for parameter errors. With the aid of co-simulation technique, the self-feedback closed-loop control is applied on a 7-DOF CDPM and it shows higher motion accuracy than the one with an open-loop control. The trajectory tracking experiment on the motion control of the 7-DOF CDPM demonstrated a good performance of the self-feedback control method.

Type Synthesis of 3-DOF Multi-Mode Translational/Spherical Parallel Mechanisms with Lockable Joints

A 3-DOF (degrees-of-freedom) multi-mode translational/spherical PM (parallel mechanism) with lockable joints is a novel reconfigurable PM. It has both 3-DOF spatial translational operation mode and 3-DOF spherical operation mode. This paper presents an approach to the type synthesis of translational/spherical PMs with lockable joints. Using the proposed approach, several 3-DOF translational/spherical PMs are obtained. It is found that these translational/spherical PMs do not encounter constraint singular configurations and self-motion of sub-chain of a leg during reconfiguration. The approach can also be used for synthesizing other classes of multi-mode PMs with lockable joints, multi-mode PMs with variable kinematic joints, partially decoupled PMs, and reconfigurable PMs with a reconfigurable platform.

Parallel Robot

Parallel robot (PR) is a mechanical system that utilized multiple computer-controlled limbs to support one common platform or end effector. Comparing to a serial robot, a PR generally has higher precision and dynamic performance and, therefore, can be applied to many applications. The PR research has attracted a lot of attention in the last three decades, but there are still many challenging issues to be solved before achieving PRs’ full potential. This chapter introduces the state-of-the-art PRs in the aspects of synthesis, design, analysis, and control. The future directions will also be discussed at the end.

Topology synthesis of three-legged spherical parallel manipulators employing Lie group theory

Driven by the requirements of the bionic joint or tracking equipment for the spherical parallel manipulators (SPMs) with three rotational degrees-of-freedom (DoFs), this paper carries out the topology synthesis of a class of three-legged SPMs employing Lie group theory. In order to achieve the intersection of the displacement subgroups, the subgroup characteristics and operation principles are defined in this paper. Mainly drawing on the Lie group theory, the topology synthesis procedure of three-legged SPMs including four stages and two functional blocks is proposed, in which the assembly principles of three legs are defined. By introducing the circular track, a novel class of three-legged SPMs is synthesized, which is the important complement to the existing SPMs. Finally, four typical examples are given to demonstrate the finite displacements of the synthesized three-legged SPMs.

Whole-mitochondrial genome sequencing in primary open-angle glaucoma using massively parallel sequencing identifies novel and known pathogenic variants

Purpose:The aim of this study was to determine whether mutations in mitochondrial DNA play a role in high-pressure primary open-angle glaucoma (OMIM 137760) by analyzing new data from massively parallel sequencing of mitochondrial DNA.
Methods:Glaucoma patients with high-tension primary open-angle glaucoma and ethnically matched and age-matched control subjects without glaucoma were recruited. The entire human mitochondrial genome was amplified in two overlapping fragments by long-range polymerase chain reaction and used as a template for massively parallel sequencing on an Ion Torrent Personal Genome Machine. All variants were confirmed by conventional Sanger sequencing.
Results:Whole-mitochondrial genome sequencing was performed in 32 patients with primary open-angle glaucoma from India (n = 16) and Ireland (n = 16). In 16 of the 32 patients with primary open-angle glaucoma (50% of cases), there were 22 mitochondrial DNA mutations consisting of 7 novel mutations and 8 previously reported disease-associated sequence variants. Eight of 22 (36.4%) of the mitochondrial DNA mutations were in complex I mitochondrial genes.
Conclusion:Massively parallel sequencing using the Ion Torrent Personal Genome Machine with confirmation by Sanger sequencing detected a pathogenic mitochondrial DNA mutation in 50% of the primary open-angle glaucoma cohort. Our findings support the emerging concept that mitochondrial dysfunction results in the development of glaucoma and, more specifically, that complex I defects play a significant role in primary open-angle glaucoma pathogenesis.

Android Malware Detection Using Parallel Machine Learning Classifiers

Mobile malware has continued to grow at an alarming rate despite on-going mitigation efforts. This has been much more prevalent on Android due to being an open platform that is rapidly overtaking other competing platforms in the mobile smart devices market. Recently, a new generation of Android malware families has emerged with advanced evasion capabilities which make them much more difficult to detect using conventional methods. This paper proposes and investigates a parallel machine learning based classification approach for early detection of Android malware. Using real malware samples and benign applications, a composite classification model is developed from parallel combination of heterogeneous classifiers. The empirical evaluation of the model under different combination schemes demonstrates its efficacy and potential to improve detection accuracy. More importantly, by utilizing several classifiers with diverse characteristics, their strengths can be harnessed not only for enhanced Android malware detection but also quicker white box analysis by means of the more interpretable constituent classifiers.

Design patterns percolating to parallel programming framework implementation

Structured parallel programming is recognised as a viable and effective means of tackling parallel programming problems. Recently, a set of simple and powerful parallel building blocks RISC pb²l) has been proposed to support modelling and implementation of parallel frameworks. In this work we demonstrate how that same parallel building block set may be used to model both general purpose parallel programming abstractions, not usually listed in classical skeleton sets, and more specialized domain specific parallel patterns. We show how an implementation of RISC pb² l can be realised via the FastFlow framework and present experimental evidence of the feasibility and efficiency of the approach.