K40 GPU Research Articles

New Adversarial Image Detection Based on Sentiment Analysis.

Deep neural networks (DNNs) are vulnerable to adversarial examples, while adversarial attack models, e.g., DeepFool, are on the rise and outrunning adversarial example detection techniques. This article presents a new adversarial example detector that outperforms state-of-the-art detectors in identifying the latest adversarial attacks on image datasets. Specifically, we propose to use sentiment analysis for adversarial example detection, qualified by the progressively manifesting impact of an adversarial perturbation on the hidden-layer feature maps of a DNN under attack. Accordingly, we design a modularized embedding layer with the minimum learnable parameters to embed the hidden-layer feature maps into word vectors and assemble sentences ready for sentiment analysis. Extensive experiments demonstrate that the new detector consistently surpasses the state-of-the-art detection algorithms in detecting the latest attacks launched against ResNet and Inception neutral networks on the CIFAR-10, CIFAR-100, and SVHN datasets. The detector only has about 2 million parameters and takes less than 4.6 ms to detect an adversarial example generated by the latest attack models using a Tesla K80 GPU card.

Reconfigurable origami hologram based on deep neural networks.

Reconfigurable and multifunctional metasurfaces are becoming indispensable in a variety of applications due to their capability to execute diverse functions across various states. However, many of these metasurfaces incorporate complex active components, thereby escalating structural complexity and bulk volume. In this research, we propose a reconfigurable passive hologram based solely on an origami structure, enabling the successful generation of holograms depicting the 'Z' and 'L' illuminated by a right-hand circular polarization (RHCP) wave in two distinct states: planar and zigzag configuration, respectively. The transformation between the 2D planar metasurface and the 3D zigzag structure with slant angles of 35 is achieved solely through mechanically stretching and compressing the origami metasurface. The phases on the origami metasurface are trained through a deep neural network which operates on the NVIDIA Tesla k80 GPU, with the total training process costing 11.88 s after 100 epochs. The reconfigurable scheme proposed in this research provides flexibility and ease of implementation in the fields of imaging and data processing.

BrutNet: A novel approach for violence detection and classification using DCNN with GRU

AbstractAutomatic Violence Detection and Classification (AVDC) with deep learning has garnered significant attention in computer vision research. This paper presents a novel approach for combining a custom Deep Convolutional Neural Network (DCNN) with a Gated Recurrent Unit (GRU) in developing a new AVDC model called BrutNet. Specifically, a time‐distributed DCNN (TD‐DCNN) is developed to generate a compact 2D representation with 512 spatial features per frame from a set of equally‐spaced frames of dimension 16090 in short video segments. Further to leverage the temporal information, a GRU layer is utilised, generating a condensed 1D vector that enables binary classification of violent or non‐violent content through multiple dense layers. Overfitting is addressed by incorporating dropout layers with a rate of 0.5, while the hidden and output layers employ rectified linear unit (ReLU) and sigmoid activations, respectively. The model is trained on the NVIDIA Tesla K80 GPU through Google Colab, demonstrating superior performance compared to existing models across various video datasets, including hockey fights, movie fights, AVD, and RWF‐2000. Notably, the model stands out by requiring only 3.416 million parameters and achieving impressive test accuracies of 97.62%, 100%, 97.22%, and 86.43% on the respective datasets. Thus, BrutNet exhibits the potential to emerge as a highly efficient and robust AVDC model in support of greater public safety, content moderation and censorship, computer‐aided investigations, and law enforcement.

Empowering edge devices: FPGA‐based 16‐bit fixed‐point accelerator with SVD for CNN on 32‐bit memory‐limited systems

