Widescreen Gaming Forum

I'm trying to get some feedback before making a purchase and hope maybe you guys can help. I'm looking to upgrade my cards in my surround setup and can't decide which is the better path: 290x or 780ti? First question: I can never ever find reviews where new video cards are tested at surround resolutions/configs. However, lately you see a lot of benchmarks for new cards on 4K displays or 3840x2160. I've been looking at 4K sli benchmark results to try and get a feel for how well the 780ti vs. 290x will perform at higher resolutions. I don't yet fully understand how a single or sli setup on a 4K monitor loads the GPU(s) vs. how the same single/sli setup might perform at the same (or higher) resolution on a surround setup? Is there something I'm missing such that I shouldn't use a 4K benchmark as a good indication of how the same card(s) would perform on a surround setup? Is there a difference?

Second, I've been limited to 2GB VRAM on my 680s and so far this really hasn't been an issue since I have to turn the in-game settings down to get playable frame rates. However, with the increased performance, I'm wondering if the 3GB vs 4GB of ram on these cards may actually come into play. Will the 3 vs. 4GB of memory on the 780ti will ever be a limiting factor?

My heartburn here is I've been a "green" fan forever & have never owned an AMD card. However, looking at the 4K performance of the 290x CF versus 780ti SLI, the "red" team is kicking butt. The performance increase in 4K benchmarks is substantial enough that I'm wondering if I shouldn't jump ship. (AnandTech Review: http://www.anandtech.com/show/7492/the-geforce-gtx-780-ti-review/6) I accept the differences in heat, power, and price.

_________________
| i7 3930K | 16GB Mushkin Redline 2133MHz | ASUS Rampage IV Formula | EVGA 2080ti SC ULTRA | EVGA Supernova 1200W PS | 2x Samsung EVO 500GB SSD RAID 0 | 2x IBM 520 SSD 240GB | Rosewill Blackhawk | 3x Acer Predator G-Sync 144MHz @ 7680x1440 |

Well to tell you the truth I find vary very few games actually take advantage of SLI or Crossfire, some games actually hinder the FPS if in SLI or Crossfire. This is one of the reasons I went from 2 5850's too 1 7970 this time around, there was just no real reason to crossfire.

But as to the OP reason, you should be able to take into account 4k somewhat. But a rule of thumb but not totally accurate is whatever the FPS is on 1 screen lest 1/2 when running 3, this can vary greatly depending on the game.

But also as to why there is not a lot of benchmarks out there for multi-monitors, is
1 - no programs really test this right, as designed for 1 screen. There is work arounds, but not totaly acurate, other then listed at bottom of post.
2 - as stated before not a lot of games handle SLI well, so for most but a rare few new cards, Nvidia did not test well as they needed 2 cards for surround, AMD only ever needed 1.

Only way to really test multi-monitors is to run a game and have everything exact same and take average FPS, that’s immensely hard to do, so numbers will never be the same, so no one has bothered. This will probably change in next few years though as multi-monitors is growing at an increasingly fast rate now.

_________________
Windows 7 x64 SP1
SABERTOOTH 990FX R1- AMD FX-8350 4.5 - Kingston Hyper X 16gb DDR3 1600
ATI R290 Eyefinity

While you have a higher pixel count on3820x2160 you do have a higher FoV on 5760x1080 resulting in more objects being rendered. it's totally game engine dependable which will result in more FPS. But in the End they're kinda similar.

Here are some Benchmarks in Both resolutions, sadly no CF/SLI
http://www.bit-tech.net/hardware/graphi ... b-review/5

The 2GB VRAM are not an issue now, they might not be an issue tomorrow, even in 3 Years you might still run 8xAA in a few games, maybe you have to reduce to 2xAA in other games. Nobody knows.

The difference between 290X and 780Ti is not that great. If you've always been a huge fan of Nvidia, stick to nVidia, except you feel the need to change!

_________________
We gonna send it to outa space!

On the topic of 4k I know the original set had to be two dvi connectors is this the same if using display port? The 780ti the Evga version has one DP two DVI and one hdmi. If I "was" going to get a 4k monitor at this moment I would have to use the two dvi is that correct or can you use one DP for 4k?

Sent from my SM-N9005 using Tapatalk

_________________
i7 3770K (Ivy bridge) | Corsair Dominator 32GB | Asus P8Z77-V LE Plus Intel Z77 | Corsair H100i | 2 x Gigabyte ATi Radeon HD 7950 in crossfire | NXZT Phantom case | OCZ 1000W Powersupply | Samsung 840 PRO 256GB | 2x OCZ agility 3 60GB SSDs | 2x 1TB HDDs |

I dont have surround setup, but you need to remember that 3840x2160 is 4 x 1920x1080. So it's not even close to "triple-wide" but more like "quadruple"

You can always simulate the behavior of your system at 3840x2160 by creating custom resolution in nvidia drivers.
Run the fps benchmark in fraps and see for yourself how big the drop of the framerate is.

EDIT:
I created custom resolutions in Windows XP.
I did a very quick test in Burnout Paradise. It natively supports 5760x1080 and 3840x2160.
Mind you i have a weak PC: C2Q q9300 @ 3.0gHz, 4GB DDR2 RAM, GF GT640 2GB.
SSAO - OFF, AA - off. The rest was maxed.

4k indeed lowers the framerate.
when the car was standing still in a very quite place, the framerate at 5760x1080 was 49-50, at 3840x2160 it was 43-44
In action i was getting:
at 5760x1080 Min - 30, Max - 58, Avg - 43.903
at 3840x2160 Min - 22, Max - 49, Avg - 36.237
another quick test:
at 5760x1080 Min - 20, Max - 53, Avg - 43.352
at 3840x2160 Min - 16, Max - 48, Avg - 37.435

You will be better off with your own setup in more modern games. Burnout Paradise PC ver. is from 2009.
Here is the DR, if you're interested http://www.wsgf.org/dr/burnout-paradise