Graph Convolutional Networks (GCNs) are increasingly adopted in large-scale graph-based recommender systems. Training GCN requires the minibatch generator traversing graphs and sampling the sparsely located neighboring nodes to obtain their features. Since real-world graphs often exceed the capacity of GPU memory, current GCN training systems keep the feature table in host memory and rely on the CPU to collect sparse features before sending them to the GPUs. This approach, however, puts tremendous pressure on host memory bandwidth and the CPU. This is because the CPU needs to (1) read sparse features from memory, (2) write features into memory as a dense format, and (3) transfer the features from memory to the GPUs. In this work, we propose a novel GPU-oriented data communication approach for GCN training, where GPU threads directly access sparse features in host memory through zero-copy accesses without much CPU help. By removing the CPU gathering stage, our method significantly reduces the consumption of the host resources and data access latency. We further present two important techniques to achieve high host memory access efficiency by the GPU: (1) automatic data access address alignment to maximize PCIe packet efficiency, and (2) asynchronous zero-copy access and kernel execution to fully overlap data transfer with training. We incorporate our method into PyTorch and evaluate its effectiveness using several graphs with sizes up to 111 million nodes and 1.6 billion edges. In a multi-GPU training setup, our method is 65-92% faster than the conventional data transfer method, and can even match the performance of all-in-GPU-memory training for some graphs that fit in GPU memory.
TorchKM: A GPU-Oriented Library for Kernel Learning and Model Selection
University of Iowa（爱荷华大学）；University of Minnesota（明尼苏达大学）；Individual Researcher（独立研究者）；AIScientists, Inc. (MorphMind)（AIScientists公司（MorphMind））；Department of Statistics and Actuarial Science, University of Iowa（爱荷华大学统计与精算科学系）
Convolutional neural networks have recently achieved significant breakthroughs in various image classification tasks. However, they are computationally expensive,which can make their feasible mplementation on embedded and low-power devices difficult. In this paper convolutional neural network binarization is implemented on GPU-based platforms for real-time inference on resource constrained devices. In binarized networks, all weights and intermediate computations between layers are quantized to +1 and -1, allowing multiplications and additions to be replaced with bit-wise operations between 32-…
Massachusetts Institute of Technology（麻省理工学院）；Princeton University（普林斯顿大学）
Sliding window convolutional networks (ConvNets) have become a popular approach to computer vision problems such as image segmentation, and object detection and localization. Here we consider the problem of inference, the application of a previously trained ConvNet, with emphasis on 3D images. Our goal is to maximize throughput, defined as average number of output voxels computed per unit time. Other things being equal, processing a larger image tends to increase throughput, because fractionally less computation is wasted on the borders of the image. It follows that an apparently slower algori…
We examine the performance profile of Convolutional Neural Network training on the current generation of NVIDIA Graphics Processing Units. We introduce two new Fast Fourier Transform convolution implementations: one based on NVIDIA's cuFFT library, and another based on a Facebook authored FFT implementation, fbfft, that provides significant speedups over cuFFT (over 1.5x) for whole CNNs. Both of these convolution implementations are available in open source, and are faster than NVIDIA's cuDNN implementation for many common convolutional layers (up to 23.5x for some synthetic kernel configurati…
Deep neural networks have achieved remarkable accuracy in many artificial intelligence applications, e.g. computer vision, at the cost of a large number of parameters and high computational complexity. Weight pruning can compress DNN models by removing redundant parameters in the networks, but it brings sparsity in the weight matrix, and therefore makes the computation inefficient on GPUs. Although pruning can remove more than 80% of the weights, it actually hurts inference performance (speed) when running models on GPUs. Two major problems cause this unsatisfactory performance on GPUs. First,…
The sophisticated structure of Convolutional Neural Network (CNN) allows for outstanding performance, but at the cost of intensive computation. As significant redundancies inevitably present in such a structure, many works have been proposed to prune the convolutional filters for computation cost reduction. Although extremely effective, most works are based only on quantitative characteristics of the convolutional filters, and highly overlook the qualitative interpretation of individual filter's specific functionality. In this work, we interpreted the functionality and redundancy of the convol…
