Single Processor Research Articles

Experiences of Civil Certification of Multi-Core Processing Systems in Commercial and Military Avionics, Integration Activities, and Analysis

<div class="section abstract"><div class="htmlview paragraph">Avionics systems are currently undergoing a transition from single core processor architectures to multi-core processor architectures. This transition enables significant advantages in reduction in size, weight, power (SWaP) and cost. However, avionics hardware and software certification policies and guidance are evolving as research and experience is gained with multi-core processor architectures. The unique challenges of using multi-core processors in certified avionics will be discussed. The requirements for a virtualization platform supporting multiple real-time operating system (RTOS) partitions on a multi-core processor used in safety-critical avionics systems are defined, including the ability to support multiple design assurance levels (DAL) on multiple cores, fault isolation and containment, static configuration as per ARINC 653, role-based development as per DO-297, and robust partitioning to reduce cost of incremental certification. The paper will present a collaborative approach undertaken by a leading avionics system supplier and a leading safety-critical commercial-off-the-shelf (COTS) RTOS supplier in the development of a multi-core real-time system with DO-178C DAL A software and DO-254 DAL A hardware safety certification on an FAA Program of Record (PoR). The approach taken to comply with FAA CAST-32A objectives will be presented. Particular focus is provided for integration activities and program specific analysis performed by the IMA application developer and integrator to guarantee determinism in the deployed system. Using the approach defined under the PoR, the application developer performs activities including foot-printing under worst-case execution time (WCET) loads and application of numerical methods to predict interference effects. The IMA integrator uses this data to define a performance restricted environment (PRE) and uses WCET verification in the PRE. Tools, analysis methods, and sample results will be presented. The method to capture results is discussed. Finally the paper includes lessons learned during the program.</div></div>

Feasibility analysis of periodic real-time systems with energy harvesting capabilities

Energy harvesting now appears as a feasible solution to power wireless devices. Such ones including sensor nodes have to be energy neutral i.e. be able to never consume more energy than available. Energy can be drawn from many ambient sources including solar, vibration, etc. Most of wireless devices perform periodic sampling and computations using a microprocessor for real-time monitoring. The software tasks that typically take place in these systems are time critical and cyclic with specified deadlines attached to finishing times. In this paper, we explore the problem of feasibility analysis of hard real-time periodic task sets on a single processor powered by renewable energy. Firstly, we investigate the notion of robustness as the capability of the system to remain stable even in case of positive changes in its operating environment. We prove that our feasibility test is robust with respect to the profile of the energy source. Secondly, we prove that for asynchronous tasks whose initial release times are not known in advance, feasibility can be checked considering the synchronous activation pattern which corresponds to the worst-case scenario. Finally, whenever accurate prediction of the incoming environmental energy is impossible, a sufficient feasibility test with high implementation efficiency is proposed.

Priority-based scheduling of mixed-critical jobs

Modern real-time systems tend to be mixed-critical, in the sense that they integrate on the same computational platform applications at different levels of criticality (e.g., safety critical and mission critical). Scheduling of such systems is a popular topic in literature due to the complexity and importance of the problem. In this paper we propose two algorithms for job scheduling in mixed critical systems: mixed criticality earliest deadline first (MCEDF) and mixed critical priority improvement (MCPI). MCEDF is a single processor algorithm that theoretically dominates state-of-the-art fixed-priority algorithm own criticality based priority (OCBP), while having a better computational complexity. The dominance is achieved by profiting from a common extension of fixed-priority online policy to mixed criticality. MCPI is a multiprocessor algorithm that supports dependency constraints. Experiments show good schedulability results. Also we formally prove that both MCEDF and MCPI are optimal in a particular class of algorithms.

