Memory Allocation Research Articles

An Approach in Big Data Analytics to improve the velocity of unstructured data Using Map Reduce

Big Data Analytics is an innovative approach for extracting the data from a huge volume of data warehouse systems. It reveals the method to compress the high volume of data into clusters by MapReduce and HDFS. However, the data processing has taken more time for extract and store in Hadoop clusters. The proposed system deals with the challenges of time delay in shuffle phase of map-reduce due to scheduling and sequencing. For improving the speed of big data, this proposed work using the Compressed Elastic Search Index (CESI) and MapReduce-Based Next Generation Sequencing Approach (MRBNGSA). This approach helps to increase the speed of data retrieval from HDFS clusters because of the way it is stored in that. this method is stored only the metadata in HDFS which takes less memory during runtime compare to big data due to the volume of data stored in HDFS. This approach is reduces the CPU utilization and memory allocation of the resource manager in Hadoop Framework and imroves data processing speed, such a way that time delay has to be reduced with minimum latency.

An Approach in Big Data Analytics to Improve the Velocity of Unstructured Data Using MapReduce

Big Data Analytics is an innovative approach for extracting the data from a huge volume of data warehouse systems. It reveals the method to compress the high volume of data into clusters by MapReduce and HDFS. However, the data processing has taken more time for extract and store in Hadoop clusters. The proposed system deals with the challenges of time delay in shuffle phase of map-reduce due to scheduling and sequencing. For improving the speed of big data, this proposed work using the Compressed Elastic Search Index (CESI) and MapReduce-Based Next Generation Sequencing Approach (MRBNGSA). This approach helps to increase the speed of data retrieval from HDFS clusters because of the way it is stored in that. this method is stored only the metadata in HDFS which takes less memory during runtime compare to big data due to the volume of data stored in HDFS. This approach is reduces the CPU utilization and memory allocation of the resource manager in Hadoop Framework and imroves data processing speed, such a way that time delay has to be reduced with minimum latency.

Simultaneous Control of Terrain Adaptation and Wheel Speed Allocation for a Planetary Rover With an Active Suspension System

Active suspensions are important features of many recent planetary rovers. For such a rover, the control strategy is crucial to its performance. This letter presents the control method for a planetary rover equipped with an active suspension system. The control algorithm is based on the estimation of terrain geometry and a mapping relation among the motions of rover body, suspension, and wheels. It realizes a simultaneous control of terrain adaptation and wheel speed allocation. The algorithm allows the posture and height of the rover body to be controlled and the slippage of driving wheels to be reduced while the rover traverse uneven, unstructured terrain. Another advantage of the proposed method is that the control system operates only based on proprioceptive sensors with improved efficiency and moderate memory allocation. The proposed control algorithm is applied to a rover prototype. The experiments conducted in a testing field of rovers validate the proposed algorithm and system.

SquadSmart — Hierarchical Planning and Coordinated Plan Execution for Squads of Characters

This paper presents an application of Hierarchical Task Network (HTN) planning to a squad-based military simulation. The hierarchical planner produces collaborative plans for the whole squad in real time, generating the type of highly coordinated behaviours typical for armed combat situations involving trained professionals. Here, we detail the extensions to HTN planning necessary to provide real-time planning and subsequent collaborative plan execution. To make full hierarchical planning feasible in a game context we employ a planner compilation technique that saves memory allocations and speeds up symbol access. Additionally, our planner can be paused and resumed, making it possible to impose a hard limit on its computation time during any single frame. For collaborative plan execution we describe several synchronization extensions to the HTN framework, allowing agents to participate in several plans at once and to act in parallel or in sequence during single plans. Overall, we demonstrate that HTN planning can be used as an expressive and powerful real-time planning framework for tightly coupled groups of in-game characters.

Kernel-Based Sampling of Arbitrary Signals

Point sampling is widely used in several Computer Graphics’ applications, such as point-based modelling and rendering, image and geometry processing. Starting from the kernel-based sampling, which approximates an input signal on a regular grid as the sum of Gaussian kernels, we introduce a set of additional variables that control the kernels’ width and height. These additional variables allow us to improve the quality of the distribution of the samples, and to achieve a higher approximation accuracy and a more accurate feature preservation, with a slightly higher computational cost. To further improve the sampling with respect to the input data, we introduce a sampling initialisation for processing high resolution signals, without incurring in limits for memory allocation, and a sampling optimisation, which adaptively selects the number and location of the samples to achieve the target approximation accuracy, without oversampling the input signal. To show the generality of the proposed approach for unstructured data of arbitrary dimension, we apply our kernel-based sampling to different types of data, such as 2D images, solutions to PDEs on arbitrary domains, and vector fields.