Arizona State University（亚利桑那州立大学）；Nanyang Technological University（南洋理工大学）
FPGA-based hardware accelerators for convolutional neural networks (CNNs) have obtained great attentions due to their higher energy efficiency than GPUs. However, it is challenging for FPGA-based solutions to achieve a higher throughput than GPU counterparts. In this paper, we demonstrate that FPGA acceleration can be a superior solution in terms of both throughput and energy efficiency when a CNN is trained with binary constraints on weights and activations. Specifically, we propose an optimized FPGA accelerator architecture tailored for bitwise convolution and normalization that features mas…
Dimensionality-reduction techniques are a fundamental tool for extracting useful information from high-dimensional data sets. Because secant sets encode manifold geometry, they are a useful tool for designing meaningful data-reduction algorithms. In one such approach, the goal is to construct a projection that maximally avoids secant directions and hence ensures that distinct data points are not mapped too close together in the reduced space. This type of algorithm is based on a mathematical framework inspired by the constructive proof of Whitney's embedding theorem from differential topology.…
North Carolina State University（北卡罗来纳州立大学）；NEC Labs America（NEC美国实验室）
Leveraging large data sets, deep Convolutional Neural Networks (CNNs) achieve state-of-the-art recognition accuracy. Due to the substantial compute and memory operations, however, they require significant execution time. The massive parallel computing capability of GPUs make them as one of the ideal platforms to accelerate CNNs and a number of GPU-based CNN libraries have been developed. While existing works mainly focus on the computational efficiency of CNNs, the memory efficiency of CNNs have been largely overlooked. Yet CNNs have intricate data structures and their memory behavior can have…
University of Edinburgh（爱丁堡大学）；Trinity College Dublin（都柏林三一学院）；University of Glasgow（格拉斯哥大学）
Convolutional Neural Networks (CNN) are becoming a common presence in many applications and services, due to their superior recognition accuracy. They are increasingly being used on mobile devices, many times just by porting large models designed for server space, although several model compression techniques have been considered. One model compression technique intended to reduce computations is channel pruning. Mobile and embedded systems now have GPUs which are ideal for the parallel computations of neural networks and for their lower energy cost per operation. Specialized libraries perform…
The freedom of fast iterations of distributed deep learning tasks is crucial for smaller companies to gain competitive advantages and market shares from big tech giants. HorovodRunner brings this process to relatively accessible spark clusters. There have been, however, no benchmark tests on HorovodRunner per se, nor specifically graph convolutional network (GCN, hereafter), and very limited scalability benchmark tests on Horovod, the predecessor requiring custom built GPU clusters. For the first time, we show that Databricks' HorovodRunner achieves significant lift in scaling efficiency for t…
We present surrogate models for heat transfer and pressure drop prediction of complex fin geometries generated using composite Bezier curves. Thermal design process includes iterative high fidelity simulation which is complex, computationally expensive, and time-consuming. With the advancement in machine learning algorithms as well as Graphics Processing Units (GPUs), we can utilize the parallel processing architecture of GPUs rather than solely relying on CPUs to accelerate the thermo-fluid simulation. In this study, Convolutional Neural Networks (CNNs) are used to predict results of Computat…
Deep convolutional neural networks take GPU days of compute time to train on large data sets. Pedestrian detection for self driving cars requires very low latency. Image recognition for mobile phones is constrained by limited processing resources. The success of convolutional neural networks in these situations is limited by how fast we can compute them. Conventional FFT based convolution is fast for large filters, but state of the art convolutional neural networks use small, 3x3 filters. We introduce a new class of fast algorithms for convolutional neural networks using Winograd's minimal fil…
I present a new way to parallelize the training of convolutional neural networks across multiple GPUs. The method scales significantly better than all alternatives when applied to modern convolutional neural networks.
