SLI GTX 970 G1 Gaming vs GTX 980 Ti Strix OC vs GTX 1070 Dual benchmarks

[EarthDog]

Ive learned to take the blue pill...

 

[Kanan]

Ive learned to take the blue pill...

Yeah the red pill lost its fun here.

 

[Artas1984]

more opinions would be welcome.. for normal gaming how many fps is enough.. i am curious..

trog

I will gladly indulge you with this.

There are games in those you need a lot of FPS in order to deliver some specific physics engine based feats and allow the player to execute specific movement tricks. From personal experience the best examples are Quake 2 and Painkiller.

In many of Quake 2 maps there are places where you need to perform specific double jumps on almost wall like objects. These jumps are much easier to make with 90 FPS vs. 60 FPS.
In Painkiller multi player it is essential to keep 120 FPS as your minimum in order to perform crucial rocket jumps, double bunny-hops and keep air control. It gets even more exaggerated in single player; there are trick jumps, which demand 300 FPS in order to be performed. Take a look at this video Painkiller Trickjumps and look at 24:17 - to perform this type of wall climb with vertical gravity control you need no less than 200 FPS.

Sure, these FPS shooters are old games and nobody makes such sophisticated games anymore with such advanced jumping techniques to be learned. These days shooters absolutely suck balls compared to the original Quake and noone needs 100+ FPS for jumping, which was not the case in the good old days.

 

[trog100]

I will gladly indulge you with this.

There are games in those you need a lot of FPS in order to deliver some specific physics engine based feats and allow the player to execute specific movement tricks. From personal experience the best examples are Quake 2 and Painkiller.

In many of Quake 2 maps there are places where you need to perform specific double jumps on almost wall like objects. These jumps are much easier to make with 90 FPS vs. 60 FPS.
In Painkiller multi player it is essential to keep 120 FPS as your minimum in order to perform crucial rocket jumps, double bunny-hops and keep air control. It gets even more exaggerated in single player; there are trick jumps, which demand 300 FPS in order to be performed. Take a look at this video Painkiller Trickjumps and look at 24:17 - to perform this type of wall climb with vertical gravity control you need no less than 200 FPS.

Sure, these FPS shooters are old games and nobody makes such sophisticated games anymore with such advanced jumping techniques to be learned. These days shooters absolutely suck balls compared to the original Quake and noone needs 100+ FPS for jumping, which was not the case in the good old days.

in such situations.. which i would say are not "normal" we would be putting high frame rates or as high as possible frame rates as a top priority.. "pretties" or good visuals would come at the bottom of the list.. resolutions would be turned right down..

these are not the kind of situations i engage in and i doubt many do.. my "how many frames rates per second" are enough question is more aimed at more normal gaming situations where good visuals and fluid game play is at the top of the priority list..

thanks for indulging me but so far i aint seeing much of answer.. i do play what i would call fast single player shooter games.. painkiller being one of my long term favorites and see think my 75 fps frame rate cap is more than adequate.. but i no longer have the reactions of a hyper active fifteen year old so my own judgment may not be a good one..

i do think "good enough" is a good approach for every day use and i must admit to not feeling any great need to run "ultimate" settings just because i can.. i do turn things like AA down and anything that gobbles gpu power without producing (to me) much of a beneficial result off.. which i think does put me in a minority on an enthusiast forum.. he he

trog

ps.. when i went 980 TI sli i knew excessive heat generation would be a problem.. i have exactly the same problem with my current "obsession".. over powering RC cars.. excessive heat generation seems to kick in with most tech type hobbies.. in fact it seems the one single common denominator.. all this stuff gets too f-cking hot and getting too f-cking hot is where it all stops..

 

[64K]

i do think "good enough" is a good approach for every day use and i must admit to not feeling any great need to run "ultimate" settings just because i can.. i do turn things like AA down and anything that gobbles gpu power without producing (to me) much of a beneficial result off.. which i think does put me in a minority on an enthusiast forum.. he he

Your SLI 980 Ti @1440p contradicts what you are professing to care about. That's $1,350 worth of GPU in your rig when you bought them.

when i went 980 TI sli i knew excessive heat generation would be a problem.

