2560x1080 monitors - How much slower than 1920x1080?

[NeilF]

Out of interest, how much performance overhead does that extra screen width add?

OK it's 33% bigger, but surely that doesn't equal a 33% frame rate drop? Is it typically 10%, 20% or indeed is it about 30%?

If someone out there flicks a game running at say 60-70fps at 2560x1080 to instead be 1920x1080, what happens to your FPS?


I gauging whether to get one or not on performance hit vs visual improvement over 16:9.

 

[wunkley]

You've already done the maths. The hit is pretty much as you've said at 30 odd percent if the game is graphically taxing enough for resolution to make a difference.

 

[NeilF]

You've already done the maths. The hit is pretty much as you've said at 30 odd percent

Is this from knowledge or guesswork?

Because having looked I'm not so sure... Look at this for example - http://www.bit-tech.net/hardware/2012/03/05/amd-radeon-hd-7870-2gb-review/7

1920x1080 = 81fps

2560x1600 = 59fps

100% bigger screen, 25% drop in performance!


Or with this one - http://www.bit-tech.net/hardware/2012/03/05/amd-radeon-hd-7870-2gb-review/5

100% bigger screen, a 40% drop.


This might suggest moving from 1920x1080 to 2560x1080 might only incur a 5-15% FPS hit!?


Got a BenQ IPS 1920x1200 monitor on order which I could run at 1920x1080. However it's not been dispatched yet so thinking about cancelling and getting a Dell U2913WM (2560x1080) instead

 

[wunkley]

Is this from knowledge or guesswork?

Because having looked I'm not so sure... Look at this for example - http://www.bit-tech.net/hardware/2012/03/05/amd-radeon-hd-7870-2gb-review/7

1920x1080 = 81fps

2560x1600 = 59fps

100% bigger screen, 25% drop in performance!


Or with this one - http://www.bit-tech.net/hardware/2012/03/05/amd-radeon-hd-7870-2gb-review/5

100% bigger screen, a 40% drop.



This might suggest moving from 1920x1080 to 2560x1080 might only incur a 5-15% FPS hit!?

Bit of both, mostly guessing

I went from a 1920 x 1200 screen to the dell 2560 x 1080 and there is an obvious impact due to more pixels having to be pushed around, even running gtx 670 in SLI. But the specific numbers depend on the game really...

 

[NeilF]

Bit of both, mostly guessing

I went from a 1920 x 1200 screen to the dell 2560 x 1080 and there is an obvious impact due to more pixels having to be pushed around, even running gtx 670 in SLI. But the specific numbers depend on the game really...

What's your impression of gaming on the 1920x1200 (even in 1920x1080) vs 2560x1080?


Can you do me a favour? Run a game (or two) in 2560x1080 and then in 1920x1080 and see what the rough speed difference is?

From the above figures it's would suggest about 10% only!? Which is nothing really and I'd happily live with. If it's in fact 50% or something, then I'll give it a miss

 

[wunkley]

What's your impression of gaming on the 1920x1200 (even in 1920x1080) vs 2560x1080?


Can you do me a favour? Run a game (or two) in 2560x1080 and then in 1920x1080 and see what the rough speed difference is?

From the above figures it's would suggest about 10% only!? Which is nothing really and I'd happily live with. If it's in fact 50% or something, then I'll give it a miss

Gaming on the new 2560 x 1080 screen is great and is definitely better than 1920 x 1200 due to the super wide aspect field of view being more immersive. I never ever gamed at 1920 x 1080 on my 1200 screen as it's non native res. I never run a screen at non native res as it looks softer, unless you force it to be one to one but then you get borders

Like others say, I do miss the vertical res for desktop and working though. The new soon to be released super wides at 1440 vertical res have now caught my eye, but they're essentially 3.5k res and will likely require a GPU upgrade, which I don't want to do yet.

I'll try run the tombraider benchmark or something for you later at each res if I can remember!

 

[NeilF]

I'll try run the tombraider benchmark or something for you later at each res if I can remember!

Thanks! Would be most appreciated!


And you don't find the response times slow/laggy on it (it does rate quite low on this. eg: > 20ms), and no blur etc?

 

[wunkley]

Thanks! Would be most appreciated!


And you don't find the response times slow/laggy on it (it does rate quite low on this. eg: > 20ms), and no blur etc?

Ok bud here you go, a quick test. Tombraider benchmark. max settings, vsync off. This is on my rig below, cpu at 4.6 and 2 overclocked GTX 670s in SLI.

2560 x 1080

min 54
max 100
average 76.6

1920 x 1080

min 70
max 118
average 96.3

Only one game, but by my bad maths its pushing towards a 30% difference, especially for minimum rates (which is most important) which is about what you thought initially.

I find the response times absolutely fine on the dell 29, but then again I don't play twitch games particularly, and I notice no motion blur. See this site here for some good info, graphs and figures on lag and response times. Lag is down to 17ms in game mode according to them: http://www.tftcentral.co.uk/reviews/dell_u2913wm.htm

Anyway, Image quality and colour rendition is far more important to me which is why I don't buy TN panels.

From reviews I read of it, it's one of the better IPS panels out there for gaming at the mo.

 

[jpod]

Hi Neil & co interesting thread, if the refresh rate on the screens is 60 to 70 hz (per second) then then the figures above are limited to that factor, my understanding is faster refresh rates 120+ give gamers an advantage.

