Benchmark Results: 3DMark Vantage
3DMark Vantage performance scales almost perfectly with the number of graphics processors, giving AMD's architecture a slight advantage when “Disable PPU” is selected. The reason we disabled PhysX for this benchmark is that it equates an added feature with added speed, and then adds this supposed performance gain to the CPU score.
GPU scores don’t change when 3DMark’s “Disable PPU” setting is chosen, and the performance scaling is nearly identical to the overall score.
With 3DMark's PPU setting disabled, the CPU scores are as consistent as we’d expect them to be.
7
Comments
Sponsored
they would have been better off overclocking the original card they have rather than downclocking the new model since this will mess with results
also it has been proven better test results can be gotten useing a pair of 280 in sli then useing a card such as a 9600 gso as a seperate card to handle physX and ECUDA
its built into the drivers and does give better performance that 3 280 in sli
its alot cheaper as well since xfx recetly started selling their own 9600 gso directly from the site for £50
which is about a £200 saveing over the 3rd 280
gaming at 2560x1600? I'll take normal 1080p res.
Exactly! Your tests mean diddly squat these days as you choose to use resolutions that nobody uses... well except for yourselves.
You have succeeded in convincing me of 2 things with your tests.
Don't bother buying crysis; it is either poorly written software or so demanding of resources it isn't worth it.
High end cards aren't worth the money as they won't do anything a midrange card can't do at the resolution I would use.
I agree with THW that any expensive graphics solution should be nenchmarked by demanding games @ high resolutions, no body spends alot of money to play @ 1280x1024!!
Personally, after I read this article I started to think about replacing my 8800GT with a single GTX 295 after 2-3 months (after the price get reduced and any issues get resolved)..I wonder if my p35 mobo (PCI-e v.1) and my E6750 @ 3.6Ghz will pose a bottlenck to its performance. Any ideas on that people? Help!
I think the observation that should have been pointed out here is that the framebuffer on the GTX 295 is limited, in comparison to the 4870x2... It's simple memory, the 4870X2 has 2 x 1GB where the GTX 295 has 896mb per card, and even with 4 cards there is no sharing of this memory.
Really, it's the only failing in the GTX 295, but how many people play at 2560x1200 anywaay?
On your specs, you state that you used Vista 32-bit, wouldn't that limit how much of your system & video ram the OS could access?
lol who spends that much on GPU to underclock it.
If you want to invest that sort of cash then its safe to assume there is a knowledge of overclocking. Shoulda cranked the cards.