Graph Convolutional Networks (GCNs) have been proved successful in the field of semi-supervised node classification by extracting structural information from graph data. However, the random selection of labeled nodes used by GCNs may lead to unstable generalization performance of GCNs. In this paper, we propose an efficient method for the deterministic selection of labeled nodes: the Determinate Node Selection (DNS) algorithm. The DNS algorithm identifies two categories of representative nodes in the graph: typical nodes and divergent nodes. These labeled nodes are selected by exploring the st…
Full batch training of Graph Convolutional Network (GCN) models is not feasible on a single GPU for large graphs containing tens of millions of vertices or more. Recent work has shown that, for the graphs used in the machine learning community, communication becomes a bottleneck and scaling is blocked outside of the single machine regime. Thus, we propose MG-GCN, a multi-GPU GCN training framework taking advantage of the high-speed communication links between the GPUs present in multi-GPU systems. MG-GCN employs multiple High-Performance Computing optimizations, including efficient re-use of m…
College of Information and Computer Science（信息与计算机科学学院）；University of Massachusetts（马萨诸塞大学）
We present diffusion-convolutional neural networks (DCNNs), a new model for graph-structured data. Through the introduction of a diffusion-convolution operation, we show how diffusion-based representations can be learned from graph-structured data and used as an effective basis for node classification. DCNNs have several attractive qualities, including a latent representation for graphical data that is invariant under isomorphism, as well as polynomial-time prediction and learning that can be represented as tensor operations and efficiently implemented on the GPU. Through several experiments w…
In this work, we present a MCTS-based Go-playing program which uses convolutional networks in all parts. Our method performs MCTS in batches, explores the Monte Carlo search tree using Thompson sampling and a convolutional network, and evaluates convnet-based rollouts on the GPU. We achieve strong win rates against open source Go programs and attain competitive results against state of the art convolutional net-based Go-playing programs.
University of Augsburg（奥格斯堡大学）；University Hospital Augsburg（奥格斯堡大学医院）
Neural networks represent a field of research that can quickly achieve very good results in the field of medical image segmentation using a GPU. A possible way to achieve good results without GPUs are random forests. For this purpose, two random forest approaches were compared with a state-of-the-art deep convolutional neural network. To make the comparison the PhC-C2DH-U373 and the retinal imaging datasets were used. The evaluation showed that the deep convolutional neutral network achieved the best results. However, one of the random forest approaches also achieved a similar high performance…
The past several years have witnessed the success of transformer-based models, and their scale and application scenarios continue to grow aggressively. The current landscape of transformer models is increasingly diverse: the model size varies drastically with the largest being of hundred-billion parameters; the model characteristics differ due to the sparsity introduced by the Mixture-of-Experts; the target application scenarios can be latency-critical or throughput-oriented; the deployment hardware could be single- or multi-GPU systems with different types of memory and storage, etc. With suc…
Courant Institute of Mathematical Sciences（库朗数学科学研究所）；New York University（纽约大学）
Convolutional networks are one of the most widely employed architectures in computer vision and machine learning. In order to leverage their ability to learn complex functions, large amounts of data are required for training. Training a large convolutional network to produce state-of-the-art results can take weeks, even when using modern GPUs. Producing labels using a trained network can also be costly when dealing with web-scale datasets. In this work, we present a simple algorithm which accelerates training and inference by a significant factor, and can yield improvements of over an order of…
Inference of Convolutional Neural Networks in time critical applications usually requires a GPU. In robotics or embedded devices these are often not available due to energy, space and cost constraints. Furthermore, installation of a deep learning framework or even a native compiler on the target platform is not possible. This paper presents a neural network code generator (NNCG) that generates from a trained CNN a plain ANSI C code file that encapsulates the inference in single a function. It can easily be included in existing projects and due to lack of dependencies, cross compilation is usua…
The University of Texas at Austin（得克萨斯大学奥斯汀分校）；NVIDIA（英伟达）