I game at 1080p 60hz, I am interested in a faster higher res, but can't figure out which!

 

[smogsy]

i would say having all the resolutions before
depending on high ultra/extreme etc on extreme these would be realistic.
1920x1080
2560x1080
2560x1440
2560x1600

2560x1080p -15-35%
2560x1440 - 35-45%
2560x1600 - 45-60%+

 

[james.miller]

expect a ~30% hit. That way anything less is a bonus and anything more will have to be adjusted for.

In reality, you cant work it out from the numbers alone, really. it's a 33% increase in pixels per frame but every graphics card will handle that differently, and that's without considering things like AA. the hit could be next to nothing if the card isnt being stressed or it could halve the framerate or worse if, say, you run out of video memory. Not very likely but possible.

 

[NeilF]

Ok bud here you go, a quick test. Tombraider benchmark. max settings, vsync off. This is on my rig below, cpu at 4.6 and 2 overclocked GTX 670s in SLI.

2560 x 1080

min 54
max 100
average 76.6

1920 x 1080

min 70
max 118
average 96.3

Only one game, but by my bad maths its pushing towards a 30% difference, especially for minimum rates (which is most important) which is about what you thought initially.

I find the response times absolutely fine on the dell 29, but then again I don't play twitch games particularly, and I notice no motion blur. See this site here for some good info, graphs and figures on lag and response times. Lag is down to 17ms in game mode according to them: http://www.tftcentral.co.uk/reviews/dell_u2913wm.htm

Anyway, Image quality and colour rendition is far more important to me which is why I don't buy TN panels.

From reviews I read of it, it's one of the better IPS panels out there for gaming at the mo.

Many thanks for that!

So if we stick with screen size increase is about the same as FPS slow down...

Well, I'm moving up from a 1680x1050 monitor to a 1920x1200, which I'll game at 1920x1080 with, which is about 17% bigger, so I'll lose say 10-15% frames.

If I was to go 2560x1080, and we stick with the same approach, that could mean, over 50%!!!

I'll stick with my BenQ IPS 1920x1200 monitor then!

Again, many thanks for the info!

 

[wunkley]

Many thanks for that!

So if we stick with screen size increase is about the same as FPS slow down...

Well, I'm moving up from a 1680x1050 monitor to a 1920x1200, which I'll game at 1920x1080 with, which is about 17% bigger, so I'll lose say 10-15% frames.

If I was to go 2560x1080, and we stick with the same approach, that could mean, over 50%!!!

I'll stick with my BenQ IPS 1920x1200 monitor then!

Again, many thanks for the info!

No worries bud, glad to help. Looking at your sig, if you are indeed using an ATI 4870 (which is pretty old and underpowered these days) and have no plans to upgrade anytime soon, you would reeeeally struggle at 2560 x 1080.

 

[NeilF]

No worries bud, glad to help. Looking at your sig, if you are indeed using an ATI 4870 (which is pretty old and underpowered these days) and have no plans to upgrade anytime soon, you would reeeeally struggle at 2560 x 1080.

I'm on a 7870xt now! The machine is 5yrs old, so I finally upgrade the GFX card.

I'll probably update my CPU etc later this year, but only if the machine starts to struggle. At the moment it's still playing everything smoothly enough for me

 

[murah]

Why don't you test it out yourself using down-sampling? I only have a 1360x768 screen at the moment but I run all my games @ 1080p things genuinely look better this way, need less AA but I have SLI now full settings ahead

 

[NeilF]

Why don't you test it out yourself using down-sampling? I only have a 1360x768 screen at the moment but I run all my games @ 1080p things genuinely look better this way, need less AA but I have SLI now full settings ahead

wunkley has actually tested 2560x1080 directly against 1920x1080. His results seem to suggest the screen is over 30% bigger and results in a slowdown in frames per second very close to that. eg: 20-30%...

If I happen to be running a game and it's struggling at around 60fps, I've rather keep it at 60fps than at around 40fps simply to have more peripheral vision


That said I could always run the Dell at lower than 2560, even down to 1920x1080 if needs be, BUT:-
1) It's £150 more than the BenQ BL2411PT.
2) Its response rate is 2-3 worse than the BenQ I believe, which concerned me slightly.

 

[AOC UK]

The question is....the monitor Wunkley used as his 1920x1080....LED or IPS?

As all the 2560x1080 29" ultra wide screens are IPS

 

[NeilF]

The question is....the monitor Wunkley used as his 1920x1080....LED or IPS?

As all the 2560x1080 29" ultra wide screens are IPS

Huh?

He was using the same Dell 2560x1080 one. He was simply running his tests in 2560x1080 or in 1920x1080 to see what the performance hit was... And that hit was a good 20-30% loss in frame rate for the additional peripheral view.

 

[skuko]

The question is....the monitor Wunkley used as his 1920x1080....LED or IPS?

As all the 2560x1080 29" ultra wide screens are IPS

no offense, but LED or IPS? why are you comparing apples to oranges?

LED is a backlighting technology while IPS is a panel type, am i right?

LED and CCFL - backlight
IPS, VA, PVA, TN, etc. - panel types

one would assume you know this if you're a vendor rep. lolz.

 

[AOC UK]

Huh?

He was using the same Dell 2560x1080 one. He was simply running his tests in 2560x1080 or in 1920x1080 to see what the performance hit was... And that hit was a good 20-30% loss in frame rate for the additional peripheral view.

ok, will read properly next time