Multi-GPU systems and Unified Virtual Memory for scientific applications: The case of the NAS multi-zone parallel benchmarks

GPU-based computing systems have become a widely accepted solution for the high-performance-computing (HPC) domain. GPUs have shown highly competitive performance-per-watt ratios and can exploit an astonishing level of parallelism. However, exploiting the peak performance of such devices is a challenge, mainly due to the combination of two essential aspects of multi-GPU execution: memory allocation and work distribution. Memory allocation determines the data mapping to GPUs, and therefore conditions all work distribution schemes and communication phases in the application. Unified Virtual Memory simplifies the codification of memory allocations, but its effects on performance depend on how data is used by the devices and how the devices' driver is going to orchestrate the data transfers across the system.In this paper we present a multi-GPU and Unified Virtual Memory (UM) implementation of the NAS Multi-Zone Parallel Benchmarks which alternate communication and computation phases offering opportunities to overlap these phases. We analyse the programmability and performance effects of the introduction of the UM support.Our experience shows that the programming efforts for introducing UM are similar to those of having a memory allocation per GPU. On an evaluation environment composed of 2 x IBM Power9 8335-GTH and 4 x GPU NVIDIA V100 (Volta), our UM-based parallelization outperforms the manual memory allocation versions by 1.10x to 1.85x. However, these improvements are highly sensitive to the information forwarded to the devices' driver describing the most convenient location for specific memory regions. We analyse these improvements in terms of the relationship between the computational and communication phases of the applications.

Gas-liquid two-phase flows simulation based on weakly compressible scheme with interface-adapted AMR method

A weakly compressible scheme using the interface-adapted adaptive mesh refinement (AMR) method is proposed for simulating low Mach number gas-liquid two-phase flows in large-scale and high-resolution simulations. A pressure evolution equation is derived from the iso-thermal process and low-Mach-number assumptions. We solve the Navier-Stokes equation and the pressure equation explicitly including acoustic waves. In the conventional semi-implicit method to solve incompressible two-phase flows, the computational cost of the linear solver increases as the problem size becomes larger. The conservative phase-field equation was solved using the finite volume method (FVM) to capture the gas-liquid interface while conserving the total mass of each phase. We developed GPU code for a tree-based AMR that greatly reduces the computational cost of assigning a high-resolution mesh around the moving interface region. The data structure of the memory pool has extra space allocated to prevent frequent memory allocations and deallocations. Results for the two-phase flow benchmark problems, 2D rising-bubble and dam-breaking onto dry floor problem, are in good agreement with the results of incompressible solvers. The two-phase flow simulation of the dam-breaking onto a wet floor problem maintained the shape interface regardless of whether the AMR method was used or not. We have computed a two-dimensional soap bubble growing from 16mm to 120mm in diameter with about 0.9 million total cells. The final thickness of the liquid film was about 0.1mm with the finest cell size being 0.0195mm, equivalent to 8,192×8,192 uniform cells. This computation showed that high-resolution two-phase flow simulation with the explicit scheme and AMR method reproduced the behavior of the thin liquid film.

An Approach in Big Data Analytics to Improve the Velocity of Unstructured Data Using MapReduce

Big Data Analytics is an innovative approach for extracting the data from a huge volume of data warehouse systems. It reveals the method to compress the high volume of data into clusters by MapReduce and HDFS. However, the data processing has taken more time for extract and store in Hadoop clusters. The proposed system deals with the challenges of time delay in shuffle phase of map-reduce due to scheduling and sequencing. For improving the speed of big data, this proposed work using the Compressed Elastic Search Index (CESI) and MapReduce-Based Next Generation Sequencing Approach (MRBNGSA). This approach helps to increase the speed of data retrieval from HDFS clusters because of the way it is stored in that. this method is stored only the metadata in HDFS which takes less memory during runtime compare to big data due to the volume of data stored in HDFS. This approach is reduces the CPU utilization and memory allocation of the resource manager in Hadoop Framework and imroves data processing speed, such a way that time delay has to be reduced with minimum latency.

DMAS Beamforming with Complementary Subset Transmit for Ultrasound Coherence-Based Power Doppler Detection in Multi-Angle Plane-Wave Imaging.

