• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

why don't video cards and motherboards migrate towards shared memory system?

draggoon01

Senior member
why don't video cards and motherboards migrate towards shared memory system as done in xbox or similar. wouldn't it be more economical for everyone that way (except ram makers), video cards would be cheaper, it'd be easier to upgrade for games, and ram upgrade would benefit system overall as well as games. and by working on a faster connections between the system ram and components, it would improve overall performance.

i was wondering after reading this article

link

the prices on the GeForce4 Ti4200 and Ti4600 chips rose up $11 and $19 respectively per one chip

surprising how low the cost of actual gpu was. and prices ROSE to those numbers so normally they were selling for less. was wondering where the cost of a $200 video card went, and thought that ram was taking a big chunk.
 
Well you can get onboard graphics cards which use system RAM. But if RAM hase 2.1GB/s as PC 2100, 2.7GB/s as PC2700 etc, and graphics cards are 8.8GB/s, 10.4GB/s etc, wouldn't that mess up the data that can go through the card meaning it doesn't work as well? And don't GFX card's have different CCD to mobos, hence why they have DDR that can run at 550Mhz, compared to the 400/333/266 whatever of system RAM? Or am I a n00b who doesn't know really what I'm talking about?
 
High end main system RAM is currently approaching speeds that the graphics industry surpassed more than 3 years ago. Speeds of Graphics DDR now exceed 600 Mhz, and run on 128 or 256 bit interfaces. This means that current high end cards like the Radeon 9700 Pro provide nearly 20GB/sec of bandwith, compared to 3.2GB/sec for DDR400 or 4.2GB/sec for dual channel PC1066 RDRAM. This doesn't even match the memory speed of the now archaic Geforce1 DDR (4.8GB/sec).
 
"wouldn't it be more economical for everyone that way"

Main system performance would only benefit up to a certain point at which bandwith would saturate the cup's requirements. The main problem would be supplying enough bandwith after that to satisfy a high-end graphics solution. Considering that most people on this forum run their rigs with about 512mb of ram, the bill fo 512mb of DDR650 or so would be huge. Trying to implement a 256-bit memory interface on a motherboard would also send a lot of motherboard makers crying home to their mothers!
 
Back
Top