Branch and Bound Algorithm for the Single Machine Scheduling Problem with Release and Delivery times

The single machine scheduling problem is one of the classic NP-hard optimization problems, and it is useful in solving flowshop and jobshop scheduling problems. The goal of this paper is to prepare algorithms for the scheduling problem, where set of tasks is performed on a single processor. Each task has a release time when it becomes available for processing, a processing time and a delivery time. At most one job can be processed at a time, but all jobs may be simultaneously delivered. Preemption is not allowed. The objective is to minimize the time, by which all tasks are delivered. We compare the nondelay schedule, in which no processor is kept idle at a time when it could begin processing a task and an inserted idle time schedule, in which a processor is kept idle at this time. In this paper, we propose an approximation algorithm and by combining this algorithm and branch and bound method, we develop branch and bound algorithm, which can find optimal solutions for the single processor scheduling problem. We use a binary branching rule, where at each branch node, a complete schedule is generated. To illustrate the effectiveness of our algorithms we tested them on randomly generated set of tasks.

Reduced-Scaling Approach for Configuration Interaction Singles and Time-Dependent Density Functional Theory Calculations Using Hybrid Functionals.

An approximation is presented which can efficiently decrease the computational expenses of configuration interaction singles (CIS) and time-dependent density functional theory (TDDFT) methods employing hybrid functionals. The approach is the adaptation of the local density fitting scheme developed for Hartree-Fock (HF) calculations for excited states and reduces the quartic scaling of the methods to cubic. It can also be applied to related methods, such as the time-dependent HF and Tamm-Dancoff approximation TDDFT approaches. Our benchmark calculations show that, for molecules of 50-100 atoms, average speedups of 2-4 can be achieved for CIS and TDDFT calculations at the expense of negligible errors in the calculated excitation energies and oscillator strengths, but for bigger systems or molecules of localized electronic structure significantly larger speedups can be gained. We also demonstrate that the approximation enables excited-state calculations on a single processor even for molecules of 1000 atoms using basis sets augmented with diffuse functions including more than 17 000 atomic orbitals.

Integrated microwave photonics

Recent advances in photonic integration have propelled microwave photonic technologies to new heights. The ability to interface hybrid material platforms to enhance light–matter interactions has led to the development of ultra-small and high-bandwidth electro-optic modulators, low-noise frequency synthesizers and chip signal processors with orders-of-magnitude enhanced spectral resolution. On the other hand, the maturity of high-volume semiconductor processing has finally enabled the complete integration of light sources, modulators and detectors in a single microwave photonic processor chip and has ushered the creation of a complex signal processor with multifunctionality and reconfigurability similar to electronic devices. Here, we review these recent advances and discuss the impact of these new frontiers for short- and long-term applications in communications and information processing. We also take a look at the future perspectives at the intersection of integrated microwave photonics and other fields including quantum and neuromorphic photonics. This Review discusses recent advances of microwave photonic technologies and their applications in communications and information processing, as well as their potential implementations in quantum and neuromorphic photonics.

Fast and Flexible Software Polar List Decoders

Flexibility is one mandatory aspect of channel coding in modern wireless communication systems. Among other things, the channel decoder has to support several code lengths and code rates. This need for flexibility applies to polar codes that are considered for control channels in the future 5G standard. This paper presents a new generic and flexible implementation of a software Successive Cancellation List (SCL) decoder. A large set of parameters can be fine-tuned dynamically without re-compiling the software source code: the code length, the code rate, the frozen bits set, the puncturing patterns, the cyclic redundancy check, the list size, the type of decoding algorithm, the tree-pruning strategy and the data quantization. This generic and flexible SCL decoder enables to explore tradeoffs between throughput, latency and decoding performance. Several optimizations are proposed to achieve a competitive decoding speed despite the constraints induced by the genericity and the flexibility. The resulting polar list decoder is about 4 times faster than a generic software decoder and only 2 times slower than a non-flexible unrolled decoder. Thanks to the flexibility of the decoder, the fully adaptive SCL algorithm can be easily implemented and achieves higher throughput than any other similar decoder in the literature (up to 425 Mb/s on a single processor core for N = 2048 and K = 1723 at 4.5 dB).

Parallel Multidimensional Lookahead Sorting Algorithm

This paper presents a new parallel structured lookahead multidimensional sorting algorithm. Our algorithm can be based on any sequential sorting algorithm. The amount of parallelism can be controlled using several parameters such as the number of threads, word size, memory/processor communication overhead, and the dimension of the algorithm. The proposed technique is ideally suited for general purpose graphic processing units and shared-memory massively parallel processor systems. It ensures that data being processed exhibits temporal and spatial locality to maximize the utilization of processor cache. The algorithm achieves a speedup even when a single processor is used. A lookahead algorithm is also proposed to achieve even higher speedup. The performance of the proposed algorithm is verified numerically and experimentally.