Conventional ultrasonic coherent plane-wave (PW) compounding corresponds to Delay-and-Sum (DAS) beamforming of low-resolution images from distinct PW transmit angles. Nonetheless, the trade-off between the level of clutter artifacts and the number of PW transmit angle may compromise the image quality in ultrafast acquisition. Delay-Multiply-and-Sum (DMAS) beamforming in the dimension of PW transmit angle is capable of suppressing clutter interference and is readily compatible with the conventional method. In DMAS, a tunable p value is used to modulate the signal coherence estimated from the low-resolution images to produce the final high-resolution output and does not require huge memory allocation to record all the received channel data in multi-angle PW imaging. In this study, DMAS beamforming is used to construct a novel coherence-based power Doppler detection together with the complementary subset transmit (CST) technique to further reduce the noise level. For p = 2.0 as an example, simulation results indicate that the DMAS beamforming alone can improve the Doppler SNR by 8.2 dB compared to DAS counterpart. Another 6-dB increase in Doppler SNR can be further obtained when the CST technique is combined with DMAS beamforming with sufficient ensemble averaging. The CST technique can also be performed with DAS beamforming, though the improvement in Doppler SNR and CNR is relatively minor. Experimental results also agree with the simulations. Nonetheless, since the DMAS beamforming involves multiplicative operation, clutter filtering in the ensemble direction has to be performed on the low-resolution images before DMAS to remove the stationary tissue without coupling from the flow signal.

LPE: Locality-Based Dead Prediction in Exclusive TLB for Large Coverage

Translation lookaside buffer (TLB) is critical to modern multi-level memory systems’ performance. However, due to the limited size of the TLB itself, its address coverage is limited. Adopting a two-level exclusive TLB hierarchy can increase the coverage [M. Swanson, L. Stoller and J. Carter, Increasing TLB reach using superpages backed by shadow memory, 25th Annual Int. Symp. Computer Architecture (1998); H.P. Chang, T. Heo, J. Jeong and J. Huh Hybrid TLB coalescing: Improving TLB translation coverage under diverse fragmented memory allocations, ACM SIGARCH Comput. Arch. News 45 (2017) 444–456] to improve memory performance. However, after analyzing the existing two-level exclusive TLBs, we find that a large number of “dead” entries (they will have no further use) exist in the last-level TLB (LLT) for a long time, which occupy much cache space and result in low TLB hit-rate. Based on this observation, we first propose exploiting temporal and spatial locality to predict and identify dead entries in the exclusive LLT and remove them as soon as possible to leave room for more valid data to increase the TLB hit rates. Extensive experiments show that our method increases the average hit rate by 8.67%, to a maximum of 19.95%, and reduces total latency by an average of 9.82%, up to 24.41%.

Synaptic plasticity-dependent competition rule influences memory formation

Memory is supported by a specific collection of neurons distributed in broad brain areas, an engram. Despite recent advances in identifying an engram, how the engram is created during memory formation remains elusive. To explore the relation between a specific pattern of input activity and memory allocation, here we target a sparse subset of neurons in the auditory cortex and thalamus. The synaptic inputs from these neurons to the lateral amygdala (LA) are not potentiated by fear conditioning. Using an optogenetic priming stimulus, we manipulate these synapses to be potentiated by the learning. In this condition, fear memory is preferentially encoded in the manipulated cell ensembles. This change, however, is abolished with optical long-term depression (LTD) delivered shortly after training. Conversely, delivering optical long-term potentiation (LTP) alone shortly after fear conditioning is sufficient to induce the preferential memory encoding. These results suggest a synaptic plasticity-dependent competition rule underlying memory formation.

Distributed and vectorized method of characteristics for fast transient simulations in water distribution systems

AbstractModeling transient flow in networked dynamical systems characterized by hyperbolic partial differential equations (PDEs) is essential to engineering applications. Solutions of hyperbolic PDEs are commonly found using the method of characteristics (MOC), particularly when modeling the water hammer phenomenon in water distribution systems (WDSs), which is critical for design and operation. For applications that require fast modeling, existing methods for speeding up traditional MOC simulations either trade off accuracy for simulation time, or do not scale properly due to memory restrictions and prolonged computational times. This work proposes a novel parallel implementation of the MOC for networked systems, which relies on vectorization and distributed parallel computing to evaluate the transient dynamics of WDSs. The proposed method, referred to as distributed and vectorized MOC (DV‐MOC), relies on aligned memory allocation for vectorization and distributed‐memory parallelization to further accelerate vectorized operations and ensure scalability for arbitrary network topologies. The algorithm has been applied to a WDS from the battle of the sensor networks (BWSN‐II) composed by 14,824 pipes and nearly solution points. Through rigorous analyses, we show that the performance of DV‐MOC surpasses that of sequential MOC, with speeding up factors in the order of thousands for sufficiently dense numerical grids.

