Jump to content

Eyeinifinity/surround And You


30 replies to this topic

#1 awdwikisi

    Member

  • PipPip
  • The Named
  • The Named
  • 36 posts

Posted 02 December 2012 - 05:58 PM

hey there guys

for anyone running eyeinfinity or nvidia surround, im wondering what your "specs"
and what type of settings/tweaks you have to better the experience

this is not an epeen thread, nor a "wahhh im not getting enough fps fix the game" thread

as an eyeinfinity user, i already understand that im not going to max the game out,
i am trying to find other users experience and possible insight into stablizing frame rates and
or possibly help others via this thread who have similar set ups

specs
8120 @ 4ghz and 1045thuban @3.2ghz
16gigs of 1600mhz ram
1x7950 and 1x6970
3x2048x1152 monitors in portrait(3456x2048)

so far this is what ive tested and experienced on windows8 and windows7 64bit versions only

8120+6970=20-30 frames on medium
8120+7950=15-25 "
1045T+6970=20-25 "
1045T=7950=15-35 "

have at it

#2 Sen

    Member

  • PipPipPipPipPipPipPip
  • 757 posts
  • LocationTexas

Posted 02 December 2012 - 07:03 PM

A single 7950 should run the game at 30+ FPS. I say SHOULD because I took mine down several months ago, and they've made a lot of performance degrading "enhancements" to the game.

TBH, trade the 7950 for another 6970 [or vice versa], Xfire, and you shouldn't have any issues... from where I'm sittin' that's your only issue atm, as you can't Xfire those things together. twin cards = no issues.

From that angle, I offer the following:

the 6970 is technically the stronger card, but with the latest enhancements to the 7k series introduced by catalyst 12.11 beta 8 + how well the 7950s tend to overclock, it's a rough call to recommend one card over the other.

Benchmark results, round 1:

Single 7950 @ 920/1300

Frozen City

average FPS @ 6058 x 1080 = 30, dipped as low as 24.7 FPS Ultra high settings.

Run 2:


Dual 7950 Xfire @ 920/1300

Caustic Valley

Average FPS 27, dipped as low as 17-19. Some chop. Ultra high settings

Run 3, frozen lake, single 7950 ultra high settings

24-30 average @ 6058 x 1080

IN all honesty though, left and right screen are all internal cockpit. All three runs were done in a CENT 9-AL , and while that one had a little extra view left and right, it was mostly all cockpit. IN spectator mode, atlas, Jenner etc FoV was no different than single monitor.

Post mortem:

I left the first part of my post stand, as I'm hopeful eventually Xfire WILL make a difference at those resolutions, but I was kinda shocked by how Xfire does NOTHING. Even though the game may not be optimized, you usually see SOME KIND of performance increase. In my case, it bombed.

Edited by Sen, 02 December 2012 - 08:08 PM.


#3 Vulpesveritas

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,003 posts
  • LocationWinsconsin, USA

Posted 02 December 2012 - 10:17 PM

View PostSen, on 02 December 2012 - 07:03 PM, said:

A single 7950 should run the game at 30+ FPS. I say SHOULD because I took mine down several months ago, and they've made a lot of performance degrading "enhancements" to the game.

TBH, trade the 7950 for another 6970 [or vice versa], Xfire, and you shouldn't have any issues... from where I'm sittin' that's your only issue atm, as you can't Xfire those things together. twin cards = no issues.

From that angle, I offer the following:

the 6970 is technically the stronger card, but with the latest enhancements to the 7k series introduced by catalyst 12.11 beta 8 + how well the 7950s tend to overclock, it's a rough call to recommend one card over the other.

Benchmark results, round 1:

Single 7950 @ 920/1300

Frozen City

average FPS @ 6058 x 1080 = 30, dipped as low as 24.7 FPS Ultra high settings.