Training convolutional neural networks (CNNs) requires intense compute throughput and high memory bandwidth. Especially, convolution layers account for the majority of the execution time of CNN training, and GPUs are commonly used to accelerate these layer workloads. GPU design optimization for efficient CNN training acceleration requires the accurate modeling of how their performance improves when computing and memory resources are increased. We present DeLTA, the first analytical model that accurately estimates the traffic at each GPU memory hierarchy level, while accounting for the complex …
Graph Neural Networks (GNNs) are a class of neural networks designed to extract information from the graphical structure of data. Graph Convolutional Networks (GCNs) are a widely used type of GNN for transductive graph learning problems which apply convolution to learn information from graphs. GCN is a challenging algorithm from an architecture perspective due to inherent sparsity, low data reuse, and massive memory capacity requirements. Traditional neural algorithms exploit the high compute capacity of GPUs to achieve high performance for both inference and training. The architectural decisi…
Deep convolutional neural networks have demonstrated promising performance on image classification tasks, but the manual design process becomes more and more complex due to the fast depth growth and the increasingly complex topologies of convolutional neural networks. As a result, neural architecture search has emerged to automatically design convolutional neural networks that outperform handcrafted counterparts. However, the computational cost is immense, e.g. 22,400 GPU-days and 2,000 GPU-days for two outstanding neural architecture search works named NAS and NASNet, respectively, which moti…
GPUs have limited memory and it is difficult to train wide and/or deep models that cause the training process to go out of memory. It is shown in this paper how an open source tool called Large Model Support (LMS) can utilize a high bandwidth NVLink connection between CPUs and GPUs to accomplish training of deep convolutional networks. LMS performs tensor swapping between CPU memory and GPU memory such that only a minimal number of tensors required in a training step are kept in the GPU memory. It is also shown how LMS can be combined with an MPI based distributed deep learning module to train…
Bern University of Applied Sciences（伯尔尼应用科学大学）；University of California Davis（加州大学戴维斯分校）
Convolutional neural networks (CNN) have achieved major breakthroughs in recent years. Their performance in computer vision have matched and in some areas even surpassed human capabilities. Deep neural networks can capture complex non-linear features; however this ability comes at the cost of high computational and memory requirements. State-of-art networks require billions of arithmetic operations and millions of parameters. To enable embedded devices such as smartphones, Google glasses and monitoring cameras with the astonishing power of deep learning, dedicated hardware accelerators can be …
Indian Institute of Technology Bombay（印度理工学院孟买分校）；Department of Electrical Engineering（电气工程系）；Mumbai, India（印度孟买）
Vision transformers (ViTs) have found only limited practical use in processing images, in spite of their state-of-the-art accuracy on certain benchmarks. The reason for their limited use include their need for larger training datasets and more computational resources compared to convolutional neural networks (CNNs), owing to the quadratic complexity of their self-attention mechanism. We propose a linear attention-convolution hybrid architecture -- Convolutional X-formers for Vision (CXV) -- to overcome these limitations. We replace the quadratic attention with linear attention mechanisms, such…
There has been significant progress in developing neural network architectures that both achieve high predictive performance and that also achieve high application-level inference throughput (e.g., frames per second). Another metric of increasing importance is GPU utilization during inference: the measurement of how well a deployed neural network uses the computational capabilities of the GPU on which it runs. Achieving high GPU utilization is critical to increasing application-level throughput and ensuring a good return on investment for deploying GPUs. This paper analyzes the GPU utilization…
Caffe provides multimedia scientists and practitioners with a clean and modifiable framework for state-of-the-art deep learning algorithms and a collection of reference models. The framework is a BSD-licensed C++ library with Python and MATLAB bindings for training and deploying general-purpose convolutional neural networks and other deep models efficiently on commodity architectures. Caffe fits industry and internet-scale media needs by CUDA GPU computation, processing over 40 million images a day on a single K40 or Titan GPU ($\approx$ 2.5 ms per image). By separating model representation fr…