Retrieval of behavior trees using map-and-reduce technique

There has been an increased interest in the creation of AI social agents who possess complex behaviors that allow them to perform social interactions. Behavior trees provide a plan model execution that has been widely used to build complex behaviors for AI social agents. Behavior trees can be represented in the form of a memory structure known as cognitive scripts, which would allow them to evolve through further development over multiple exposure to repeated enactment of a particular behavior or similar ones. Behavior trees that share the same context will then be able to learn from each other resulting in new behavior trees with richer experience. The main challenge appears in the expensive cost of retrieving contextually similar behavior trees (scripts) from a repertoire of scripts to allow for that learning process to occur. This paper introduces a novel application of map-and-reduce technique to retrieve cognitive with low computational time and memory allocation. The paper focuses on the design of a corpus of cognitive scripts, as a knowledge engineering key challenge, and the application of map-and-reduce with semantic information to retrieve contextually similar cognitive scripts. The results are compared to other techniques used to retrieve cognitive scripts in the literature, such as Pharaoh which uses the least common parent (LCP) technique in its core. The results show that the map-and-reduce technique can be successfully used to retrieve cognitive scripts with high retrieval accuracy of 92.6%, in addition to being cost effective.

Recognition of handwritten MNIST digits on low-memory 2 Kb RAM Arduino board using LogNNet reservoir neural network

The presented compact algorithm for recognizing handwritten digits of the MNIST database, created on the LogNNet reservoir neural network, reaches the recognition accuracy of 82%. The algorithm was tested on a low-memory Arduino board with 2 Kb static RAM low-power microcontroller. The dependences of the accuracy and time of image recognition on the number of neurons in the reservoir have been investigated. The memory allocation demonstrates that the algorithm stores all the necessary information in RAM without using additional data storage, and operates with original images without preliminary processing. The simple structure of the algorithm, with appropriate training, can be adapted for wide practical application, for example, for creating mobile biosensors for early diagnosis of adverse events in medicine. The study results are important for the implementation of artificial intelligence on peripheral constrained IoT devices and for edge computing.

Penalty- and Locality-aware Memory Allocation in Redis Using Enhanced AET

Due to large data volume and low latency requirements of modern web services, the use of an in-memory key-value (KV) cache often becomes an inevitable choice (e.g., Redis and Memcached). The in-memory cache holds hot data, reduces request latency, and alleviates the load on background databases. Inheriting from the traditional hardware cache design, many existing KV cache systems still use recency-based cache replacement algorithms, e.g., least recently used or its approximations. However, the diversity of miss penalty distinguishes a KV cache from a hardware cache. Inadequate consideration of penalty can substantially compromise space utilization and request service time. KV accesses also demonstrate locality, which needs to be coordinated with miss penalty to guide cache management. In this article, we first discuss how to enhance the existing cache model, the Average Eviction Time model, so that it can adapt to modeling a KV cache. After that, we apply the model to Redis and propose pRedis, Penalty- and Locality-aware Memory Allocation in Redis, which synthesizes data locality and miss penalty, in a quantitative manner, to guide memory allocation and replacement in Redis. At the same time, we also explore the diurnal behavior of a KV store and exploit long-term reuse. We replace the original passive eviction mechanism with an automatic dump/load mechanism, to smooth the transition between access peaks and valleys. Our evaluation shows that pRedis effectively reduces the average and tail access latency with minimal time and space overhead. For both real-world and synthetic workloads, our approach delivers an average of 14.0%∼52.3% latency reduction over a state-of-the-art penalty-aware cache management scheme, Hyperbolic Caching (HC), and shows more quantitative predictability of performance. Moreover, we can obtain even lower average latency (1.1%∼5.5%) when dynamically switching policies between pRedis and HC.

Adaptive Layered Segregated Fit Scheme for Dynamic Memory Allocation

Embedded applications are becoming more complex and are required to utilize computing platform resources more efficiently. Existing dynamic memory allocation (DSA) schemes cannot adaptively perform memory management according to the environment in which they are located or integrate various memory allocation strategies, making it impossible to guarantee a constant execution time. Efficient memory utilization is a crucial challenge for developers, especially in embedded OSs (operating systems). In this paper, we propose an adaptive layered segregated fit (ALSF) scheme for DSA. The ALSF scheme combines dynamic two-dimensional arrays and bitmaps, completes the allocation and freeing of memory blocks within constant execution time, and uses memory splitting technology to reduce internal fragmentation. The proposed scheme also adjusts the number of segregated lists by analyzing the system’s allocation of different memory sizes, which improves the matching accuracy of memory blocks. We conducted a comparative experimental analysis and investigation of the ALSF and two-level segregated fit (TLSF) schemes in the Zephyr OS. Experiments show that the average memory utilization of the proposed ALSF scheme reaches 94.95%. Compared with the TLSF scheme, our scheme has a 12.99% higher allocation success rate in the memory-scarce environment, and the execution speeds of the two are similar.

Robust segmentation of exudates from retinal surface using M-CapsNet via EM routing

Retinopathy is any damage to the retina of the eyes, which causes vision impairment and may lead to blindness. The initial manifestation of retinopathy is identified by the presence of exudates, microaneurysms on the retinal surface. So, the early detection of exudates prevents the further spread and simultaneously reduces the severity of the disease. However, automatic detection of exudates is a challenging task as the exudates vary from each other in terms of shape and size. This paper proposes a novel approach for the automatic segmentation of exudates using an encoder-decoder style network termed as “deep M-CapsNet using Expectation-Maximization (EM) Routing,” which reduces the memory allocation problems in semantic segmenting of objects. In M-CapsNet, every child capsule connects with every parent capsule at every location. Thus the predictions are forwarded to parent capsules using a shared kernel through a matrix capsule. Due to similar intensities between exudates and the optic disc, the M-CapsNet extracts the exudates along with the optic disc from the retinal surface. The optic disc is eliminated from the segmented output using regional and morphological features. This paper achieves an average accuracy of 94%, the specificity of 100%, the sensitivity of 100%, and the F1 score of 95% when tested over the images selected from publicly available datasets randomly. The experiment results demonstrate that M-CapsNet outperformed previous networks in detecting exudates.

Memory State Verification Based on Inductive and Deductive Reasoning

Memory allocation and deallocation are the fundamental operations of embedded operating systems, which have been extensively used in many safety critical systems. The correctness of the operations is of paramount importance because their failure could incur severe consequences. While the system is running, the memory state can easily grow to a gigantic amount, which means that it is impossible to verify the huge memory states one by one. Therefore, it is a challenge how to verify the correctness of running memory state of the system. In this article, we propose a novel memory state verification method based on inductive and deductive reasoning. First, we abstract the memory state as a list of memory blocks, which will transform in memory operations. Second, we construct the generic model based on the transition function of the memory management and summarize the invariant properties of the memory state. Third, we use the inductive method to calculate the changes between the memory states, and verify that the memory state of the system always satisfy the global properties. All the proofs are implemented in the interactive theorem prover Coq. On the basis of our proposed model, we verify the correctness of a two-level segregated fit (TLSF) algorithm through some extensions, and we also apply this method to verify the correctness of the memory state of the embedded system at runtime.

Mobile Assisted Language Learning (MALL): Trends from 2010 to 2020 Using Text Analysis Techniques

Throughout the decade of 2010-2020, the widespread use of mobile devices of any type (smartphones, tablets) has encouraged and strengthened their use in different learning processes and in different ways. Latest improvements in devices’ processing power, in storage capacity, in memory allocation, in wireless connectivity, in GPS and in Bluetooth capabilities, has led to their wider integration and smoother use in the field of learning and in the field of language learning as well. This study examined the trend of a large number of academic studies concerning Mobile Assisted Language Learning (MALL) using text analysis techniques and tools, published in the decade 2010-2020. Over three hundred and forty (340) publications such as journal articles, conferences proceedings papers, books chapters and books were retrieved and analyzed. The preliminary analysis presents the main characteristics and the research trends of this decade and discuss how the field of mobile assisted language learning has evolved in these years.

A Neighborhood Aware Caching and Interest Dissemination Scheme for Content Centric Networks

Content-Centric Networking (CCN) is a promising framework for the next generation Internet architecture that exploits ubiquitous in-network caching to minimize content delivery latency and reduce network traffic. In this paper, we introduce a neighborhood aware mechanism for content caching, named Neighborhood Aware Caching and Interest Dissemination (NACID) that accounts for the popularity of contents and how close the content copies are in the neighborhood. We use a very low-overhead, Bloom Filter based dissemination of caching information in the neighborhood. Given the neighborhood cached contents, the proposed scheme decides when and how to handle the additional caching of content and its eviction. Simulation results show that NACID performs substantially better than the existing CCN caching policies. We also study different heterogeneous cache memory allocation strategies and show that the simpler homogeneous allocation strategies work almost as well.