Large-scale sound field rendering with graphics processing unit cluster for three-dimensional audio with loudspeaker array

Overlapping computation and communication of three-dimensional FDTD on a GPU cluster

Recently, general-purpose computation on graphics processing units (GPGPU) has become an increasingly popular field of study as graphics processing units (GPUs) continue to be proposed as high performance and relatively low cost implementation platforms for scientific computing applications. Among these applications figure astrophysical N-bodysimulations, which form one of the most challenging problems in computational science. However, in most reported studies, a simple $ \mathcal{O}(N^{2})$ algorithm was used for GPGPUs, and the resulting performances were not observed to be better than those of conventional CPUs that were based on more optimized $ \mathcal{O}(N \log N)$ algorithms such as the tree algorithm or the particle-particle particle-mesh algorithm. Because of the difficulty in getting efficient implementations of such algorithms on GPUs, a GPU cluster had no practical advantage over general-purpose PC clusters for N-bodysimulations. In this paper, we report a new method for efficient parallel implementation of the tree algorithm on GPUs. Our novel tree code allows the realization of an N-bodysimulation on a GPU cluster at a much higher performance than that on general PC clusters. We practically performed a cosmological simulation with 562 million particles on a GPU cluster using 128 NVIDIA GeForce 8800GTS GPUs at an overall cost of 168172 $. We obtained a sustained performance of 20.1 Tflops, which when normalized against a general-purpose CPU implementation leads to a performance of 8.50 Tflops. The achieved cost/performance was hence a mere $19.8 /Gflops which shows the high competitiveness of GPGPUs.

<strong class="journal-contentHeaderColor">Abstract.</strong> With semiconductor technology gradually approaching its physical and thermal limits, Graphics processing unit (GPU) is becoming an attractive solution in many scientific applications due to their high performance. This paper presents an application of GPU accelerators in air quality model. We endeavor to demonstrate an approach that runs a PPM solver of horizontal advection (HADVPPM) for air quality model CAMx on GPU clusters. Specifically, we first convert the HADVPPM to a new Compute Unified Device Architecture C (CUDA C) code to make it computable on the GPU (GPU-HADVPPM). Then, a series of optimization measures are taken, including reducing the CPU-GPU communication frequency, increasing the size of data computation on GPU, optimizing the GPU memory access, and using thread and block indices in order to improve the overall computing performance of CAMx model coupled with GPU-HADVPPM (named as CAMx-CUDA model). Finally, a heterogeneous, hybrid programming paradigm is presented and utilized with the GPU-HADVPPM on GPU clusters with Massage Passing Interface (MPI) and CUDA. Offline experiment results show that running GPU-HADVPPM on one NVIDIA Tesla K40m and NVIDIA Tesla V100 GPU can achieve up to 845.4x and 1113.6x acceleration. By implementing a series of optimization schemes, the CAMx-CUDA model resulted in a 29.0x and 128.4x improvement in computational efficiency using a GPU accelerator card on a K40m and V100 cluster, respectively. In terms of the single-module computational efficiency of GPU-HADVPPM, it can achieve 1.3x and 19.4x speedup on NVIDIA Tesla K40m GPU and NVIDA Tesla V100 GPU respectively. The multi-GPU acceleration algorithm enables 4.5x speedup with 8 CPU cores and 8 GPU accelerators on V100 cluster.

<strong class="journal-contentHeaderColor">Abstract.</strong> With semiconductor technology gradually approaching its physical and thermal limits, Graphics processing unit (GPU) is becoming an attractive solution in many scientific applications due to their high performance. This paper presents an application of GPU accelerators in air quality model. We endeavor to demonstrate an approach that runs a PPM solver of horizontal advection (HADVPPM) for air quality model CAMx on GPU clusters. Specifically, we first convert the HADVPPM to a new Compute Unified Device Architecture C (CUDA C) code to make it computable on the GPU (GPU-HADVPPM). Then, a series of optimization measures are taken, including reducing the CPU-GPU communication frequency, increasing the size of data computation on GPU, optimizing the GPU memory access, and using thread and block indices in order to improve the overall computing performance of CAMx model coupled with GPU-HADVPPM (named as CAMx-CUDA model). Finally, a heterogeneous, hybrid programming paradigm is presented and utilized with the GPU-HADVPPM on GPU clusters with Massage Passing Interface (MPI) and CUDA. Offline experiment results show that running GPU-HADVPPM on one NVIDIA Tesla K40m and NVIDIA Tesla V100 GPU can achieve up to 845.4x and 1113.6x acceleration. By implementing a series of optimization schemes, the CAMx-CUDA model resulted in a 29.0x and 128.4x improvement in computational efficiency using a GPU accelerator card on a K40m and V100 cluster, respectively. In terms of the single-module computational efficiency of GPU-HADVPPM, it can achieve 1.3x and 19.4x speedup on NVIDIA Tesla K40m GPU and NVIDA Tesla V100 GPU respectively. The multi-GPU acceleration algorithm enables 4.5x speedup with 8 CPU cores and 8 GPU accelerators on V100 cluster.

