I am running the Asus 4870x2 on Vista 64x , 8gb memory , Asus P5Q-E Motherboard, Catalyts 8.9.
If you look at my Hardware information on Catalyst, it shows I have 767mb HYPERmemory, and not 2gb?
I then sent the card back to the supplier, he tested it on another machine and it showed just under 2gb hypermemory, and then tested another new card, and got the same result. He then 3d marked, all tested 100%
Why am I seeing 767mb hypermemory? When I owned the saphire , it showed 1024mb memory?
It shows 767mb across both cards.
I really have no idea if this is an issue or a normal? Any experts brave enough to take this on?
What does GPU-Z say? How do you like your card and what kind of performance are you getting in general usage and games etc.? Also in the bios or during the boot what does the system report for GPU mem?
Its advanced memory management technology makes more efficient use of all the available memory in computers with PCI Express® technology. Graphics applications can share system memory and use fast on-board memory for real-time processing between the graphics card and the computer’s memory. This ultimately increases the affordability of ATI’s Radeon™ graphics technology.
--- Therefore how do I know If I am sharing system memory or not, my memory reading is too low. Can this be disabled?
You have 1 gig of VRAM, no need to worry. Youll never use the hyper memory, that would mean youd have to use your vram first. Even if you somehow managed to use more than whats on your card, the hypermemory would only allow your system to run faster. Dont worry
Even if my card runs out of memory, there is NO way that using the system's RAM would be *near* the same as using the VRAM. The GPU would waste more time waiting for the CPU to allocate his memory needs than waiting on his own address space.
It's a neat tech for slow video cards, but hi end can't even THINK of using that. Maybe, and only maybe, on "2D over 3D" apps (video and non-real-time rendering).
The other way around sounds like a smarter move though, but it would still have performance penalties of memory intensive apps cause of the paths it has to take. The CPU would have to wait even more cycles to get that memory in practical terms.
Firstly, I wasnt claiming itd run faster than it does with vram, let me clarify that. What I was saying is that itll run faster thru the pci using system ram than going all thwe way around and still using system ram. This usually shows up in 2 areas, low end low ram cards or mobiles. Not sure why hes seeing this