École Nationale Supérieure d’Informatique（国立高等信息学院）；Université Polytechnique Hauts-de-France（上法兰西理工大学）
Running Convolutional Neural Network (CNN) based applications on edge devices near the source of data can meet the latency and privacy challenges. However due to their reduced computing resources and their energy constraints, these edge devices can hardly satisfy CNN needs in processing and data storage. For these platforms, choosing the CNN with the best trade-off between accuracy and execution time while respecting Hardware constraints is crucial. In this paper, we present and compare five (5) of the widely used Machine Learning based methods for execution time prediction of CNNs on two (2) …
The growth in the complexity of Convolutional Neural Networks (CNNs) is increasing interest in partitioning a network across multiple accelerators during training and pipelining the backpropagation computations over the accelerators. Existing approaches avoid or limit the use of stale weights through techniques such as micro-batching or weight stashing. These techniques either underutilize of accelerators or increase memory footprint. We explore the impact of stale weights on the statistical efficiency and performance in a pipelined backpropagation scheme that maximizes accelerator utilization…
School of Computing, Newcastle University（纽卡斯尔大学计算机学院）；Department of Computer Science, Durham University（杜伦大学计算机科学系）
With the growing significance of graphs as an effective representation of data in numerous applications, efficient graph analysis using modern machine learning is receiving a growing level of attention. Deep learning approaches often operate over the entire adjacency matrix -- as the input and intermediate network layers are all designed in proportion to the size of the adjacency matrix -- leading to intensive computation and large memory requirements as the graph size increases. It is therefore desirable to identify efficient measures to reduce both run-time and memory requirements allowing f…
As GPUs scale their low precision matrix math throughput to boost deep learning (DL) performance, they upset the balance between math throughput and memory system capabilities. We demonstrate that converged GPU design trying to address diverging architectural requirements between FP32 (or larger) based HPC and FP16 (or smaller) based DL workloads results in sub-optimal configuration for either of the application domains. We argue that a Composable On-PAckage GPU (COPAGPU) architecture to provide domain-specialized GPU products is the most practical solution to these diverging requirements. A C…
Autoregressive convolutional neural networks (CNNs) have been widely exploited for sequence generation tasks such as audio synthesis, language modeling and neural machine translation. WaveNet is a deep autoregressive CNN composed of several stacked layers of dilated convolution that is used for sequence generation. While WaveNet produces state-of-the art audio generation results, the naive inference implementation is quite slow; it takes a few minutes to generate just one second of audio on a high-end GPU. In this work, we develop the first accelerator platform~\textit{FastWave} for autoregres…
Training deep learning models is compute-intensive and there is an industry-wide trend towards hardware specialization to improve performance. To systematically benchmark deep learning platforms, we introduce ParaDnn, a parameterized benchmark suite for deep learning that generates end-to-end models for fully connected (FC), convolutional (CNN), and recurrent (RNN) neural networks. Along with six real-world models, we benchmark Google's Cloud TPU v2/v3, NVIDIA's V100 GPU, and an Intel Skylake CPU platform. We take a deep dive into TPU architecture, reveal its bottlenecks, and highlight valuabl…
University of Illinois at Urbana-Champaign（伊利诺伊大学厄巴纳-香槟分校）；Adobe Research（奥多比研究院）
The ability to train large-scale neural networks has resulted in state-of-the-art performance in many areas of computer vision. These results have largely come from computational break throughs of two forms: model parallelism, e.g. GPU accelerated training, which has seen quick adoption in computer vision circles, and data parallelism, e.g. A-SGD, whose large scale has been used mostly in industry. We report early experiments with a system that makes use of both model parallelism and data parallelism, we call GPU A-SGD. We show using GPU A-SGD it is possible to speed up training of large convo…
Hardware-Software Co-Design is a highly successful strategy for improving performance of domain-specific computing systems. We argue for the application of the same methodology to deep learning; specifically, we propose to extend neural architecture search with information about the hardware to ensure that the model designs produced are highly efficient in addition to the typical criteria around accuracy. Using the task of keyword spotting in audio on edge computing devices, we demonstrate that our approach results in neural architecture that is not only highly accurate, but also efficiently m…
