header image
Home
Contact Us
xoxide
PC Hardware
Login Form
Username

Password

Remember me
Password Reminder
Home arrow intel arrow Intel's Clarkdale: CPU & iGFX Part IV
Intel's Clarkdale: CPU & iGFX Part IV Print E-mail
Written by Michael Schuette   
Jan 16, 2010 at 05:00 PM


OpenGL and More Gaming

At one time, OpenGL (Open Graphics Library ) used to be the predominant API for 2D and 3D applications. OpenGL is a standard cross platform API for computer graphics originally developed by Silicon Graphics in 1992, featuring over 250 function calls. While OpenGL has been pushed somewhat into the back by DirectX powering most games in the Microsoft Windows environment, it remains a powerful factor in CAD and scientific computing. OpenGL is currently managed by a non-profit technology consortium known as the Khronos Group. OpenGL allows each vendor to add functions through extensions of the standard that are designated by specific identifiers (for example “NV” for nVidia).

Integrated graphics, regardless of how good and powerful they are, cannot compete with standalone discrete graphics like those provided by nVidia and AMD, at least not in the mid and high-end gaming environment where game developers are competing with each other with respect to visual effects and level of detail for photorealistic displays of the game contents. In that particular environment, also frame rates count, nobody wants to play computer games at frame rates the drop below 30 fps, particularly since aside from a competitive disadvantage in first person shooter games, it is plain and simply annoying having to wait for the next screen to be rendered.

An additional issue in this regard is the demand on Anti Aliasing procedures. There is no doubt that full-screen anti aliasing improves the perceived image quality by smoothing jagged edges and taking out the annoying crawling of diagonal lines, yet, strictly speaking, it is averaging of data points that may not be required or even desired in scientific or technical applications where any shifting of a line could have significant consequences on the interpretation and validity of data. At the same time, the supersampling and multisampling techniques used by commodity graphics cards manufacturers significantly increase the demand on the graphics memory buffer size since sub-pixels have to be taken into account. In a graphics configuration relying on any united memory architecture (UMA), this adds a tremendous amount of data traffic, thereby stressing the bandwidth limitations of a system memory bus, which, in comparison to the ultra-wide, ultra-high speed on-board memory interfaces on discrete graphics cards is extremely limited, even on Neahlem systems.



Last Updated ( Mar 12, 2010 at 09:01 AM )
<Previous Article   Next Article>
NIA in action, what fun!

badge
We have 72 guests online