Adaptive Hierarchical Scheduling Framework for TiRTOS

The complexity of embedded real-time systems has increased, and most applications have large diversity in execution times of their tasks. Therefore, most traditional scheduling techniques do not satisfy requirements of such applications. This article proposes an adaptive hierarchical scheduling framework for a set of independent concurrent applications composing of soft and hard real time tasks, that run on a single processor. It ensures temporal partitioning between independent applications with budget adaption feature, where CPU time of each application is periodically and dynamically assigned. Implemented in the kernel of TI-RTOS on a resource constrained platform, experiments show that proposed scheme provides good performance for multiple applications with dynamic tasks under overload conditions. Compared with traditional priority scheduler originally implemented in TI-RTOS and EDF scheduler, it achieves low miss ratio with minimal overhead while yielding temporal partitioning.

AES Overhead Mitigation Using OpenMP

The Advanced Encryption Standard (AES) has been extensively used in many systems to provide security. However, many of the implementations are susceptible to Side Channel Attacks (SCA). To address the susceptibility, several researchers have proposed a number of countermeasures. Masking scheme is one of the countermeasures that are commonly used to counteract such attacks. In this paper, we investigate the overhead of the countermeasure in terms of execution time of the first order-masking scheme for the AES “Encryption part only”. To achieve that, we have implemented the sequential algorithm of the AES with single processor, and after that used OpenMP directives to reduce the overhead of the countermeasure. Subsequently, the result of the sequential algorithm has been compared with its parallel implementations using 2, 4, 6, and 8 processors. We show how parallel implementation of the AES offers better performance yet flexible enough for cryptographic algorithms. Compared to the sequential unmasked AES implementation, the best masking scheme for the first order using data parallelism shows a performance in terms of speed up around 5x when 8 threads are used.

The Generation and Forecast of Extreme Winds during the Origin and Progression of the 2017 Tubbs Fire

On 8–9 October 2017, fourteen wildfires developed rapidly during a strong Diablo wind event in northern California including the Tubbs Fire, which travelled over 19 km in 3.25 h. Here, we applied the CAWFE® coupled numerical weather prediction-fire modeling system to investigate the airflow regime and extreme wind peaks underlying the extreme fire behavior using simulations that refine from a 10 km to a 185 m horizontal grid spacing. We found that as Diablo winds travelled south down the Sacramento Valley and fanned out southwestward over the Wine Country, their strength waxed and waned and their direction wavered, creating varying locations near fire origins where wind overrunning topography reached 30–40 m/s, along with streaks and bursts of strong winds in the lee of some topographic features and stagnation downstream of others. Despite a statically stable layer in the lowest 1.5 km, the high Froude number flow sometimes resembled a hydraulic jump. Elsewhere, the flow behaved similarly to neutrally-stratified flow over small hills, creating wind extrema that exceeded 40 m/s at the crest of some lesser hills including near the Tubbs fire ignition, but which shed bursts of high speed winds that travel downstream at approximately 5–7-min intervals. Nonetheless, simulated fire growth lagged satellite detection of fire arrival in Santa Rosa by up to 1 h, although whether the data detect fire line or spotting is ambiguous. A forecast simulation with a 370 m horizontal grid spacing produced an on-time fire line arrival in Santa Rosa, with calculations executed 4 times faster than real time on a single computer processor.

