ATi Radeon Graphics - Hilbert Hagedoorn @ 18:53

Hey Folks! While reading the Ati PR I wanted to give you a little insight on what this baby can do.

Frankly I'm not too surprised to see that ATi will release a product as they do now, rumors had been floating the web for weeks now. ATi is hitting back really hard with this new Videocard (ATi Radeon 256).

In short I'll adres the big advantages of this new videocard. The most important technology that ATi pursued was speed vs. quality the Radeon 256 is based up-on the Charisma™ engine, a engine that indeed is a GPU with Hardware T&L, advanced vertex skinning (smooth joining of objects) just like the GeForce, and 3D key framing (fluid animation through 3D morphing).

Secondly, speed. this card is going to blazingly fast and competitive with the GeForce 2 ! In theory (we have to see some benchmarks first of course) this card will push 1.6 GigaTexels/sec just like the GeForce 2. You can compare this with a 'plain' GeForce which pushes 0.48 Gigatexels/sec.

Third, Pixel Tapestry Architecture™. ATI's RADEON GRAPHICS will utilize the most advanced and flexible texturing system available, it will render 3D surfaces in very high detail through option like Dot bump mapping, Environment Mapped Bump Mapping (EMBM) !!, Texture Transformations,


Dual-Paraboloid Environment Mapping


Cubic Environment Mapping


Shadow Mapping


Now there slogan: Ati, the new face of graphics After seeing the specs I must admit it's possible ! The Radeon 256 is going to be a serious competitor towards the geForce 256 and most certainly towards the Voodoo 5 5500 videocards since it will be faster and since I truly believe that EMBM and all other cinematic effect on this card is way more important than FSAA or Motion Blur.

Well, that was my little scoop !

 

Let the games begin!! - Kyle Morrison @ 15:29

This will officially be the craziest sping graphics card showdown EVER! The big news tonight is the unveiling of ATI's Radeon chip. After just a quick look at the details I have to say that I am very impressed with the features of the Radeon. ATI is pushing huge a huge texel rate (1.6 Giga-Texels) and an advanced geometry engine.

Charisma Engine: ATI's Geometry Engine
Charisma combines the world's fastest geometry engine along with revolutionary new techniques for bringing 3D characters to life. The geometry engine supports full transformation, clipping, and lighting (T&L) at 30 million per second processing capability, resulting in a 10x improvement in 3D modeling detail relative to current CPU based geometry. Charisma supports cutting-edge character animation features, such as advanced vertex skinning and key frame interpolation. These effects are required in bringing game characters to life with realistic behaviors, fluid movements, and believable facial expressions including lip-synching.

Pixel Tapestry: ATI's Rendering Engine
Pixel Tapestry combines the world's fastest rendering engine along with an unprecedented new set of 3D special effects. It is the first graphics chip to break through the Gigatexel barrier with an awesome 1.5 Gigatexel per second rendering engine. The new chip's ability to process up to 3-textures concurrently, means unsurpassed performance in accelerating the coming wave of multi-textured game content. Because of its awesome rendering capability, RADEON 256 enables full performance in 32-bit color, which finally makes 16-bit color obsolete.

Those are the major points that ATI is pushing with the Radeon, all I have to say is SEND ME ONE! For me this has taken the steam out of 3dfx's Voodoo 5. I'll stop there and save all judgment until I have seen all the production boards in action. Till then read the rest of ATI's press release here: Radeon Press Release

DOH!! - Nicholas Vasquez @ 15:01

In my hurry to get the scoop on the new ATI chip, I forgot to include a link to the article. Please forgive me. Please follow this link to skip past the flash intro.

Love your RADEON - Nicholas Vasquez @ 14:31

Hah, just out from ATI is the new RADEON chip, Rage 6, and charisma engine, type quick , gotta scoop Hilbert. Copy, paste lovin':

"Three Filtered Textures Per Pixel - At Full Speed
Current generation graphics processors have just one or two texture units per rendering pipeline. They are capable of applying up to two textures to a pixel in a single clock cycle, but with one texture unit per pipeline this requires using two rendering pipelines to generate a single pixel. Thus, the fill rate of these products is effectively cut in half when multitexturing with two textures per pixel. In order to add a third texture to a pixel, these chips must use a second clock cycle, which cuts the fill rate in half again. Thus, it takes most current generation processors four times as long to apply three textures to a pixel than it does to apply a single texture. Even those that have two texture units per pipeline take twice as long to apply three textures per pixel. Thus, the number of rendering passes required to output a single pixel is becoming increasingly critical to providing acceptable graphics performance.


One of the most talked about factors for determining graphics performance today is "peak fill rate". This refers to the maximum number of pixels a graphics product can render to the screen per second. However, peak fill rate is calculated with the assumption that each pixel has only a single, unfiltered texture applied. In reality, 3D games today make heavy use of filtering and multitexturing, and future games will use these techniques even more extensively. The result is that graphics architectures that do not fully take multitexturing requirements into account will suffer rapidly decreasing performance as more textures are applied to each pixel. As the following chart illustrates, including more texture units per rendering pipeline allows a graphics chip to maintain high performance even with large numbers of textures per pixel."

Maybe this one will get up there with the big boys.

home

Copyright 1999 - All rights reserved Hilbert Hagedoorn

 

 

trans468x15.gif (182 bytes) right.gif (163 bytes)