We present a set of algorithms for simulating and visualizing real-time crowds in GPU (Graphics Processing Units) clusters. First we will present crowd simulation and rendering techniques that take advantage of single GPU machines, then using as an example a wandering crowd behavior simulation algorithm, we explain how this kind of algorithms can be extended for their use in GPU cluster environments. We also present a visualization architecture that renders the simulation results using detailed 3D virtual characters. This architecture is adaptable in order to support the Barcelona Supercomputing Center (BSC) infrastructure. Results show that our algorithms are scalable in different hardware platforms including embedded systems, desktop GPUs and GPU clusters, in particular, the BSC's Minotauro cluster.

We evaluate a novel implementation of a Self-Organizing Map (SOM) on a Graphics Processing Unit (GPU) cluster. Using various combinations of OpenCL, CUDA, and two different graphics cards, we demonstrate the scalability of the SOM implementation on one to eight GPUs. Results indicate that while the algorithm scales well with the number of training samples and the map size, the benefits from using the data-parallel approaches offered by the GPU are severely limited when combined with the Message Passing Interface (MPI) in this setting, and comparable to speedups of GPU-based implementations as compared to optimized sequential code. Speedups achieved range from 3 to 32, for various map and training data sizes. We also observed a performance penalty for the OpenCL implementation as compared to CUDA.

The calculation of pairwise correlation coefficient on a dataset, known as the correlation matrix, is often used in data analysis, signal processing, pattern recognition, image processing, and bioinformatics. With the state-of-the-art Graphic Processing Units (GPUs) that consist of massive cores capable to do processing up to several Gflops, the calculation of correlation matrix can be accelerated several times over traditional CPUs. However, due to the rapid growth of the data in the digital era, the correlation matrix calculation becomes computing intensive which needs to be executed on multiple GPUs. As of now, GPUs are common components in data center at many institutions. Their GPU deployment tends towards a GPU cluster which each node is equipped with GPUs. In this paper, we propose a parallel computing based on the hybrid MPI/CUDA programming for fast and efficient Pearson correlation matrix calculation on GPU clusters. At coarse grain parallelism, the correlation matrix is partitioned into tiles which are distributed to execute concurrently on many GPUs using MPI. At fine grain level, the CUDA kernel function on each node performs massively parallel computing on a GPU. To balance load across all GPUs, we adopt the work pool model which there is a master node that manages tasks in the work pool and dynamically assign tasks to worker nodes. The result of the evaluation shows that the proposed work can ensure the load balance across different GPUs and thus gives better execution time than using a simple static data partitioning.

In this paper,a scalable graphics processing unit (GPU) cluster solution for the acceleration of FDTD for large-scale simulations is proposed. The hardware and software implementations are described. To illustrate the speed performance of the cluster, the simulation results of a cubic resonator with PEC boundaries is presented. A realistic large-scale simulation performed using SEMCAD X from SPEAG, accelerated by Acceleware's software libraries and GPU cluster solution is also presented.

It has been possible to reproduce point sound sources between listeners and a loudspeaker array by using the focused-source method. However, this method requires physical parameters of the sound sources to be reproduced, such as source positions, directions, and original signals. This fact makes it difficult to apply the method to real-time reproduction systems because decomposing received signals into such parameters is not a trivial task. This paper proposes a method for recreating virtual sound sources in front of a planar or linear loudspeaker array. The method is based on wave field synthesis but extended to include the inverse wave propagator often used in acoustical holography. Virtual sound sources can be placed between listeners and a loudspeaker array even when the received signals of a microphone array equally aligned with the loudspeaker array are only known. Numerical simulation results are presented to compare the proposed and focused-source methods. A system was constructed consisting of linear microphone and loudspeaker arrays and measurement experiments were conducted in an anechoic room. When comparing the sound field reproduced using the proposed method with that using the focused-source method, it was found that the proposed method could reproduce the sound field at almost the same accuracy.

In recent years, computation power to simulate sound fields from the three-dimensional numerical models has progressed fast, for example, using GPU cluster systems. However, the intervals among loudspeakers used in reproduction of the rendered sound field are insufficient to avoid spatial aliasing. Therefore, it is impossible to reproduce the rendered sound field accurately with current loudspeaker systems. Nevertheless, subjective impression of reproduction field with an insufficient system might be similar to intended sound field. The intended spatial attributes to render, such as sound position, sound directivity, reverberation, should be evaluated subjectively. For this report we performed a subjective evaluation of the sound fields rendered with a graphical processing unit (GPU) cluster. The sound fields were reproduced by a surrounding loudspeaker array. The number of loudspeakers in the array was 157. However the distance among loudspeakers was 50 cm and was insufficient to reproduce frequencies higher than 340 Hz. Listeners answered their subjective impressions on distance and directivity of sound source of a singing voice. We confirmed listeners could perceive the rendered distances in part when the directivity of the sound source was also rendered.