AbstractConvolutional neural networks (CNNs) are now often used in deep learning and computer vision applications. Its convolutional layer accounts for most calculations and should be computed fast in a local edge device. Field‐programmable gate arrays (FPGAs) have been adequately explored as promising hardware accelerators for CNNs due to their high performance, energy efficiency, and reconfigurability. This paper developed an efficient FPGA‐based 16‐bit fixed‐point hardware accelerator unit for deep learning applications on the 32‐bit low‐memory edge device (PYNQ‐Z2 board). Additionally, singular value decomposition is applied to the fully connected layer for dimensionality reduction of weight parameters. The accelerator unit was designed for all five layers and employed eight processing elements in convolution layers 1 and 2 for parallel computations. In addition, array partitioning, loop unrolling, and pipelining are the techniques used to increase the speed of calculations. The AXI‐Lite interface was also used to communicate between IP and other blocks. Moreover, the design is tested with grayscale image classification on MNIST handwritten digit dataset and color image classification on the Tumor dataset. The experimental results show that the proposed accelerator unit implementation performs faster than the software‐based implementation. Its inference speed is 89.03% more than INTEL 3‐core CPU, 86.12% higher than Haswell 2‐core CPU, and 82.45% more than NVIDIA Tesla K80 GPU. Furthermore, the throughput of the proposed design is 4.33GOP/s, which is better than the conventional CNN accelerator architectures.

ReSTiNet: On Improving the Performance of Tiny-YOLO-Based CNN Architecture for Applications in Human Detection

Human detection is a special application of object recognition and is considered one of the greatest challenges in computer vision. It is the starting point of a number of applications, including public safety and security surveillance around the world. Human detection technologies have advanced significantly in recent years due to the rapid development of deep learning techniques. Despite recent advances, we still need to adopt the best network-design practices that enable compact sizes, deep designs, and fast training times while maintaining high accuracies. In this article, we propose ReSTiNet, a novel compressed convolutional neural network that addresses the issues of size, detection speed, and accuracy. Following SqueezeNet, ReSTiNet adopts the fire modules by examining the number of fire modules and their placement within the model to reduce the number of parameters and thus the model size. The residual connections within the fire modules in ReSTiNet are interpolated and finely constructed to improve feature propagation and ensure the largest possible information flow in the model, with the goal of further improving the proposed ReSTiNet in terms of detection speed and accuracy. The proposed algorithm downsizes the previously popular Tiny-YOLO model and improves the following features: (1) faster detection speed; (2) compact model size; (3) solving the overfitting problems; and (4) superior performance than other lightweight models such as MobileNet and SqueezeNet in terms of mAP. The proposed model was trained and tested using MS COCO and Pascal VOC datasets. The resulting ReSTiNet model is 10.7 MB in size (almost five times smaller than Tiny-YOLO), but it achieves an mAP of 63.74% on PASCAL VOC and 27.3% on MS COCO datasets using Tesla k80 GPU.

GPU-Accelerated Computation of EM Scattering of a Time-Evolving Oceanic Surface Model II: EM Scattering of Actual Oceanic Surface

Based on marine environmental factors of different sea areas, a high-performance sea clutter time series modeling algorithm for the real sea surface is developed to study the amplitude mean and Doppler spectrum characteristics of sea clutter. The European Centre for Medium-Range Weather Forecasts (ECMWF) data set (ERA-Interim) and ESA’s soil moisture and ocean salinity (SMOS) data set are utilized to establish databases of different marine environmental factors. Combined with the mixed spectrum model, the geometric fine structure of wind-driven sea surface with swell superposition is established by using the double-superposition method (DSM) and comprehensively considering small-scale capillary ripples, large-scale gravity waves and swell. A triangle facet-based sea clutter series modeling algorithm is developed, in which the quasi-specular scattering based on a triangle and the scattering based on gravity wave modulation capillary spectrum are calculated, respectively, and compared with the measured results. For high-resolution radar, dynamic sea surface modeling and sea clutter calculation are very time consuming. In this paper, the Tesla K80 GPU manufactured by NVIDIA in Santa Clara, Computed Unified Device architecture (CUDA) high-performance parallel technology and some optimization strategies are adopted to improve the efficiency of sea clutter modeling. The results can be used to analyze the distribution characteristics of marine factors, the average amplitude and Doppler characteristics of sea clutter in different sea areas.

