SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Admittance values for infinite dipole antenna immersed in magnetoplasma obtained from numerical integration of Fourier integral for antenna current
No abstract available
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Color Doppler flow mapping has played an important role in clinical echocardiography. Most of the clinical work, however, has been primarily qualitative. Although qualitative information is very valuable, there is considerable quantitative information stored within the velocity map that has not been extensively exploited so far. Recently, many researchers have shown interest in using the encoded velocities to address the clinical problems such as quantification of valvular regurgitation, calculation of cardiac output, and characterization of ventricular filling. In this article, we review some basic physics and engineering aspects of color Doppler echocardiography, as well as drawbacks of trying to retrieve velocities from video tape data. Digital storage, which plays a critical role in performing quantitative analysis, is discussed in some detail with special attention to velocity encoding in DICOM 3.0 (medical image storage standard) and the use of digital compression. Lossy compression can considerably reduce file size with minimal loss of information (mostly redundant); this is critical for digital storage because of the enormous amount of data generated (a 10 minute study could require 18 Gigabytes of storage capacity). Lossy JPEG compression and its impact on quantitative analysis has been studied, showing that images compressed at 27:1 using the JPEG algorithm compares favorably with directly digitized video images, the current goldstandard. Some potential applications of these velocities in analyzing the proximal convergence zones, mitral inflow, and some areas of future development are also discussed in the article.
This paper proposes a new method that combines checkpointing methods with error-controlled lossy compression for large-scale high-performance full-waveform inversion (FWI), an inverse problem commonly used in geophysical exploration. This combination can significantly reduce data movement, allowing a reduction in run time as well as peak memory. In the exascale computing era, frequent data transfer (e.g., memory bandwidth, PCIe bandwidth for GPUs, or network) is the performance bottleneck rather than the peak FLOPS of the processing unit. Like many other adjoint-based optimization problems, FWI is costly in terms of the number of floating-point operations, large memory footprint during backpropagation, and data transfer overheads. Past work for adjoint methods has developed checkpointing methods that reduce the peak memory requirements during backpropagation at the cost of additional floating-point computations. Combining this traditional checkpointing with error-controlled lossy compression, we explore the three-way tradeoff between memory, precision, and time to solution. We investigate how approximation errors introduced by lossy compression of the forward solution impact the objective function gradient and final inverted solution. Empirical results from these numerical experiments indicate that high lossy-compression rates (compression factors ranging up to 100) have a relatively minor impact on convergence rates and the quality of the final solution.
Over the last nearly 20 years, lossy compression has become an essential aspect of HPC applications’ data pipelines, allowing them to overcome limitations in storage capacity and bandwidth and, in some cases, increase computational throughput and capacity. However, with the adoption of lossy compression comes the requirement to assess and control the impact lossy compression has on scientific outcomes. In this work, we take a major step forward in describing the state of practice and by characterizing workloads. We examine applications’ needs and compressors’ capabilities across 9 different supercomputing application domains. We present 24 takeaways that provide best practices for applications, operational impacts for facilities achieving compressed data, and gaps in application needs not addressed by production compressors that point towards opportunities for future compression research.
Fast reliable data compression is urgently needed for many leading-edge scientific instruments and for exascale high-performance computing applications because they produce vast amounts of data at extremely high rates. The goal of this project has been to develop a framework named LC that is able to automatically generate high-speed lossless and reliable lossy compression and decompression algorithms that can be customized for different kinds of data. The resulting LC framework is freely available on GitHub. To achieve high-speed operation, LC outputs optimized and parallelized CPU and GPU implementations of the generated algorithms. To ensure the quality of lossily compressed data, LC guarantees the user-provided error bound. To be able to customize the compression algorithm to various use cases, LC can synthesize millions of different algorithms and automatically search for the one that works best for the given data. We have already employed LC to create state-of-the-art lossless and lossy compressors for scientific data as well as leading lossless compressors for images. We hope that LC and the customized, fast, reliable, and CPU/GPU-compatible compression algorithms that it can generate will greatly benefit the many scientific applications that need not only high trustworthiness but also high performance.
