Is it possible to use system RAM as "VRAM" or graphics memory?

[maguserroris]

Hello all, thank you for your time. Here's my situation:

I've been playing with Dataram/AMD Radeon RAMDisk to see how useful it is. (No consensus on that yet...) I also do a bit of distributed computing (Folding@home, etc.). One of the programs I use relies heavily on GPU and the GPU's memory. So now I'm wondering if it's possible to use system RAM to supplement graphics RAM. RAMDisk uses system RAM to create a virtual disk that operates extremely quickly. But can it append system RAM to the GPU's memory? Is there some software out there that does this? Or is there another way to increase GPU memory similarly to the way a Page File supplements system RAM and overflow?

I currently have 8GB DDR3 @ 1866 MHz, and a Radeon HD 6950 (flashed to 6970) 1GB.

My ultimate goal is to simulate having more like 3GB or 5GB of GDDR5 instead of just 1. Is there any known way to do such a thing without buying another GPU to put in CrossfireX?

Thanks again,
MagusErroris

 

[Kevin Authenthic]

I read somewhere that ram can not ever be as fast as VRAM because it has memory that was intened for that but when your vram runs out it looks for your regular ram

 

[bryonhowley]

None that I have ever heard of no. And putting another card in CrossfireX will not increase your ram one bit. Both cards would then use the lowest amount of ram the card has so if yours has only 1gig that is what you still have 1gig.

 

[maguserroris]

None that I have ever heard of no. And putting another card in CrossfireX will not increase your ram one bit. Both cards would then use the lowest amount of ram the card has so if yours has only 1gig that is what you still have 1gig.

Thanks for the quick replies. I realize they'd be using 1GB each, but with distributed computing (and to some extent gaming as well,) you would see benefits from the shared load handling which allows the cards to be using those separate 1GB modules simultaneously.

Again, I'm just trying to increase performance for applications that rely heavily on GPU and its memory.