虚拟示波器快速捕获技术研究及应用

发布时间：2018-03-10 11:05

本文选题：虚拟示波器　切入点：波形捕获率　出处：《电子科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：由于多元化和高效化检测的需要,虚拟示波器应用需求逐年增多。不同于直接通过硬件处理的台式示波器,虚拟示波器需要通过PC机进行分析处理,所以在处理波形数据的耗时上,虚拟示波器比台式示波器长,导致虚拟示波器的死区时间较长,波形捕获率低。因为波形捕获率决定了示波器捕获波形完整信息的能力,所以提高虚拟示波器的波形捕获能力具有较大的理论研究和实用价值。提高虚拟示波器波形捕获能力的方法主要有两种,即硬件升级和软件优化。本文主要研究使用软件提升波形捕获率的方法。通过设计一种虚拟示波器软件优化方法,加快程序数据处理以及绘制波形的速率,减少死区时间,提高波形捕获率,最大限度还原真实信号内容。并为其添加波形快速定位功能,在高速捕获模式下能更好地观查被测波形。本文主要通过绘图优化和并行化的方法提高虚拟示波器的波形捕获能力:(1)在绘图方法上,通过调用Direct2D绘图编程接口,借助现在性能不断提升的图形处理器,使Direct2D的硬件加速效果更加明显,从而提高虚拟示波器的波形绘制速率,使波形刷新率提高。并且还对波形数据进行批量处理,缩短死区时间,显示更多波形细节。最后通过波形定位功能找出需要的单幅波形并显示出来,方便对波形数据的检测。(2)在数据处理及任务划分上,利用Intel多核处理器的特点,将虚拟示波器程序进行并行化改造,使之能在多核处理器上并行执行各个任务,缩短任务执行时间,提升程序运行速率。并行化改造主要利用Intel线程构造块TBB(Intel Thread Building Blocks),它能实现多核处理器下的串行程序的并行化,以及通过Intel Parallel Studio工具集对并行程序进行调试及调优。缩短程序中数据处理时间,达到提高波形捕获能力的目的。通过对软件优化的研究及实施,实现虚拟示波器在相同的硬件配置及操作环境下波形捕获率的提升。本文研究的软件优化方法适用于不同核心数的处理器,在双核处理器平台上其波形捕获率可提高1.33倍,在四核处理器的平台上优化带来的加速比能高达1.7。
[Abstract]:The demand for virtual oscilloscope application is increasing year by year because of the need of multiplex and high efficiency detection. Different from desktop oscilloscope which is processed directly by hardware, virtual oscilloscope needs to be analyzed and processed by PC. Therefore, in the processing of waveform data, the virtual oscilloscope is longer than the desktop oscilloscope, which leads to the longer dead time of virtual oscilloscope and the lower waveform capture rate, because the waveform acquisition rate determines the ability of oscilloscope to capture the complete waveform information. Therefore, improving the waveform acquisition ability of virtual oscilloscope has great theoretical research and practical value. There are two main methods to improve the waveform acquisition ability of virtual oscilloscope. This paper mainly studies the method of using software to improve the waveform capture rate. By designing a software optimization method of virtual oscilloscope, the speed of program data processing and waveform drawing is accelerated, and the dead time is reduced. Improve the waveform capture rate, restore the real signal content to the maximum extent, and add the waveform fast localization function for it, In the high speed capture mode, we can better observe and check the measured waveforms. In this paper, we improve the waveform acquisition ability of virtual oscilloscopes by drawing optimization and parallelization. In drawing method, we call the Direct2D drawing programming interface. With the help of the GPU, the hardware acceleration effect of Direct2D is more obvious, thus the waveform rendering rate of virtual oscilloscope is improved, the waveform refresh rate is improved, and the waveform data is processed in batches. Finally, through the waveform localization function to find out the needed single waveform and display it, it is convenient to detect the waveform data in data processing and task division, using the characteristics of Intel multi-core processor. The virtual oscilloscope program is parallelized so that it can execute each task in parallel on the multi-core processor and shorten the task execution time. The parallelization transformation mainly uses Intel thread to construct TBB(Intel Thread Building locks, which can realize the parallelization of serial program under multi-core processor. The parallel program is debugged and optimized by the Intel Parallel Studio toolset. The data processing time in the program is shortened, and the ability of waveform acquisition is improved. The waveform acquisition rate of virtual oscilloscope is improved under the same hardware configuration and operating environment. The software optimization method studied in this paper is suitable for processors with different core numbers, and the waveform capture rate can be increased by 1.33 times on the dual-core processor platform. Optimization on a four-core processor platform results in a speedup of up to 1. 7.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TM935.3

【相似文献】