The paper briefly describes reconfigurable architecture classifications and comparisons, as well as an FPGA implementation of a simplified version of the reconfigurable data path processor (RDPP) using Xilinx design t...
详细信息
The paper briefly describes reconfigurable architecture classifications and comparisons, as well as an FPGA implementation of a simplified version of the reconfigurable data path processor (RDPP) using Xilinx design tools. The FPGA configured as RDPP can be employed for a wide variety of signalprocessing applications. That is, the inherent parallel nature of the architecture lends itself to signal processing algorithms, which are discussed. Simulations of operations on the input data stream, which include multiply-and-accumulate (MAC) operations used in filters and matrix manipulation for image processing applications, are presented.
The objective of the presentation is to illustrate that the principle of analytic continuation can be utilized to interpolate/extrapolate either computational results or measured data in an efficient fashion instead o...
详细信息
The objective of the presentation is to illustrate that the principle of analytic continuation can be utilized to interpolate/extrapolate either computational results or measured data in an efficient fashion instead of utilizing sheer brute force.
This paper presents an execution controller generation methodology for data centric coarse grain data flows. This methodology assumes that processing blocks are designed and their execution characteristics are well de...
详细信息
This paper presents an execution controller generation methodology for data centric coarse grain data flows. This methodology assumes that processing blocks are designed and their execution characteristics are well defined. The overall controller generation is based on block level pipelining where concurrency among processing blocks is fully maintained. The methodology isolates controls for buffer, global sequencing, and logic such that system integration is simplified while controllers are locally configured from orthogonal global information.
This paper introduces TINA, a novel framework for implementing non Neural Network (NN) signal processing algorithms on NN accelerators such as GPUs, TPUs or FPGAs. The key to this approach is the concept of mapping ma...
详细信息
ISBN:
(数字)9798350372250
ISBN:
(纸本)9798350372267
This paper introduces TINA, a novel framework for implementing non Neural Network (NN) signal processing algorithms on NN accelerators such as GPUs, TPUs or FPGAs. The key to this approach is the concept of mapping mathematical and logic functions as a series of convolutional and fully connected layers. By mapping functions into such a small sub stack ofNN layers, it becomes possible to execute non-NN algorithms on NN hardware (HW) accelerators efficiently, as well as to ensure the portability of TINA implementations to any platform that supports such NN accelerators. Results show that TINA is highly competitive vs alternative frame-works, specifically for complex functions with iterations. For a Polyphase Filter Bank use case TINA shows GPU speedups of up to 80x vs a CPU baseline with NumPy compared to 8x speedup achieved by alternative frameworks. The frame-work is open source and publicly available at httPs://***/ChristiaanBoe/TINA.
The design and analysis of software which implements digital signalprocessing (DSP) algorithms on multiprocessor systems is examined. The DSP algorithms that are considered exhibit certain characteristics which are c...
详细信息
The design and analysis of software which implements digital signalprocessing (DSP) algorithms on multiprocessor systems is examined. The DSP algorithms that are considered exhibit certain characteristics which are common of DSP algorithms used in communication and control systems. The analysis methods developed use linear complexity algorithms to analyze the software that implements the DSP algorithms. The first step involves finding the repeat factors associated with a basic cycle of the algorithm. Using these repeat factors it is straightforward to (1) find the precedence relationships between tasks and finally a feasible schedule for the multiprocessor, and (2) coordinate and organize the efforts to optimize the software.< >
This paper presents a method for evaluating design complexity of a class of algorithms with characteristics that are common for many loop-based signalprocessing real-time applications. The method is not only used for...
详细信息
This paper presents a method for evaluating design complexity of a class of algorithms with characteristics that are common for many loop-based signalprocessing real-time applications. The method is not only used for evaluations, but can also be transformed to the actual implementation. The model transforms the data-flow structure of the algorithms to hierarchical pipelined architecture where control structure derivation is straightforward. The proposed method is used to estimate design complexity of two particle filtering algorithm: the sample importance resampling particle filter (SIRF) and the Gaussian Particle Filter (GPF) applied to the bearings-only tracking problem.
This work presents an approach at integrating novel methodologies for teaching graduate level courses in the areas of high performance computing (HPC) and advanced signal processing algorithms (ASPA) for computer engi...
详细信息
This work presents an approach at integrating novel methodologies for teaching graduate level courses in the areas of high performance computing (HPC) and advanced signal processing algorithms (ASPA) for computer engineering and computer science and engineering curricula. The novel teaching methodology presented here in high performance computing centers on the use of innovative empirical methods, i.e., exploratory data analysis, experiment design, etc., for studying computer performance, whereas an operator signal algebra approach is considered a novel methodology for the studying of advanced signal processing algorithms. The work also discusses an on going concerted effort at utilizing common tools and IT resources in both courses to provide students a holistic learning experience.
A review of work in the field of space-time signalprocessing (STSP), which takes into consideration complicated antenna motion, noises, and medium all together, is presented. The works led in three directions. The fi...
详细信息
A review of work in the field of space-time signalprocessing (STSP), which takes into consideration complicated antenna motion, noises, and medium all together, is presented. The works led in three directions. The first one is the development of the theory of STSP for moving antennas, the second one is the development of fast STSP algorithms realized complicated multichannel processing with a few calculating efforts, and the third one is the implementation of the results in sonar. A new approach for optimizing STSP in complicated dynamic conditions is proposed. By using it new STSP algorithms for coherent and stochastic signals are developed. Research of these algorithms shows that in some cases complicated antenna motion plays a great positive role. STSP systems with mobile antennas may have essentially higher space-time selectivity and noise immunity than systems with static antennas. These effects may be used for system improvement. In systems with line arrays they can be used, in particular, for alienating ambiguity in the estimation of signal direction. A series of fast STSP algorithms for different types of moving antennas are developed. The results are obtained both for coherent and for stochastic signals. It is found that in the case of parallel processing in millions of space-time channels the new fast algorithms decrease calculating efforts by a factor of ten times and more. The main positions of the theory are confirmed by experimental research.
The fading envelope encountered in narrowband mobile communications often dominates the impairments to communications usage of the channel. signalprocessing techniques are extensively used to mitigate the limitations...
详细信息
The fading envelope encountered in narrowband mobile communications often dominates the impairments to communications usage of the channel. signalprocessing techniques are extensively used to mitigate the limitations imposed by the fading, and many of the techniques require knowledge of the channel transfer function. The possibility of predicting the changing channel behaviour, in particular the position of the short-term fades, is therefore of interest for the subsequent communications signalprocessing. Different models of the multipath propagation encourage different algorithms for the model parameter estimation and channel behaviour prediction. These algorithms are reviewed and we present results from simulations showing the performance and limitations of prediction behaviour.
While the traditional deployment platform for digital hearing aids has been custom ICs, recent advances in the power efficiency of digital signal processors (DSPs) have generated interest in their use. A primary motiv...
详细信息
While the traditional deployment platform for digital hearing aids has been custom ICs, recent advances in the power efficiency of digital signal processors (DSPs) have generated interest in their use. A primary motivation for this interest is the increased algorithmic flexibility that a DSP provides over a fixed-function IC. The C54xx family of DSPs from Texas Instruments is among the leaders in power efficiency, and TI provides a development platform (the DHP-100) aimed at exploring the hearing aid design space. Here, we report on our experience porting hearing aid signal processing algorithms (J.L. Goldstein, August 2001) to the DHP-100 and report on the resulting performance, both for real-time response and power consumption.
暂无评论