Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Heterogeneous platform neuron positioning three-stage pipeline parallel method, system and medium

A heterogeneous platform and neuron technology, applied in the field of analysis of the fine structure of neural circuits, can solve problems such as technical solutions without neuron positioning, affecting program operating efficiency, and occupying program running time, so as to achieve easy transplantation and promotion, program The effect of shortening the total execution time and increasing the processing speed

Active Publication Date: 2019-08-16
NAT UNIV OF DEFENSE TECH
View PDF6 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Frequent data transmission between CPU and GPU takes up a lot of program running time, which greatly affects the running efficiency of the program
There is no public report on the technical solution of using CPU-GPU for neuron localization

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
  • Heterogeneous platform neuron positioning three-stage pipeline parallel method, system and medium
  • Heterogeneous platform neuron positioning three-stage pipeline parallel method, system and medium
  • Heterogeneous platform neuron positioning three-stage pipeline parallel method, system and medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The following will take a server equipped with a dual-channel twelve-core 2.4GHz CPU and an NVIDIA GTX 1080Ti GPU as an example of a heterogeneous platform to further describe the three-stage pipeline parallel method, system and medium for neuron positioning on a heterogeneous platform according to the present invention. The hard disk capacity of the server is 24TB, the memory capacity is 256GB, and the GPU memory space is 11GB. The input data consists of 10,000 single-layer image sequences with a resolution of 40,000 × 40,000.

[0039] Such as figure 1 As shown, the steps of the three-stage pipeline parallel method for neuron positioning on heterogeneous platforms in this embodiment include:

[0040] 1) Calculate the block parameters of the sliced ​​image data according to the image size and calculation granularity;

[0041] 2) Based on the block parameters, the storage space is allocated on the CPU side and the GPU side respectively;

[0042] 3) variable and storag...

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 heterogeneous platform neuron positioning three-stage pipeline parallel method and system and a medium. According to the method, slice image data are calculated into block parameters according to the image size and the calculation granularity; storage space allocation is performed on the CPU end and the GPU end respectively based on the partitioning parameters; variablesand storage space are initialized; task scheduling is carried out by CPU, a three-level assembly line mode is adopted by a CPU and a GPU at the same time to execute a calculation task, wherein each calculation task comprises three steps of parallel data reading, positioning calculation and data writing-back, and each round of calculation task in the middle executes positioning calculation while executing data writing-back of the previous round of calculation task, and executes data reading-in of the next round of calculation task. The processing speed of neuron positioning can be improved, and the method has the advantages of high neuron positioning speed, short program execution total time, flexible three-level assembly line implementation, parameter configuration supporting and easinessin transplantation and popularization.

Description

technical field [0001] The present invention relates to a method for analyzing the fine structure of a neural circuit, in particular to a three-stage pipeline parallel method, system and medium for neuron positioning on a heterogeneous platform, which are used to realize parallel computing of neuron positioning based on a CPU-GPU heterogeneous computing platform. Background technique [0002] Neural circuit information is the key to understanding brain function and the mechanism of brain diseases. How to realize the automatic tracking of neural circuit big data is one of the key scientific problems faced by research in the field of neurology such as brain science. Neuron localization is the key to the analysis of neural circuit data. Accurate location of neuron cell bodies is obtained through the analysis of neural circuit image data, which is the basis for subsequent quantitative analysis. [0003] The typical large-scale neuron localization method is based on the biologica...

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 Applications(China)
IPC IPC(8): G06N3/06G06F9/38
CPCG06F9/3877G06N3/061
Inventor 邹丹朱小谦朱敏王文珂李金才汪祥陆丽娜甘新标孟祥飞夏飞
Owner NAT UNIV OF DEFENSE TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products