This paper presents an implementation of the brute-force exact k-Nearest Neighbor Graph (k-NNG) construction for ultra-large high-dimensional data cloud. The proposed method uses Graphics Processing Units (GPUs) and is scalable with multi-levels of parallelism (between nodes of a cluster, between different GPUs on a single node, and within a GPU). The method is applicable to homogeneous computing clusters with a varying number of nodes and GPUs per node. We achieve a 6-fold speedup in data processing as compared with an optimized method running on a cluster of CPUs and bring a hitherto impossible -NNG generation for a dataset of twenty million images with 15 k dimensionality into the realm of practical possibility.

Broadcast operations (e.g. MPI_Bcast) have been widely used in deep learning applications to exchange a large amount of data among multiple graphics processing units (GPUs). Recent studies have shown that leveraging the InfiniBand hardware-based multicast (IB-MCAST) protocol can enhance scalability of GPU-based broadcast operations. However, these initial designs with IB-MCAST are not optimized for multi-source broadcast operations with large messages, which is the common communication scenario for deep learning applications. In this paper, we first model existing broadcast schemes and analyze their performance bottlenecks on GPU clusters. Then, we propose a novel broadcast design based on message streaming to better exploit IB-MCAST and NVIDIA GPUDirect RDMA (GDR) technology for efficient large message transfer operation. The proposed design can provide high overlap among multi-source broadcast operations. Experimental results show up to 68% reduction of latency compared to state-of-the-art solutions in a benchmark-level evaluation. The proposed design also shows near-constant latency for a single broadcast operation as a system grows. Furthermore, it yields up to 24% performance improvement in the popular deep learning framework, Microsoft CNTK, which uses multi-source broadcast operations; notably, the performance gains are achieved without modifications to applications. Our model validation shows that the proposed analytical model and experimental results match within a 10% range. Our model also predicts that the proposed design outperforms existing schemes for multi-source broadcast scenarios with increasing numbers of broadcast sources in large-scale GPU clusters.

ABSTRACTThe mean shift image segmentation algorithm is very computation-intensive. To address the need to deal with a large number of remote sensing (RS) image segmentations in real-world applications, this study has investigated the parallelization of the mean shift algorithm on a single graphics processing unit (GPU) and a task-scheduling method with message passing interface (MPI)+OpenCL programming model on a GPU cluster platform. This paper presents the test results of the parallel mean shift image segmentation algorithm on Shelob, a GPU cluster platform at Louisiana State University, with different datasets and parameters. The experimental results show that the proposed parallel algorithm can achieve good speedups with different configurations and RS data and can provide an effective solution for RS image processing on a GPU cluster.

In order to investigate the effect of hall response on players, field measurements on the stage of a concert hall and laboratory experiment using digital simulation technique were performed. In the field experiment, the subject, a professional violinist, was asked to play and to make comments on her acoustical impression of five points on the stage. As a physical measurement, impulse responses were obtained at the same points by using omni-directional loudspeakers as a sound source and an omni-directional microphone and directional microphones as receivers. As a result, it has been found that not only the strength of the early reflections but also their direction influences the subject’s impression. In the laboratory experiment (anechoic chamber), the sound field was modeled and synthesized by using a 13 channel reproduction system; ambient reverberation judged as being natural was provided by simple digital reverberators and different strength and direction of early reflections were obtained by real-time convolvers. For a constant value of reverberation, several conditions with a different level and direction of the early reflections were created. For each condition, the violin player was asked to make similar judgments as in the field experiment. The results of two experiments were examined.

Late reverberation involves the superposition of many sound reflections, approaching the properties of a diffuse sound field. Since the spatially resolved perception of individual late reflections is impossible, simplifications can potentially be made for modelling late reverberation in room acoustics simulations with reduced spatial resolution. Such simplifications are desired for interactive, real-time virtual acoustic environments with applications in hearing research and for the evaluation of hearing supportive devices. In this context, the number and spatial arrangement of loudspeakers used for playback additionally affect spatial resolution. The current study assessed the minimum number of spatially evenly distributed virtual late reverberation sources required to perceptually approximate spatially highly resolved isotropic and anisotropic late reverberation and to technically approximate a spherically isotropic sound field. The spatial resolution of the rendering was systematically reduced by using subsets of the loudspeakers of an 86-channel spherical loudspeaker array in an anechoic chamber, onto which virtual reverberation sources were mapped using vector base amplitude panning. It was tested whether listeners can distinguish lower spatial resolutions of reproduction of late reverberation from the highest achievable spatial resolution in different simulated rooms. The rendering of early reflections remained unchanged. The coherence of the sound field across a pair of microphones at ear and behind-the-ear hearing device distance was assessed to separate the effects of number of virtual sources and loudspeaker array geometry. Results show that between 12 and 24 reverberation sources are required for the rendering of late reverberation in virtual acoustic environments.