Quote:
Originally posted by Han
He's already explained it. An explanation probably 3 times better than one I could do. 620MHz 256bit DDR memory is basically 1240MHz 128bit DDR memory, or 620MHz 128bit DDR2 memory. 
You cant "upgrade" a Video card's memory, they come with a fixed amount. I havent seen a 256MB 9700 though.
|
Umm, not exactly. For the most part you're right but not the DDR2 part. DDR2 is just faster DDR memory. It doesn't offer anything new in terms of technology. 620 Mhz 256-bit DDR is equal to 1240 Mhz 128-bit DDR as well as 1240 Mhz 128-bit DDR2. The upcoming GeForce FX will use 1 Ghz DDR2 memory so it'll actually have less raw memory bandwidth than the Radeon 9700 Pro (GF FX = 16 GB/s, R300 = 19.8 GB/s).
I'm not entirely sure on FSAA algorithms. There seems to be a lot of conflicting information. I think supersampling is taking an image at a higher resolution and sampling it down, probably using bilinear filtering. Either that or it breaks each pixel into sub-pixels, calculates the color of each subpixel and averages it to get the final color. Multisample, I believe is rendering an image multiple times and merging them on a sub-pixel level. I'm not entirely sure about them. I'm going to have to do more research on it to get a definite answer. There also seems to be conflicting information on the FSAA method of the Radeon 8500. In its first set of reviews (at either THG or Anandtech) it's stated that it uses multisampling but in later articles it's shown as supersampling.
If anyone else knows the answer feel free to jump in.
__________________
CPU: Intel Core 2 Quad Q9450 Mobo: Intel DX48BT2 Memory: 8192 MB PC10600 DDR3 Videocard: PNY Geforce 9800 GX2 Soundcard: On-board SigmaTel High Definition Audio Hard drive: 1 TB Seagate & 1 TB Hitachi Optical drive: LG GGW-H20L (2x BD-R DL) PSU: Nexus 1000 Watt OS: Microsoft Windows 7 Ultimate (64-bit)
Last edited by Demigod; January 14th, 2003 at 20:37..