Scientific computation-oriented high performance DMA (Direct Memory Access) part in GPDSP (General-Purpose Digital Signal Processor)

A scientific computing, high-performance technology, applied in the direction of instrumentation, electrical digital data processing, etc., can solve the problems of increasing the load of the on-chip interconnection network, reducing the efficiency of DDR access, reducing the speed of data transmission, etc., to avoid system loop blocking, Avoid frequent page turning and improve execution efficiency

Active Publication Date: 2016-03-09
NAT UNIV OF DEFENSE TECH
View PDF4 Cites 19 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

When multiple cores access (read or write) a peripheral at the same time, it will increase the load on the on-chip interconnection network and reduce the speed of data transmission
At the same time, if the peripheral is double-rate synchronous dy

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
  • Scientific computation-oriented high performance DMA (Direct Memory Access) part in GPDSP (General-Purpose Digital Signal Processor)
  • Scientific computation-oriented high performance DMA (Direct Memory Access) part in GPDSP (General-Purpose Digital Signal Processor)
  • Scientific computation-oriented high performance DMA (Direct Memory Access) part in GPDSP (General-Purpose Digital Signal Processor)

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0039] Such as figure 1 Shown is a schematic diagram of the architecture of the microprocessor in a specific application example of the present invention. The multi-core microprocessor includes a DSP supernode, an on-chip interconnect network and peripherals. Among them, each DSP super node is composed of a Matrix2 core, and the on-chip interconnection network completes the data communication between each DSP super node, and between the DSP super node and the peripheral, and the peripheral is DDR.

[0040] Such as figure 2Shown is a schematic diagram of the position of the DMA component of the present invention in the Matrix2 kernel. The DMA part is connected with peripheral hardware configuration bus PBUS, vector memory VM (VectorMemory, VM) and scalar memory SM (ScalarMemory, SM) in the Matrix2 core. Among them, PBUS is used to...

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 present invention discloses a scientific computation-oriented high performance DMA (Direct Memory Access) part in a GPDSP (General-Purpose Digital Signal Processor). The scientific computation oriented high performance DMA part comprises: a host portion for completing read/write request calculation according to a configured transmission parameter, and comprising a logical channel controller and two general physical channels; a slave computer portion for simultaneously providing a channel for the DMA to return out-of-kernel storage volume data read by the host to an in-kernel storage volume and providing a channel for a out-of-kernel device to read/write the in-kernel storage volume, and comprising an FIFO (First In, First Out) for buffering a out-of-kernel read/write request, a special channel arbiter, a matrix transpose module, a read/write AM special channel and a read/write SM special channel, wherein when an in-kernel request flows out of the read/write request FIFO, the special channel arbiter sends the request to different modules according to a mark signal and a target address of a request packet; and a bus controller, wherein both the host portion and the slave computer portion are connected to the bus controller. The scientific computation-oriented high performance DMA part in the GPDSP provided by the present invention has the advantages of supporting multiple types of transmission modes, improving the data transmission speed, improving the program execution efficiency and the like.

Description

technical field [0001] The present invention mainly relates to the field of direct memory access components (DirectMemoryAccess, DMA) of multi-core general-purpose digital processors (General-Purpose Digital Signal Processor, referred to as GPDSP), and specifically refers to a high-performance DMA component for scientific computing in GPDSP, which can support A variety of special transmission modes are used to increase the data transmission speed between the source and destination storage banks in GPDSP and improve the program execution efficiency. Background technique [0002] GPDSP is mainly used in high-performance applications such as scientific computing. The amount of data to be processed in these fields is very large, and the data processing capability of GPDSP is extremely high. In order to meet the computing needs, GPDSP generally integrates multiple cores, and each core has more parallel computing units. These computing units read the source data from the periphera...

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
IPC IPC(8): G06F13/30
CPCG06F13/30G06F2213/28
Inventor 陈书明马胜万江华郭阳刘宗林孙书为刘仲彭元喜杨柳丁一博刘胜
Owner NAT UNIV OF DEFENSE TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap