Unlock instant, AI-driven research and patent intelligence for your innovation.

Parallel generation method of pulse compression reference matrix based on GPU (Graphic Processing Unit) core platform

A technology of pulse compression and reference matrix, which is applied in radio wave reflection/re-radiation, image data processing, measurement devices, etc., can solve problems such as low operating efficiency, low integration, and poor algorithm parallelism, and achieve improved operating efficiency , Improve operating efficiency and reduce computing time

Inactive Publication Date: 2013-07-31
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

At present, the generation of the reference matrix adopts the method of double loop, which is serially generated on the CPU, but the integration degree of the CPU is relatively low, the parallelism of the algorithm is relatively poor, and the operation efficiency is relatively low, and the GPGPU is a processing-intensive data processing method. Processors that can contain hundreds or thousands of processing cores and parallel data are more suitable for large-scale parallel computing
However, the generation of pulse compression reference matrices is not currently done in parallel on hundreds of processing cores

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Parallel generation method of pulse compression reference matrix based on GPU (Graphic Processing Unit) core platform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0020] Taking the generation of the azimuth pulse compression reference matrix for 4096×4096 point target imaging in the SAR system as an example, the GPU platform is NVIDIA’s GPU, model quadro fx 5600, and the implementation of this patent mainly includes the following processes:

[0021] 1. First write the string constants to be used as a header file, identified as RefMatrixConst.h, and include this file in the main program;

[0022] 2. Calculate the size of the storage space: In this system application, the data type of the matrix element is a floating point number, then the storage space occupied by the one-dimensional array generated by the GPU is: mem_size=sizeof(float)*4096*4096=67108864b ytes;

[0023] 3. Allocate one-dimensional arrays of mem_size in CPU memory and GPU global memory respectively: CPU real part RefMatrix_Re_h[16777216], CPU imaginary part RefMatrix_Im_h[16777216], GPU real part RefMatrix_Re_d[16777216], GPU imaginary part RefMatrix_Im_d[16777216];

[0...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a parallel generation method of a pulse compression reference matrix based on a core platform. The parallel generation method comprises the following steps of: realizing the paralleling of the pulse compression reference matrix on a GPU (Graphic Processing Unit) through parallel programming, generating a used temporary independent variable in a shared memory of each threadblock, directly finding the independent variable in the shared memory when in calculation without fetching data in a global memory, fully utilizing a structure of an inside shared memory, ensuring that all elements are successfully generated through a synchronous mechanism of the GPU and then writing a result back to a CPU (Central Processing Unit) internal memory. By adopting the strategy which combining the mass calculation amount and single communication, the invention greatly reduces the communication consumption, improves the parallelism degree of a program, ensures the great improvementof an operating efficiency and significantly reduces the calculation time.

Description

technical field [0001] The invention relates to a parallel generation method of a pulse compression reference matrix based on a GPU many-core platform and its application in engineering practice. Background technique [0002] Synthetic aperture radar is a two-dimensional microwave remote sensing imaging radar. This radar has the characteristics of long-distance all-weather high-resolution imaging, automatic target identification and advanced digital processing capabilities, and the high resolution is the combination of azimuth and direction. The result of the aperture technology, on the other hand, is the result of the pulse compression in the range direction. It can be seen that the maturity and accuracy of the pulse compression technology greatly affect the quality of the final SAR imaging. In the pulse compression technique, the generation of the pulse compression reference matrix is ​​an important step and often takes up a large part of the time of the SAR imaging algori...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T1/00G06T1/20G06T1/60G01S13/90
Inventor 姚迪龙腾靳星星刘峰
Owner BEIJING INSTITUTE OF TECHNOLOGYGY