Research On The Performance Of Graphic Accelerated Card Based On The Model Of Linear Regression

With the fast development of electronic technology , many kinds of graphic processor units have appeared . Most of them come from two famous companies : Nvidia Co.ltd and Ati Co.ltd , their products are used in broad field with advanced technology . Using the products are convenient in a lot of applications and fields , but it is hard to judge the performance of them .The main body of graphic accelerated card is graphic processor unit . The performance of graphic accelerated card is mainly judged by the frequency of graphic processor unit , the frequency of graphic memory , the number of pixel shader and the bit width of graphic memory . The performance of graphic accelerated card has more important influence to the personal computer graphic system than the central processor unit' . The initial ideas to judge the the performance of graphic accelerated card are the two followings : the first depends on the parameters of graphic accelerated card and the second depends on the testing of graphic accelerated card . Nether of Them is perfect to use .The most important part of this paper is using the model of linear regression to solve the problem initially . By changing the different parameters of graphic accelerated card , the result and data can be got . After that , using the model of linear regression to process and get the regression equation with the performance and the parameters . This method is convenient to choose or study the graphic accelerated cards . It is a useful supplementation and a relatively good method to the practical problem .At last , a software for analyzing the performance of graphic accelerated card has been designed . The software is based on the regression equations . The interface and operation of the software is very friendly and clearly . By inputting some setups and parameters , the software will get the effective result automatically .