Optimizing the GPU based method calculating energy deposition of beams coupling with discrete materials in dynamical and thermal simulations for higher computing efficiency

In order to obtain more accurate energy deposition simulation results of beams coupling with discrete materials especially the granular materials in dynamical and thermal simulations, a discrete energy deposition calculation method was proposed previously to replace the equivalent homogenization method and accelerated with GPUs (Tian et al., 2021 [10]). However, it was found that the computing efficiency drops so severely with the increasing of the energy space size and the time required for simulations increases greatly unacceptably. In this work, for higher computing performance, based on the bottleneck analyses of CUDA (Compute Unified Device Architecture) kernels of the previous method, two improvements were made to the space cell marking phase and the energy deposition phase. In the space cell marking phase, a new proposed Lagrange-Euler Mapping Method searching the fixed space cells through flowing grains replaced the previous Euler Searching Method searching the flowing grains from fixed space cells. In the energy deposition phase, a warp aggregation method was used, assigning fewer threads to perform atomic operations. In the simulation of a granular-flow target bombarded by a 1 GeV proton beam setting a large energy computing space with a number of cells up to ∼108, the improved algorithm can effectively reduce the number of memory access instructions of CUDA warps. As a result, the computing performance was improved by 85% on A100 GPUs and 206% on K80 GPUs, making the dynamical and thermal simulations of granular-flow targets more efficient. Our method also could be beneficial for simulating the interactions between arbitrary source (or field) with a fixed spatial distribution and discrete materials.

Analysis of the Leaky Integrate-and-Fire neuron model for GPU implementation

Understanding how neurons perform, when they are organized in interacting networks, is a key to understanding how the brain performs complex functions. Different models that approximate the behavior of interconnected neurons have been proposed in the literature. Implementing these models to simulate neuron behavior at an appropriately detailed level to observe collective phenomena is computationally intensive. In this study we analyze the coupled Leaky Integrate-and-Fire model and report on the issues that affect performance when the model is implemented on a GPU. We conclude that the problem is heavily memory-bound. Advances in memory technology at the hardware level seem to be the deciding factor to achieve better performance on the GPU. Our results show that using an NVidia K40 GPU a modest 2x speedup can be achieved compared to a parallel implementation running on a modern multi-core CPU. However, a substantial speedup of 11.1x can be achieved using an NVidia V100 GPU, mainly due to the improvements in its memory subsystem.

Translating the InChI: adapting neural machine translation to predict IUPAC names from a chemical identifier

We present a sequence-to-sequence machine learning model for predicting the IUPAC name of a chemical from its standard International Chemical Identifier (InChI). The model uses two stacks of transformers in an encoder-decoder architecture, a setup similar to the neural networks used in state-of-the-art machine translation. Unlike neural machine translation, which usually tokenizes input and output into words or sub-words, our model processes the InChI and predicts the IUPAC name character by character. The model was trained on a dataset of 10 million InChI/IUPAC name pairs freely downloaded from the National Library of Medicine’s online PubChem service. Training took seven days on a Tesla K80 GPU, and the model achieved a test set accuracy of 91%. The model performed particularly well on organics, with the exception of macrocycles, and was comparable to commercial IUPAC name generation software. The predictions were less accurate for inorganic and organometallic compounds. This can be explained by inherent limitations of standard InChI for representing inorganics, as well as low coverage in the training data.

Quantum algorithm for quicker clinical prognostic analysis: an application and experimental study using CT scan images of COVID-19 patients

BackgroundIn medical diagnosis and clinical practice, diagnosing a disease early is crucial for accurate treatment, lessening the stress on the healthcare system. In medical imaging research, image processing techniques tend to be vital in analyzing and resolving diseases with a high degree of accuracy. This paper establishes a new image classification and segmentation method through simulation techniques, conducted over images of COVID-19 patients in India, introducing the use of Quantum Machine Learning (QML) in medical practice.MethodsThis study establishes a prototype model for classifying COVID-19, comparing it with non-COVID pneumonia signals in Computed tomography (CT) images. The simulation work evaluates the usage of quantum machine learning algorithms, while assessing the efficacy for deep learning models for image classification problems, and thereby establishes performance quality that is required for improved prediction rate when dealing with complex clinical image data exhibiting high biases.ResultsThe study considers a novel algorithmic implementation leveraging quantum neural network (QNN). The proposed model outperformed the conventional deep learning models for specific classification task. The performance was evident because of the efficiency of quantum simulation and faster convergence property solving for an optimization problem for network training particularly for large-scale biased image classification task. The model run-time observed on quantum optimized hardware was 52 min, while on K80 GPU hardware it was 1 h 30 min for similar sample size. The simulation shows that QNN outperforms DNN, CNN, 2D CNN by more than 2.92% in gain in accuracy measure with an average recall of around 97.7%.ConclusionThe results suggest that quantum neural networks outperform in COVID-19 traits’ classification task, comparing to deep learning w.r.t model efficacy and training time. However, a further study needs to be conducted to evaluate implementation scenarios by integrating the model within medical devices.