Run 2:


Dual 7950 Xfire @ 920/1300

Caustic Valley

Average FPS 27, dipped as low as 17-19. Some chop. Ultra high settings

Run 3, frozen lake, single 7950 ultra high settings

24-30 average @ 6058 x 1080

IN all honesty though, left and right screen are all internal cockpit. All three runs were done in a CENT 9-AL , and while that one had a little extra view left and right, it was mostly all cockpit. IN spectator mode, atlas, Jenner etc FoV was no different than single monitor.

Post mortem:

I left the first part of my post stand, as I'm hopeful eventually Xfire WILL make a difference at those resolutions, but I was kinda shocked by how Xfire does NOTHING. Even though the game may not be optimized, you usually see SOME KIND of performance increase. In my case, it bombed.

Umm... how is it that a 6970 = a 7950 when the 7950 was faster than a Geforce GTX 580 when it first came out, before it got all the catalyst boosts, and a 580 is faster than a 6970. Please explain.

Now if you say 2x 6970s in Xfire > single 7950 I would agree. However that's a different story.

#4 Hakkukakt

    Member

  • PipPipPipPipPipPip
  • 264 posts
  • LocationSuisse / Lausanne

Posted 03 December 2012 - 08:25 AM

hum for me

i5 3750K @ 4Ghz
16Go Ram
SLI GTX680 4Go
3x 27" 1920x1080

5880x1080 ultra-hight detail (motion blur off, AA off) => 25-35 fps in game

the fps is normal because sli is not working for this moment...

Edited by Hakkukakt, 03 December 2012 - 08:28 AM.


#5 Sen

    Member

  • PipPipPipPipPipPipPip
  • 757 posts
  • LocationTexas

Posted 03 December 2012 - 09:54 AM

Quote

[color=#959595]Umm... how is it that a 6970 = a 7950[/color]


clocks on the Saphpire 6970 are 900/1300. on the xfx 7950 they're 800/1250. you're talking 5-15 FPS improvement stock for stock in benchmarks, but at 1920 x 1080 both cards run the majority of gaming benchmark titles at well over 30-60 FPS, making the additional FPS counts benchmarking E-peen numbers.

Yes, the 7950 is the superior card, but there's that whole price component [I've seen some cheap 6970s on ebay] Price/performance the 6970 is the better buy

Hey, that sounds familiar, I wonder where I've read stuff like that before? :)

The big argument for the 7950 is that it has more GDDR5, which in most cases isn't an issue but becomes important in Eyefinity-land.

Edited by Sen, 03 December 2012 - 09:55 AM.


#6 Vulpesveritas

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,003 posts
  • LocationWinsconsin, USA

Posted 03 December 2012 - 10:35 AM

View PostSen, on 03 December 2012 - 09:54 AM, said:


clocks on the Saphpire 6970 are 900/1300. on the xfx 7950 they're 800/1250. you're talking 5-15 FPS improvement stock for stock in benchmarks, but at 1920 x 1080 both cards run the majority of gaming benchmark titles at well over 30-60 FPS, making the additional FPS counts benchmarking E-peen numbers.

Yes, the 7950 is the superior card, but there's that whole price component [I've seen some cheap 6970s on ebay] Price/performance the 6970 is the better buy

Hey, that sounds familiar, I wonder where I've read stuff like that before? :)

The big argument for the 7950 is that it has more GDDR5, which in most cases isn't an issue but becomes important in Eyefinity-land.

Where can you find a 6970 cheaper than a 7950 new? That's one of my points. You can get a 7950 new at $300 with boost clocks. 6970s go for $300+ new. (I've been seeing them at $400+) Used they're $230+ on Amazon, but for that you may as well get a Radeon HD 7870 for about the same cost (Slightly more expensive before MIR, cheaper after MIR) for slightly higher performance and lower power usage.
http://www.amazon.co...rnid=2057431011

#7 awdwikisi

    Member

  • PipPip
  • The Named
  • The Named
  • 36 posts

Posted 03 December 2012 - 02:18 PM

reason for running eyeinfinity is 2 fold, ive been running multiple monitors for a long while via various methods and am used to it,
and to take most if not all load of the cpu, which i feel is VERY underperforming for its clock
MAYBE at some point theyll code for bulldozer....

