from 670 sli to 290 cf @ 6160x1200 - performance gains

[digitaldurandal]

i5 2500k @ 4.0 ghz
8gb ddr3 1600 8-8-8

Two recent games tested at 6160*1200 (bezel corrected resolution)

Metro Last Light
CARD ----| AVG | MAX | MIN
670 sli |40.33 |144.09 |14.13
290 cf |59.67 |121.65|16.03
47.95% average framerate increase

Thief (used mantle for 290cf)
CARD ----| AVG | MAX | MIN
670 sli | 36.4 |52.1 |24.6
290 cf | 47.4 |54.7 |42.3
30.21% average framerate increase
(note the minimum frame increase using mantle here though, even though it is a lesser increase compared to Metro LL, it feels like a much bigger jump in playability. I had a lot of slow downs with these settings in Thief on the 670 sli.)

Metro LL Settings used
6160*1200
Quality High
SSAA Off
Texture 16x AF
Motion Blur Normal
Tesselation Normal
Adv Physx Off

Thief Settings used
6160*1200
Textures High
AF 8x
Shadows Low
S. Reflection Off
Depth of Field Normal
Parallax On
SSAA Off
FXAA Off
Tesselation Off

 

[monstercameron]

hmmm so the faster cards are...faster?

 

[SlowSpyder]

Nice, not bad increases. I like seeing benches at settings not everyone uses. I have to think at that resolution more vram will be good to have for future games. Interesting that the max FPS on the 670's is higher than with the 290's in your fist bench. Not that those numbers make a difference in any practical playability way, but somewhat unexpected none the less.

 

[digitaldurandal]

Nice, not bad increases. I like seeing benches at settings not everyone uses. I have to think at that resolution more vram will be good to have for future games. Interesting that the max FPS on the 670's is higher than with the 290's in your fist bench. Not that those numbers make a difference in any practical playability way, but somewhat unexpected none the less.

I noticed this as well. I think the max fps is taking place on a transition frame or something. Most important to me are the minimums (not the lowest ever frame but like the lowest 1 - 5% and the average.

@monstercameron

The point of the post was to help people see how much the difference is. I have a fairly average system save for the 2 card configuration and eyefinity setup. I know from experience that it is difficult to get performance comparisons at these resolutions. I expected the cards would be faster.

Also of note is the noise level of the cards is not really much more noticeable than the 670s and I must say the build quality of these cards is great. They have a nice backplate and the shroud is well made. The cards are about an inch longer than the 670s. I had to remove the easy pin readout from my motherboard because it was blocking the second card. I also had to flash my bios to get any display.

Haven't got to game on them yet, the 670s already made the room heat up so it will be interesting to see what the 290s do, at least I won't need to heat the room for winter .

 

[Silverforce11]

You have reference CF R290s?

They are definitely throttling if you leave settings on default.

 

[digitaldurandal]

You have reference CF R290s?

They are definitely throttling if you leave settings on default.

They are MSI 290 Gaming cards. Have not monitored the temperatures yet.

 

[digitaldurandal]

I was also able to get about the same FPS (very close actually almost identical) on the 290s but set to Very High Quality instead of High. Not sure if I prefer the quality or the FPS.. that is always the question.

It is just on the edge of how playable I like. I guess I will have to overclock these cards now.

 

[digitaldurandal]

You have reference CF R290s?

They are definitely throttling if you leave settings on default.

Turns out card 1 is throttling. I forgot the new motherboard I have only has 2 PCIE and they are fairly close. Am going to have to check out those fans you suggested.