Numerical values for admittance of infinite cylindrical antenna in lossy, isotropic, compressible plasma
Error-bounded lossy compression is becoming an indispensable technique for the success of today's scientific projects with vast volumes of data produced during the simulations or instrument data acquisitions. Not only can it significantly reduce data size, but it also can control the compression errors based on user-specified error bounds. Autoencoder (AE) models have been widely used in image compression, but few AE-based compression approaches support error-bounding features, which are highly required by scientific applications. To address this issue, we explore using convolutional autoencoders to improve error-bounded lossy compression for scientific data, with the following three key contributions. (1) We provide an in-depth investigation of the characteristics of various autoencoder models and develop an error-bounded autoencoder-based framework in terms of the SZ model. (2) We optimize the compression quality for main stages in our designed AE-based error-bounded compression framework, fine-tuning the block sizes and latent sizes and also optimizing the compression efficiency of latent vectors. (3) We evaluate our proposed solution using five real-world scientific datasets and comparing them with six other related works. Experiments show that our solution exhibits a very competitive compression quality from among all the compressors in our tests. In absolute terms, it can obtain a much better compression quality (100%similar to 800% improvement in compression ratio with the same data distortion) compared with SZ2.1 and ZFP in cases with a high compression ratio.
HPC trends favor algorithms and implementations that reduce data motion relative to FLOPS. We investigate the use of lossy compressed data arrays in place of traditional IEEE floating point arrays to store the primary data of calculations. Simulation is fundamentally an exercise in controlled approximation, and error introduced by finite-precision arithmetic (or lossy compression) is just one of several sources of error that need to be managed to ensure sufficient accuracy in a computed result. We describe ZFP, a compressed numerical format designed for in-memory storage of multidimensional arrays, and summarize theoretical results that demonstrate that the error of repeated lossy compression can be bounded and controlled. Furthermore, we establish a relationship between grid resolution and compression-induced errors and show that, contrary to conventional floating point, ZFP reduces finite-difference errors with finer grids. We present example calculations that demonstrate data reduction by 4x or more with negligible impact on solution accuracy. Our results further demonstrate several orders-of-magnitude increase in accuracy using ZFP over IEEE floating point and Posits for the same storage budget.
Cross-Silo federated learning is widely used for scaling deep neural network (DNN) training over data silos from different locations worldwide while guaranteeing data privacy. Communication has been identified as the main bottleneck when training large-scale models due to large-volume model parameters and gradient transmission across public networks with limited bandwidth. Most previous works focus on gradient compression, while limited work tries to compress parameters that can not be ignored and extremely affect communication performance during the training. Here, to bridge this gap, we propose FedCSpc: an efficient cross-silo federated learning system with an XAI-driven adaptive parameter compression strategy for large-scale model training. Our work substantially differs from existing gradient compression techniques due to the distinct data features of gradient and parameter. The key contributions of this paper are fourfold. (1) Our designed FedCSpc proposes to compress the parameter during the training using the state-of-the-art error-bounded lossy compressor – SZ3. (2) We develop an adaptive compression error bound adjustment algorithm to guarantee the model accuracy effectively. (3) We exploit an efficient approach to utilize the idle CPU resources of clients to compress the parameters. (4) We perform a comprehensive evaluation with a wide range of models and benchmarks on a GPU cluster with 65 GPUs. Results show that FedCSpc can achieve the same model accuracy as FedAvg while reducing the data volume of parameters and gradients in communication by up to 7.39× and 288×, respectively. With 32 clients on a 4 Gb size model, FedCSpc significantly outperforms FedAvg in wall-clock time in the emulated WAN environment (at the bandwidth of 1 Gbps or lower without loss of generality).
During this reporting period we have worked on three somewhat different problems. These are modeling of video traffic in packet networks, low rate video compression, and the development of a lossy + lossless image compression algorithm, which might have some application in browsing algorithms. The lossy + lossless scheme is an extension of work previously done under this grant. It provides a simple technique for incorporating browsing capability. The low rate coding scheme is also a simple variation on the standard discrete cosine transform (DCT) coding approach. In spite of its simplicity, the approach provides surprisingly high quality reconstructions. The modeling approach is borrowed from the speech recognition literature, and seems to be promising in that it provides a simple way of obtaining an idea about the second order behavior of a particular coding scheme. Details about these are presented.