@ sen
here is where i run into an issue, when i swap to a single 1920x1080 monitor and it runs similar to your first post
i get nowhere near those frame rates with a 7950, it struggles to get 10-15, ive scratched my head on that with no solid lead or conclusion
though i have not dl'd and installed the newest catalyst yet..i will update in a day or so
now with a 6970@1920x1080 i get 30+ no problem
i will have a second 7950 whenever msi decides to finish the rma proccess...and ill have another round of tests and update this thread

@hakkukakt
thank you, ive been looking for nvidia surround results
i tried nvidia when they first introduced surround, reason i have ati cards
though it peaks my intrest that they run so smoothly, i would like an update when they get sli working

@vulpesveritas
power savings is not something i care about at all, not that it doesnt have its place
imo 6970/7870 for any resolution up to or around 2560x1440, and that is pushing it based on experience from a 6970
now ive never used a 7870,
but having used 3x5870's 6ed(cards made for 6 monitors)
i can safely say that for those 2gig cards it is nearly impossiple to run 6 monitors, and with 3 monitors its 10-15 in mwo at the lowest setting via ingame tab

#8 Sen

    Member

  • PipPipPipPipPipPipPip
  • 757 posts
  • LocationTexas

Posted 03 December 2012 - 03:06 PM

Vulp:

http://www.ebay.com/...=item2ec3eeb183

http://www.ebay.com/...=item3f1f93d273

http://www.ebay.com/...=item416ed80a14

Never said you could get them NEW, said you could GET THEM ^^


I only know this because I was going to try to unload my 2 6970s and leave my wife with a single 6870. . but I couldn't justify the sale after what I Paid for them ;)
(The fact that she would've disemboweled me could have something to do with it too >.> )

@3rd:

The ONLY thing I can think of is that something in the 6 series drivers is lingering causing the 7k series to not perform properly. It's a guess, but it's all I've got. single 7950 gets 45-60 FPS [v sync on] in CCC everything is set to "Application controlled". In my wife's rig, single 6970 performs the same way. Try blowing all the drivers out with driver sweeper and installing 12.11 beta 8 with the 7950 and see what happens. If you still have performance issues. . with JUST that card. . it kinda starts to point at the card as the problem. REALLY hope that's not the case, esp. if you already had to RMA one card. Been there, not fun.

Edited by Sen, 03 December 2012 - 03:09 PM.


#9 awdwikisi

    Member

  • PipPip
  • The Named
  • The Named
  • 36 posts

Posted 04 December 2012 - 03:02 PM

i will just reformate, as that is the best option when going that direction, fri/sat ill update the thread with findings

"off topic" my crosshair5 went up in smoke with current 7950 in it, ALL pci/pciE/lan/usb died on that board except for the slot the card was in, i wouldnt be suprised if it did do something.
the 7950 runs 880/1250, havent oc'd either of them ever, maybe ill try that

currently i have 2 donated 990fxud3's with my bulldozer and a donated thuban, hence my first curiosity into this, and research into different systems and the performance in eyeinfinity or surround.

idk ive never had a problem with rma'ing anything, which seems like a miracle.

#10 Sen

    Member

  • PipPipPipPipPipPipPip
  • 757 posts
  • LocationTexas

Posted 04 December 2012 - 03:26 PM

What did you do, fill the water loop with battery acid by mistake???????

If that card was in the board when it went, I'd RMA it IMMEDIATELY if possible. There's no telling what kind of damage was done internally. For the price of the card, it's not worth the risk.

Edited by Sen, 04 December 2012 - 03:26 PM.