Acceleration of Parallel-Blocked QR Decomposition of Tall-and-Skinny Matrices on FPGAs

QR decomposition is one of the most useful factorization kernels in modern numerical linear algebra algorithms. In particular, the decomposition of tall-and-skinny matrices (TSMs) has major applications in areas including scientific computing, machine learning, image processing, wireless networks, and numerical methods. Traditionally, CPUs and GPUs have achieved better throughput on these applications by using large cache hierarchies and compute cores running at a high frequency, leading to high power consumption. With the advent of heterogeneous platforms, however, FPGAs are emerging as a promising viable alternative. In this work, we propose a high-throughput FPGA-based engine that has a very high computational efficiency (ratio of achieved to peak throughput) compared to similar QR solvers running on FPGAs. Although comparable QR solvers achieve an efficiency of 36%, our design exhibits an efficiency of 54%. For TSMs, our experimental results show that our design can outperform highly optimized QR solvers running on CPUs and GPUs. For TSMs with more than 50K rows, our design outperforms the Intel MKL solver running on an Intel quad-core processor by a factor of 1.5×. For TSMs containing 256 columns or less, our design outperforms the NVIDIA CUBLAS solver running on a K40 GPU by a factor of 3.0×. In addition to being fast, our design is energy efficient—competing platforms execute up to 0.6 GFLOPS/Joule, whereas our design executes more than 1.0 GFLOPS/Joule.

A novel Ship detection method from SAR image with reduced false alarm

Many research works using deep learning techniques for automatic ship detection from SAR images have good detection accuracy. But the main problem in these methods is false detection, mostly due to speckle presence. Therefore, we propose a new deep learning model with a novel preprocessing stage to address this problem. We are introducing a deep learning architecture to detect and localize ships in the SAR image. First, generate a three-channel image from gray-scale SAR image. Then, this image is used to train the model to predict the ship’s position in the SAR image. We experimented on the public SAR ship detection dataset (SSDD) and Dataset of Ship Detection for Deep Learning under Complex Backgrounds (SDCD) to validate the proposed method’s feasibility. We used python 3.5 for coding with the Keras framework in the NVIDIA Tesla K80 GPU hardware platform. The experimental results indicated that our proposed method’s ship detection accuracy has increased with reduced false detection percentage.

GG-Net: Gaze Guided Network for Self-driving Cars

Imitation learning is used massively in autonomous driving for training networks to predict steering commands from frames using annotated data collected by an expert driver. Believing that the frames taken from a front-facing camera are completely mimicking the driver’s eyes raises the question of how eyes and the complex human vision system attention mechanisms perceive the scene. This paper proposes the idea of incorporating eye gaze information with the frames into an end-to-end deep neural network in the lane-following task. The proposed novel architecture, GG-Net, is composed of a spatial transformer network (STN), and a multitask network to predict steering angle as well as the gaze map for the input frame. The experimental results of this architecture show a great improvement in steering angle prediction accuracy of 36% over the baseline with inference time of 0.015 seconds per frame (66 fps) using NVIDIA K80 GPU enabling the proposed model to operate in real-time. We argue that incorporating gaze maps enhances the model generalization capability to the unseen environments. Additionally, a novel course-steering angle conversion algorithm with a complementing mathematical proof is proposed.