GenMatcher

Packet classification methods rely upon packet content/header matching against rules. Thus, throughput of matching operations is critical in many networking applications. Further, with the advent of Software Defined Networking (SDN), efficient implementation of software approaches to matching are critical for the overall system performance. This article presents 1 GenMatcher, a generic, software-only, arbitrary matching framework for fast, efficient searches. The key idea of our approach is to represent arbitrary rules with efficient prefix-based tries. To support arbitrary wildcards, we rearrange bits within the rules such that wildcards accumulate to one side of the bitstring. Since many non-contiguous wildcards often remain, we use multiple prefix-based tries. The main challenge in this context is to generate efficient trie groupings and expansions to support all arbitrary rules. Finding an optimal mix of grouping and expansion is an NP-complete problem. Our contribution includes a novel, clustering-based grouping algorithm to group rules based upon their bit-level similarities. Our algorithm generates near-optimal trie groupings with low configuration times and provides significantly higher match throughput compared to prior techniques. Experiments with synthetic traffic show that our method can achieve a 58.9X speedup compared to the baseline on a single core processor under a given memory constraint.

On a certain class of power- and energy-related scheduling problems

We consider problems of power and energy allocation to computational jobs performed on identical processors, where jobs may be performed with different speeds. Each job requires for its execution a processor as well as an amount (unknown in advance) of electrical power. Power is a continuous, doubly-constrained resource whose consumption over time is represented by energy. The amount of power allocated to a job is nonlinearly related to a particular speed of the job. Jobs are preemptable and independent, and all are available at the start of the process. Two general types of scheduling problems are considered: laptop problems, where the objective is to minimize the schedule length subject to power and/or energy constraints, and server problems, where power usage or energy consumption is to be minimized subject to a given schedule deadline. We present formulations of the analyzed problems, as well as methodologies for solving them for the cases of parallel processors and a single processor.

A New Parallel Intel Xeon Phi Hydrodynamics Code for Massively Parallel Supercomputers

In this paper, a new hydrodynamics code called gooPhi to simulate astrophysical flows on modern Intel Xeon Phi processors with KNL architecture is presented. A new vector numerical method implemented in the form of a program code for massively parallel architectures is proposed. A detailed description is given and a parallel implementation of the code is made. A performance of 173 gigaflops and 48 speedup are obtained on a single Intel Xeon Phi processor. A 97 per cent scalability is reached with 16 processors.

Langevin dynamics simulation with dipole–dipole interactions: Massive performance improvements and advanced analytical integrator

The required computation duration is a crucial element for the computer simulation of both molecular and suspended nanoparticle dynamics. The latter one is already highly-optimized on a theoretical level by the implicit solvent notion within the Langevin and Brownian Dynamics techniques which still requires a more clear theoretical connection though. Also, the calculation performance issue becomes much more serious in case of the long-range interaction within the magnetostatics and/or electrostatics problems. Hence, both conceptual aspects and a technical challenge require a resolution for the real experimental applications. In order to do so, the analytical fluctuative–dissipative integrator has been derived from basic theoretical foundation. A numerical simulation complexity is partly reduced by the analytical solution. The method is reinforced by the Barnes–Hut general-purpose graphics processing unit (GPU) based dipole–dipole interaction algorithm. As a real application close to an experiment, the Langevin dynamics simulation of up to 1 million ferrofluid nanoparticles evolution have been performed at the all-with-all dipole–dipole long-range interactions without any cutoff radius and with real coefficients of translational and rotational viscous friction. Computational time scales comparisons of different methods have been made: 100 ns in silico could take between minutes till years of a computation Unix epoch time depending on a method. The 850000 times calculation duration speed up from slowest (well-known) towards fastest (suggested here) simulation method has been achieved. Due to a GPU parallel computation the required calculation time scales up proportionally to a number of nanoparticles N which is much better compared to N*N or N*log(N) in case of a direct dipole summation or Barnes–Hut on the single processor method, respectively. The overall implementation is available and ready to be used within the wider open-source software package.

Development of a framework for parallel reservoir simulation

The development of a parallel reservoir simulator is a more complex task than nonparallel simulators. Problems related to parallel implementation such as parallel communication, model division among processors, and the management of data distributed among processors, and other issues should be addressed and solved in addition to the already complex task of reservoir simulator development. Hence, the development of parallel reservoir simulators is more time intensive than the traditional development on single processor computers. In this work, a framework was developed to aid the development of parallel reservoir simulators. The framework developed in this work implements and handles the parallel processing complexity and provides easy-to-use programming interfaces to accelerate the development of new parallel reservoir simulators or the parallelization of existing ones. Additionally, The University of Texas Compositional Simulator was used in conjunction with the framework developed in this work to create a parallel reservoir simulator. Case studies are used to verify accuracy and to test parallel performance of our new parallel reservoir simulator.