#11 awdwikisi

    Member

  • PipPip
  • The Named
  • The Named
  • 36 posts

Posted 08 December 2012 - 09:32 PM

updating with further/new information
received my second 7950 back from msi
reformated in windows7 and 8 for each test ensuring clean os's at the start of the testing cycle PER card WITH beta cat 12.11
- denotes through
7950 runs stock @ 880/1500
6970 runs stock @ 880/1500
8120 runs @ 4.0ghz and 4.2ghz(interesting to note is that the difference in oc did not give any boost in frames)
win8 (vsync on or off does not change a thing)
8120+7950= 20-35 med-high settings @ 1080p, 15-30 med @ 3456x2048
8120+6970= 25-40 med-high settings @ 1080p, 20-30 med @ 3456x2048
8120+2x7950= 45 high settings @ 1080p, 20-40 med @ 3456x2048
win7 (vsync on or off does not change a thing)
nothing really differed from above except frame rates "seemed" to fluctuate far less.

did not test the 1045thuban, will be sunday's project

OFF TOPIC
the "new" 7950 performs exactly the same as the one caught in the borked ch5 board
i might oc the cards and see how much a difference that makes, as your post indicated 920/1300 that might make the difference
could be my goofy rez monitors...but they are the same aspect ratio as 1920x1080
im unsure what has cause the ch5 to poof, it just decided it was working no more.
i had the 8120 up to 4.7ghz 3 hours of prime95 stable and then had been running it at 4.2ghz for the last 4 months.
the first ch5 board i purchased the ram slots bugged out in 1 week, and i had done nothing but put in the 8120
then almost a week ago the replacement board smokes and i lose the usb/pci/pcie slots...i just have bad luck with asus it seems
OR the 8120 is borked in some way and this could be causing the underwhelming performance..
(during prime95@4.7ghz/1.4volts, the cpu never went above 45c, +1 for a single delta fan and tuniqtower)

@sen
might i inquire as to what cpu you have?

#12 Greyson Carlyle

    Member

  • PipPip
  • Elite Founder
  • Elite Founder
  • 25 posts

Posted 09 December 2012 - 12:00 PM

does xfire even work in the game yet?? I have 2 6870's and havnt noticed a diff from 1 card to 2...

#13 awdwikisi

    Member

  • PipPip
  • The Named
  • The Named
  • 36 posts

Posted 09 December 2012 - 01:24 PM

well, it doesn't seem like its fully implemented, considering I have received a slightly higher framerate and in gpuz+vom the second gpu is pegged my guess is that it does, but not well atm. I can stand corrected if someone knows positively, but it would seem strange to have the gpu at full load and not being used.

Edited by Fightin the 3rd, 09 December 2012 - 01:26 PM.


#14 MidNite27

    Member

  • PipPip
  • Elite Founder
  • Elite Founder
  • 25 posts

Posted 11 December 2012 - 11:05 PM

I have 2 6850's in crossfire. When i run at 5292x1050 on medium settings with AA off i get maximum 20FPS with 100%gpu usage on both cards. It is for that reason that i dont get to enjoy eyefinity. I play on single screen cuz i need the FPS ;)

#15 Zashel

    Rookie

  • 8 posts

Posted 15 December 2012 - 05:37 PM

View PostMidNite27, on 11 December 2012 - 11:05 PM, said:

I have 2 6850's in crossfire. When i run at 5292x1050 on medium settings with AA off i get maximum 20FPS with 100%gpu usage on both cards. It is for that reason that i dont get to enjoy eyefinity. I play on single screen cuz i need the FPS :)


Strange, i have 2x Diamond 6950's running 3x 1920x1200 (in portrait for 3600x1920) and i get 25-30 fps

specs are:

Intel i5 2500k
2x AMD Radeon 6950's
4x G. Skill 4gb 1600MHz Ram
Patriot Pyro 60gb SSD (one for sys and one for game)
Gigabyte z68xp-ud3p mobo
3x Dell Ultrasharp 2412's 24" Monitor

Edit: windows 7 home premium 64 bit

Edited by Zashel, 15 December 2012 - 05:37 PM.