Cloud GPU servers have become the de facto way for deep learning practitioners to train complex models on large-scale datasets. However, it is challenging to determine the appropriate cluster configuration---e.g., server type and number---for different training workloads while balancing the trade-offs in training time, cost, and model accuracy. Adding to the complexity is the potential to reduce the monetary cost by using cheaper, but revocable, transient GPU servers. In this work, we analyze distributed training performance under diverse cluster configurations using CM-DARE, a cloud-based mea…
University of California, Los Angeles（加利福尼亚大学洛杉矶分校）
Graph convolutional networks (GCNs) have been introduced to effectively process non-euclidean graph data. However, GCNs incur large amounts of irregularity in computation and memory access, which prevents efficient use of traditional neural network accelerators. Moreover, existing dedicated GCN accelerators demand high memory volumes and are difficult to implement onto resource limited edge devices. In this work, we propose LW-GCN, a lightweight FPGA-based accelerator with a software-hardware co-designed process to tackle irregularity in computation and memory access in GCN inference. LW-GCN d…
Sun Yat-sen University（中山大学）；The University of Hong Kong（香港大学）
Deep convolutional neural networks have become a key element in the recent breakthrough of salient object detection. However, existing CNN-based methods are based on either patch-wise (region-wise) training and inference or fully convolutional networks. Methods in the former category are generally time-consuming due to severe storage and computational redundancies among overlapping patches. To overcome this deficiency, methods in the second category attempt to directly map a raw input image to a predicted dense saliency map in a single network forward pass. Though being very efficient, it is a…
MatConvNet is an implementation of Convolutional Neural Networks (CNNs) for MATLAB. The toolbox is designed with an emphasis on simplicity and flexibility. It exposes the building blocks of CNNs as easy-to-use MATLAB functions, providing routines for computing linear convolutions with filter banks, feature pooling, and many more. In this manner, MatConvNet allows fast prototyping of new CNN architectures; at the same time, it supports efficient computation on CPU and GPU allowing to train complex models on large datasets such as ImageNet ILSVRC. This document provides an overview of CNNs and h…
Intensive computation is entering data centers with multiple workloads of deep learning. To balance the compute efficiency, performance, and total cost of ownership (TCO), the use of a field-programmable gate array (FPGA) with reconfigurable logic provides an acceptable acceleration capacity and is compatible with diverse computation-sensitive tasks in the cloud. In this paper, we develop an FPGA acceleration platform that leverages a unified framework architecture for general-purpose convolutional neural network (CNN) inference acceleration at a data center. To overcome the computation bound,…
The convolutional layers are core building blocks of neural network architectures. In general, a convolutional filter applies to the entire frequency spectrum of the input data. We explore artificially constraining the frequency spectra of these filters and data, called band-limiting, during training. The frequency domain constraints apply to both the feed-forward and back-propagation steps. Experimentally, we observe that Convolutional Neural Networks (CNNs) are resilient to this compression scheme and results suggest that CNNs learn to leverage lower-frequency components. In particular, we f…
In machine learning for fluid mechanics, fully-connected neural network (FNN) only uses the local features for modelling, while the convolutional neural network (CNN) cannot be applied to data on structured/unstructured mesh. In order to overcome the limitations of FNN and CNN, the unstructured convolutional neural network (UCNN) is proposed, which aggregates and effectively exploits the features of neighbour nodes through the weight function. Adjoint vector modelling is taken as the task to study the performance of UCNN. The mapping function from flow-field features to adjoint vector is const…
This paper describes a novel deep learning-based method for mitigating the effects of atmospheric distortion. We have built an end-to-end supervised convolutional neural network (CNN) to reconstruct turbulence-corrupted video sequence. Our framework has been developed on the residual learning concept, where the spatio-temporal distortions are learnt and predicted. Our experiments demonstrate that the proposed method can deblur, remove ripple effect and enhance contrast of the video sequences simultaneously. Our model was trained and tested with both simulated and real distortions. Experimental…