High-throughput Protein Sequence Alignment on Multi-core Systems

Rapid evolution in sequencing technologies results in generating data on an enormous scale. A focal and main challenge in analyzing data at such a large scale is the alignment of the DNA/Protein sequences, whereby reads are compared to the reference sequences. To find similar sequences, alignment algorithms are used to align a query sequence with the database. Alignment algorithms can be utilized to classify the source of a sequence, to discover similarities among the organisms, or to deduce a progenitor connection. A wide range of algorithms for alignment has been developed in recent years. In this paper, an accurate method of accelerating such algorithms using GPUs has been investigated. A Swiss-Prot database has been processed using GPU implemented Smith-Waterman Sequence Alignment Algorithm. The first step in the process generates the alignment scores but not the actual alignment. Various available alignment tools like ssearch2 are then utilized to align the output file generated during the first step. The performance of GPU-accelerated implementation as compared to other techniques is then evaluated for performance /throughput improvement. Swiss-Prot database was aligned using various alignment tools. NVIDIA TESLA K40 GPU is being utilized for generating the results for this research. This implementation achieves the performance of 44.3 Giga cell updates per second (GCUPS), which is 22.9 times better than its implementation on GTX 275. Performance is improved as the workload of sequences of equal length is equally distributed among all the threads on Multiprocessors of GPU.

EBrainII: a 3\xa0kW Realtime Custom 3D DRAM Integrated ASIC Implementation of a Biologically Plausible Model of a Human Scale Cortex

The Artificial Neural Networks (ANNs), like CNN/DNN and LSTM, are not biologically plausible. Despite their initial success, they cannot attain the cognitive capabilities enabled by the dynamic hierarchical associative memory systems of biological brains. The biologically plausible spiking brain models, e.g., cortex, basal ganglia, and amygdala, have a greater potential to achieve biological brain like cognitive capabilities. Bayesian Confidence Propagation Neural Network (BCPNN) is a biologically plausible spiking model of the cortex. A human-scale model of BCPNN in real-time requires 162 TFlop/s, 50 TBs of synaptic weight storage to be accessed with a bandwidth of 200 TBs. The spiking bandwidth is relatively modest at 250 GBs/s. A hand-optimized implementation of rodent scale BCPNN has been done on Tesla K80 GPUs require 3 kWs, we extrapolate from that a human scale network will require 3 MWs. These power numbers rule out such implementations for field deployment as cognition engines in embedded systems.The key innovation that this paper reports is that it is feasible and affordable to implement real-time BCPNN as a custom tiled application-specific integrated circuit (ASIC) in 28 nm technology with custom 3D DRAM - eBrainII - that consumes 3 kW for human scale and 12 watts for rodent scale. Such implementations eminently fulfill the demands for field deployment.

Efficient Hardware Architectures for 1D- and MD-LSTM Networks

Recurrent Neural Networks, in particular One-dimensional and Multidimensional Long Short-Term Memory (1D-LSTM and MD-LSTM) have achieved state-of-the-art classification accuracy in many applications such as machine translation, image caption generation, handwritten text recognition, medical imaging and many more. However, high classification accuracy comes at high compute, storage, and memory bandwidth requirements, which make their deployment challenging, especially for energy-constrained platforms such as portable devices. In comparison to CNNs, not so many investigations exist on efficient hardware implementations for 1D-LSTM especially under energy constraints, and there is no research publication on hardware architecture for MD-LSTM. In this article, we present two novel architectures for LSTM inference: a hardware architecture for MD-LSTM, and a DRAM-based Processing-in-Memory (DRAM-PIM) hardware architecture for 1D-LSTM. We present for the first time a hardware architecture for MD-LSTM, and show a trade-off analysis for accuracy and hardware cost for various precisions. We implement the new architecture as an FPGA-based accelerator that outperforms NVIDIA K80 GPU implementation in terms of runtime by up to 84× and energy efficiency by up to 1238× for a challenging dataset for historical document image binarization from DIBCO 2017 contest, and a well known MNIST dataset for handwritten digits recognition. Our accelerator demonstrates highest accuracy and comparable throughput in comparison to state-of-the-art FPGA-based implementations of multilayer perceptron for MNIST dataset. Furthermore, we present a new DRAM-PIM architecture for 1D-LSTM targeting energy efficient compute platforms such as portable devices. The DRAM-PIM architecture integrates the computation units in a close proximity to the DRAM cells in order to maximize the data parallelism and energy efficiency. The proposed DRAM-PIM design is 16.19 × more energy efficient as compared to FPGA implementation. The total chip area overhead of this design is 18 % compared to a commodity 8 Gb DRAM chip. Our experiments show that the DRAM-PIM implementation delivers a throughput of 1309.16 GOp/s for an optical character recognition application.