#16 Sen

    Member

  • PipPipPipPipPipPipPip
  • 757 posts
  • LocationTexas

Posted 15 December 2012 - 07:30 PM

Quote

[color=#959595]Strange, i have 2x Diamond 6950's running 3x 1920x1200 (in portrait for 3600x1920) and i get 25-30 fps[/color]


Do me a favor, disable eyefinity, restart, load in and see if there is any change in performance with only one card.

I'm just curious :blush:

#17 awdwikisi

    Member

  • PipPip
  • The Named
  • The Named
  • 36 posts

Posted 16 December 2012 - 12:27 PM

testing my thuban @ 3.2ghz in windows 7/8 with a 6970/7950 @ 3456x2048 netted similar enough results that i wont bore people with them

my xfire result with the 7950's appears to be a fluke, i cannot reproduce it anywhere but on my bulldozer, however in the other test
computers it still pegs the watt draw and gpu usage despite not having a performance difference

@sen, thank you, i hadnt oc'd my 7950 at all, going from 880 to 900 gives about 3-5 frame increase, which is awesome!

@zashel, thank you for the information, as the closest to the rez i run, this is very interesting,
seems i might be out of the vidcards range with what i have.

to anyone with eyeinfinity/surround that have not replied we would like to hear your thoughts and specs to further the knowledge base for fellow mwo'ers interested in this subject

#18 repete

    Member

  • PipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 522 posts
  • LocationNew Zealand

Posted 17 December 2012 - 06:53 PM

Specs:

CPU: AMD Phenom II X4 of some sort (Not with me at the moment)
GPU: Single ATI Radeon 6870
Display*: 3840 x 1024
Settings: I think everything is turned up to the max
FPS (Average): 30 - Some times more, some times less.

* Middle monitor is 22" 1680 x 1050 and peripheral monitors are 19" 1280 x 1024. The middle monitor doesn't look too bad stretching 5:4 image to a 16:10 aspect ratio. I don't have three matching monitors, and if I did, I'd match the middle 22", but I'd be concerned about how well the 6870 would cope with 5040 x 1050. Future plans are either another 6870 when / if MWO supports CFX or a single 7970, and three matching 22" monitors with whichever GPU.

Edited by repete, 17 December 2012 - 06:59 PM.


#19 Zashel

    Rookie

  • 8 posts

Posted 19 December 2012 - 06:29 PM

View PostSen, on 15 December 2012 - 07:30 PM, said:


Do me a favor, disable eyefinity, restart, load in and see if there is any change in performance with only one card.

I'm just curious :wub:

i get about 66 fps with nothing going on, 44 looking at my team at the start and down to about 30 when action is going. i did update to the 12.10 drivers too. i should note i have all settings turned down on my card so i can play at quasi-smooth frames when the guns start firing

#20 krazypoloc

    Member

  • Pip
  • 18 posts

Posted 20 December 2012 - 01:13 PM

This game does not seem to be optimized very well yet. The last patch I did see a performance increase though when running surround. System is below.

3x 2560x1440 monitors (8020x1440 Bezel corrected)
I7 860@4.1Ghz (Just purchased a 3570k and will post back if my performance increases or not)
2x 670 4GB (Have another 670 on the way that I ordered today)
16GB memory

I usually hover right around 30FPS with mostly medium settings with the exception of texture detail and view distance I think are maxed. Also AA is off. Lowest framerate I have seen is 18FPS. Using Precision X to watch FPS in realtime.

From what I can tell this is utilizing both cards, but my performance tells a different story - I would think to get better than 30FPS in this game with 2 670's....maybe not though. I guess we will find out when I get my 3rd card installed.

Anybody running my rez running a different FoV or something to fix the fisheye. Its not bad but it could be a lot better - I am going to try a modified FoV when I get home in a bit.

Edited by krazypoloc, 20 December 2012 - 01:16 PM.






1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users