Integrating neural networks for data compression directly in the Read-Out Integrated Circuits (ROICs), i.e. the pixelated front-end, would result in a significant reduction in off-chip data transfer, overcoming the I/O bottleneck. Our ROIC test chip (AI-In-Pixel-65) is designed in a 65 nm Low Power CMOS process for the readout of pixelated X-ray detectors. Each pixel consists of an analog front-end for signal processing and a 10b analog-to-digital converter operating at 100KSPS. Here, we compare two non-reconfigurable techniques, Principal Component Analysis (PCA) and an AutoEncoder (AE) as lossy data compression engines implemented within the pixelated area. The PCA algorithm achieves 50$×$ compression, adds one clock cycle latency, and results in a 21% increase in the pixel area. The AE achieves 70$×$ compression, adds 30 clock cycle latency, and results in a similar area increase.
Future space-based, remote sensing systems will have data transmission requirements that exceed available downlinks necessitating the use of lossy compression techniques for multispectral data. In this paper, we describe several algorithms for lossy compression of multispectral data which combine spectral decorrelation techniques with an adaptive, wavelet-based, image compression algorithm to exploit both spectral and spatial correlation. We compare the performance of several different spectral decorrelation techniques including wavelet transformation in the spectral dimension. The performance of each technique is evaluated at compression ratios ranging from 4:1 to 16:1. Performance measures used are visual examination, conventional distortion measures, and multispectral classification results. We also introduce a family of distortion metrics that are designed to quantify and predict the effect of compression artifacts on multi spectral classification of the reconstructed data.
Astronomical images have some rather unusual characteristics that make many existing image compression techniques either ineffective or inapplicable. A typical image consists of a nearly flat background sprinkled with point sources and occasional extended sources. The images are often noisy, so that lossless compression does not work very well; furthermore, the images are usually subjected to stringent quantitative analysis, so any lossy compression method must be proven not to discard useful information, but must instead discard only the noise. Finally, the images can be extremely large. For example, the Space Telescope Science Institute has digitized photographic plates covering the entire sky, generating 1500 images each having 14000 x 14000 16-bit pixels. Several astronomical groups are now constructing cameras with mosaics of large CCD's (each 2048 x 2048 or larger); these instruments will be used in projects that generate data at a rate exceeding 100 MBytes every 5 minutes for many years. An effective technique for image compression may be based on the H-transform (Fritze et al. 1977). The method that we have developed can be used for either lossless or lossy compression. The digitized sky survey images can be compressed by at least a factor of 10 with no noticeable losses in the astrometric and photometric properties of the compressed images. The method has been designed to be computationally efficient: compression or decompression of a 512 x 512 image requires only 4 seconds on a Sun SPARCstation 1. The algorithm uses only integer arithmetic, so it is completely reversible in its lossless mode, and it could easily be implemented in hardware for space applications.
New higher-count-rate, integrating, large-area X-ray detectors with framing rates as high as 17400 images per second are beginning to be available. These will soon be used for specialized macromolecular crystallography experiments but will require optimal lossy compression algorithms to enable systems to keep up with data throughput. Some information may be lost. Can we minimize this loss with acceptable impact on structural information? To explore this question, we have considered several approaches: summing short sequences of images, binning to create the effect of larger pixels, use of JPEG-2000 lossy wavelet-based compression, and use of Hcompress, which is a Haar-wavelet-based lossy compression borrowed from astronomy. We also explore the effect of the combination of summing, binning, and Hcompress or JPEG-2000. In each of these last two methods one can specify approximately how much one wants the result to be compressed from the starting file size. These provide particularly effective lossy compressions that retain essential information for structure solution from Bragg reflections.
The enormous size of the data holding and the complexity of the information system resulting from the EOS system pose several challenges to computer scientists, one of which is data archival and dissemination. More than ninety percent of the data holdings of NASA is in the form of images which will be accessed by users across the computer networks. Accessing the image data in its full resolution creates data traffic problems. Image browsing using a lossy compression reduces this data traffic, as well as storage by factor of 30-40. Of the several image compression techniques, VQ is most appropriate for this application since the decompression of the VQ compressed images is a table lookup process which makes minimal additional demands on the user's computational resources. Lossy compression of image data needs expert level knowledge in general and is not straightforward to use. This is especially true in the case of VQ. It involves the selection of appropriate codebooks for a given data set and vector dimensions for each compression ratio, etc. A planning and scheduling system is described for using the VQ compression technique in the data access and ingest of raw satellite data.