A Monte Carlo volumetric-ray-casting estimator for global fluence tallies on GPUs

A Monte Carlo fluence estimator has been designed to take advantage of the computational power of graphical processing units (GPUs). This new estimator, termed the volumetric-ray-casting estimator, is an extension of the expectation estimator. It can be used as a replacement of the track-length estimator for the estimation of global fluence. Calculations for this estimator are performed on the GPU while the Monte Carlo random walk is performed on the central processing unit (CPU). This method lowers the implementation cost for GPU acceleration of existing Monte Carlo particle transport codes as there is little modification of the particle history logic flow. Three test problems have been evaluated to assess the performance of the volumetric-ray-casting estimator for neutron transport on GPU hardware in comparison to the standard track-length estimator on CPU hardware. Evaluation of neutron transport through air in a criticality accident scenario showed that the volumetric-ray-casting estimator achieved 23 times the performance of the track-length estimator using a single core CPU paired with a GPU and 15 times the performance of the track-length estimator using an eight core CPU paired with a GPU. Simulation of a pressurized water reactor fuel assembly showed that the performance improvement was 6 times within the fuel and 7 times within the control rods using an eight core CPU paired with a single GPU.

Application of Parallel Vector Space Model for Large-Scale DNA Sequence Analysis

With rapid advancement in the field of bioinformatics and computational biology, the collected DNA dataset is growing exponentially, doubling after every 18 months. Due to large-scale and complex structure of the DNA dataset, the analysis of DNA sequence is becoming computationally a challenging issue in bioinformatics field and computational biology. Fast algorithms, capable of analyzing large-scale DNA sequence, are now required in the field of bioinformatics. This paper presents a novel Parallel Vector Space Model (PVSM) approach that supports the analysis of large-scale DNA sequence by taking advantages of multi-core system. The proposed approach is built on top of modified Vector Space Model (VSM). In order to evaluate the performance of PVSM, the proposed technique is extensively evaluated using varied size of DNA sequences in the context of computational efficiency and accuracy. The performance of PVSM is compared with sequential modified VSM. The sequential VSM is implemented on a single processor whereas, the proposed method is initially parallelized on 4 processors and subsequently on 12 processors. The experimental results show that the PVSM performed better than the sequential VSM. The proposed method achieved approximately 2× speedup compared with sequential approach, without affecting the accuracy level. Moreover, the proposed PVSM is highly scalable with an increase in the number of processing cores and support the analysis of large-scale DNA sequences.

Zero cross algorithm performance on raspberry pi machine for ECG QRS detection

The electrocardiogram (ECG) is one of the most popular instruments in the medical field. Have been developed a smart IoT system for ECG Arrhythmia classification using Android device as the IoT Hub. In order to minimize the computing load on the Android device, some parts of the signal processing are moved to the data acquisition module by using Raspberry Pi 3 as the processor. In the Raspberry machine, a QRS detection system with Zero Cross Algorithm was implemented and its performance was measured. To measure the performance of the algorithm, a timer was placed before and after the execution of the process. The result shows that the algorithm can be run in real-time use in Raspberry Pi 3, although it is a bit slow and produced a high computational load of a single core processor in the machine.

The single-processor scheduling problem with time restrictions: complexity and related problems

We consider the single-processor scheduling problem with time restrictions (STR) in order to minimize the makespan, where a set of independent jobs has to be processed on a single processor, subject only to the following constraint: During any time period of length $$\alpha $$ , the number of jobs being executed is less than or equal to a given integer value B. First, we prove that the STR problem is binary NP-hard even for $$B=2$$ by pointing out an interesting analogy of this problem to the parallel machine scheduling problem with a single server with equal processing times. Next, we give a formal description of the STR problem by providing two mixed integer programming formulations to solve it optimally. The first is based on time-indexed variables and the second is based on assignment and positional date variables (APV). Both models were tested on randomly generated instances, and the experimental results show that the APV model performs very well and can solve problem instances of up to $$n=500$$ jobs.