The Research of Acceleration Methods which Based on GPU Hardware

Jin Xu; Zheng Mao Mei; Yu Ling Zhou; Qiao Wang

doi:10.4028/www.scientific.net/AMR.760-762.1699

Paper Titles

Design of Tiny TTY for Multi-Process Micro-Kernel Embedded OS on ARM
p.1677

A Novel Construction Method of QC-LDPC Codes Based on Generalized ACE Constrained
p.1684

Design of 3D-Route Planning Based on the Improvement of Simple Genetic Algorithm
p.1690

The Study of the PCI Bus Expansion Based on PCI Bridge Chip
p.1695

The Research of Acceleration Methods which Based on GPU Hardware
p.1699

Semantic Inner Product Based Web Service Matchmaking Method
p.1703

The Research of Semantic Web Service Matching Algorithm Based on the Formal Concept Analysis
p.1708

A TIFF-Tree Based High Utility Itemset Mining Algorithm
p.1713

Constructing Secure Cloud Computing Platform for Strong Smart Grid
p.1718

HomeAdvanced Materials ResearchAdvanced Materials Research Vols. 760-762The Research of Acceleration Methods which Based...

The Research of Acceleration Methods which Based on GPU Hardware

Abstract:

In todays world, the web pages are both descriptive and graphic. The videos are more vivid, and the games more realistic, and so on, which all demand for the ability of concurrently processing greater data flow. Considering that the CPU has more control unit but less calculation unit, so the CPU can not meet multitasking and more concurrently processing greater data flow. But GPU has amount of calculation unit, and it can meet easily. In order to achieve the effect of hardware acceleration, this paper dealing with the redundancy comes from data transmission between CPU and GPU,optimized communication module, raised a VIP channel to multistage process the signals of 2D and 3D,and increased the logic of hardware and the ability of processing the driver from user layer.