ZMCintegral-v5: Support for Integrations with the scanning of large parameter space on multi-GPUs

In this updated version of ZMCintegral, we have added the functionality of parameter scan for integrations with a large parameter space (up to 1010 points to be scanned). The Python API is kept the same as the previous ones and users have full flexibility to define their own integrands. The performance of the new functionality is tested for multi-nodes conditions. Program summaryProgram Title: ZMCintegralProgram Files doi:http://dx.doi.org/10.17632/p7wc7k6mpp.2Licensing provisions: Apache-2.0Programming language: PythonJournal reference of previous version: Comput. Phys. Commun. 248 (2020) 106962Does the new version supersede the previous version?: YesReasons for the new version: For many physical cases, one usually encounters integrations containing a large parameter space [1,2,3,4]. In these conditions, integrations of the same form (but with different parameters) need be evaluated for millions of times. However, our previous versions [5] mainly focus on a single integration of high dimension, hence lack the good performance for many integrations of relatively high dimensions (e.g., the scan of parameter space). Therefore, we have added the demanding functionality of parameter scan. For each point in the parameter space, an integration will be evaluated, and all these integrations will be allocated automatically.Summary of revisions:•Integration with parameter space. We use the six-dimensional integration (1)f(x1,x2,x3,x4)=(∏k=16∫01dyk)sin(∑j6yj+∑l=14xl),to demonstrate the functionality of parameter scan. Here, the component of each parameter point xl∈{x1,x2,x3,x4} takes values in list [0,1,2,⋯,99], thus forming a parameter space of size 1004 (i.e., 108 points to be scanned). The allocation of the thread and block configurations will be handled by the code automatically. •Test on multiple (three) nodes. The hardware conditions for the three nodes are Intel(R) Xeon(R) CPU E5-2620 v3@2.40 GHz CPU with 24 processors + 4 Nvidia Tesla K40m GPUs, Intel(R) Xeon(R) CPU E5-2680 V4@2.40 GHz CPU with 10 processors + 2 Nvidia Tesla K80 GPUs, and Intel(R) Xeon(R) Silver 4110 CPU@2.10 GHz CPU with 10 processors + 1 Nvidia Tesla V100 GPU. These three nodes are in a local area network. We perform the integration of Eq. (1) independently for 10 times. The results are shown in Table 1. Note that the accuracy of the integrations purely depends on the number of samples in direct Monte Carlo method.Nature of problem: ZMCintegral is an easy to use Python package for doing high dimensional integrations on distributed GPU clusters. With the Python libraries Numba [6] and Ray [7], as well as the NVIDIA CUDA [8], ZMCintegral provides a succinct Python interface for evaluating numerical integrations in physical problems. In this new version, we have mainly focused on one kind of problem where the integration contains a large parameter space (up to 1010 parameter points).Solution method: This new version of ZMCintegral contains two Python classes. One class (ZMCintegral_normal, similar to our previous version), which is for high-dimensional integration, uses stratified-sampling and heuristic-tree-search technics. The other class (ZMCintegral_functional, the new functionality) uses direct Monte Carlo method and distributes the tasks of integrations (with each containing a different parameter point) on various devices automatically. In ZMCintegral_functional, the integration is performed on a single GPU thread and each thread gives the result of the integration for one specific parameter point.Additional comments: If the integrations are high-dimensional (e.g. dimensionality of 8–12), we encourage users to use ZMCintegral_normal. If the integrations are middle-dimensional (e.g. dimensionality of 1-7) but have a large parameter space, we suggest users to try ZMCintegral_functional. The detailed instructions can be found here: [9]. Declaration of Competing InterestThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgment The authors are supported in part by the Major State Basic Research Development Program (973 Program) in China under Grant No. 2015CB856902 and by the National Natural Science Foundation of China (NSFC) under Grant No. 11535012. The Computations are performed at the GPU servers of department of modern physics at USTC. We are thankful for the valuable discussions with Prof. Qun Wang of department of modern physics at USTC.

A Comparative Analysis of Data Augmentation Approaches for Magnetic Resonance Imaging (MRI) Scan Images of Brain Tumor.

Introduction:Machine Learning (ML) is a rapidly growing subfield of Artificial Intelligence (AI). It is used for different purposes in our daily life such as face recognition, speech recognition, text translation in different languages, weather prediction, and business prediction. In parallel, ML also plays an important role in the medical domain such as in medical imaging. ML has various algorithms that need to be trained with large volumes of data to produce a well-trained model for prediction.Aim:The aim of this study is to highlight the most suitable Data Augmentation (DA) technique(s) for medical imaging based on their results.Methods:DA refers to different approaches that are used to increase the size of datasets. In this study, eight DA approaches were used on publicly available low-grade glioma tumor datasets obtained from the Tumor Cancer Imaging Archive (TCIA) repository. The dataset included 1961 MRI brain scan images of low-grade glioma patients. You Only Look Once (YOLO) version 3 model was trained on the original dataset and the augmented datasets separately. A neural network training/testing ecosystem named as supervisely with Tesla K80 GPU was used for YOLO v3 model training on all datasets.Results:The results showed that the DA techniques rotate at 180o and rotate at 90o performed the best as data enhancement techniques for medical imaging.Conclusion:Rotation techniques are found significant to enhance the low volume of medical imaging datasets.

Fast Acquisition of Spread Spectrum Signals Using Multiple GPUs

Spread spectrum (SS) signal acquisition in satellite communication is a very computation intensive technique, which hinders the development of real-time spread spectrum signal acquisition. In this paper, in order to achieve real-time acquisition, we propose a multi-GPU based SS signal acquisition algorithm. First, sliding correlation, the computation kernel, is formulated and efficiently parallelized by CUDA. Second, a CUDA-enabled SS signal acquisition algorithm is implemented by adopting the CUDA-enabled sliding correlation. Third, a multi-GPU based algorithm is implemented by using multi-GPU programming. The performance is evaluated in a real SS telemetry system. Real-time acquisition is achieved in all cases by using a single K40 GPU. Furthermore, an average of 374.7 $\times$ speedup (max 473 $\times$ ) in 6 datasets is achieved when using four K40 GPUs. Good scalability is observed when varying the parameters.

Exploiting Bank Conflict-based Side-channel Timing Leakage of GPUs

To prevent information leakage during program execution, modern software cryptographic implementations target constant-time function, where the number of instructions executed remains the same when program inputs change. However, the underlying microarchitecture behaves differently when processing different data inputs, impacting the execution time of the same instructions. These differences in execution time can covertly leak confidential information through a timing channel. Given the recent reports of covert channels present on commercial microprocessors, a number of microarchitectural features on CPUs have been re-examined from a timing leakage perspective. Unfortunately, a similar microarchitectural evaluation of the potential attack surfaces on GPUs has not been adequately performed. Several prior work has considered a timing channel based on the behavior of a GPU’s coalescing unit. In this article, we identify a second finer-grained microarchitectural timing channel, related to the banking structure of the GPU’s Shared Memory. By considering the timing channel caused by Shared Memory bank conflicts, we have developed a differential timing attack that can compromise table-based cryptographic algorithms. We implement our timing attack on an Nvidia Kepler K40 GPU and successfully recover the complete 128-bit encryption key of an Advanced Encryption Standard (AES) GPU implementation using 900,000 timing samples. We also evaluate the scalability of our attack method by attacking an implementation of the AES encryption algorithm that fully occupies the compute resources of the GPU. We extend our timing analysis onto other Nvidia architectures: Maxwell, Pascal, Volta, and Turing GPUs. We also discuss countermeasures and experiment with a novel multi-key implementation, evaluating its resistance to our side-channel timing attack and its associated performance overhead.