If too much heat is an issue then sell those two 980 Ti add a couple of hundred cash too it and buy a single 1080 Ti. That's plenty of GPU for 1440p.

If you show up here with SLI 1080 Ti saying the things you are saying then I won't know what to think of you.

 

[EarthDog]

Not to mention, he turns settings down... which i would consider 'good enough' versus he has plenty of horsepower to run default ultra...inckuding AA. I can still easily see jaggies without AA at 2560x1440...

Its trog....*hands 64k the blue pill, takes mine so as not to get tortured further*

 

[Kanan]

Guys you're wasting your time, he is too special. Let him do whatever he wants, don't argue.

 

[rtwjunkie]

Yeah "normal gaming", whatever that is, you can see it in fast games such as your everyday shooter, racing games and you can see it on the desktop when you move things. You can pretty much see it everywhere, the difference is big. If you can't see it, you're blind. he he Shooter players know this for decades now, they used to play on low resolution on a CRT monitor to have high FPS and 100 Hz. Are you seriously doubting this? Go and try it again.

I can't tell the difference above 50fps. I'm not blind, either.

 

[Kanan]

I can't tell the difference above 50fps. I'm not blind, either.

Are you sure? I could demonstrate it in real life to you, too bad I live a few thousand km away.

 

[Artas1984]

I can't be sure myself in the matter of this topic. I probably don't see the visual difference between 60 and 120 Hz in games, because the monitor's image quality and mouse input lag are far more important things than monitor refresh rate for shooter games, but i sure as hell feel (not see) the difference in gameplay 60 vs 120 FPS.

 

[Kanan]

I see and feel it, even the mouse pointer is far smoother, the windows are smoother, popping up of tasks looks better. Games look world's better. How many examples do you guys need? I feel a bit of ignorance here. Maybe it's true after all and I'm simply more sensible. I have a past as a pro level player as well it would suit the profile.

 

[rtwjunkie]

No sir, we're not ignorant. I don't doubt you see a difference. A great many people, though, including me, aren't so sensitive with their vision, and don't see such minute differences. It's simple biological differences.

 

[CAPSLOCKSTUCK]

same here, i cant tell any difference between 40- 60 fps. I feel i drive better with higher fps though.

Turning the frame counter off helps a lot.

 

[64K]

I wonder if some people run FRAPS and get so preoccupied with FPS that they convince themselves that they are losing out if the FPS drops below a certain target rate.

I ran FRAPS a few times over the years and tbh I couldn't point to the sections of the game where the FPS dropped from 60 to 50 in Shooters even though FRAPS recorded it as such.

Dropping to around 30 FPS in was very, very obvious though and below 30 FPS average is unplayable to me. I don't need FRAPS to tell me that.

 

[Artas1984]

I have a second Asus GeForce GTX980 Ti Strix OC Gaming now.

While this is not about SLI GTX980 Ti performance, i am going to compare SLI GTX970 VS SLI GTX980 Ti scaling in general at 2560X1440. I already expected worse scaling on SLI GTX980 Ti than on SLI GTX970, because SLI GTX970 had to work harder, so no surprises. I deselected 3 games from the list: Alan Wake American Nightmare, Left 4 Dead 2 and Thief. First two were getting FPS wall limit, and Thief is just badly optimized for GPU testing.

So, after benchmarking the rest 17 games, i've got SLI GTX980 Ti scaling at 46 %, which is 24 % worse than on SLI GTX970.

 

[EarthDog]

4ghz broadwell holding both sets of cards back...

What would be interesting is seeing what that 5775c overclocked to over 4.5ghz does to the scaling....ill bet it chamges the story a bit.

 

[Artas1984]

4ghz broadwell holding both sets of cards back...

What would be interesting is seeing what that 5775c overclocked to over 4.5ghz does to the scaling....ill bet it chamges the story a bit.

Overclocking Core i7 5775C to 4.4 GHz shrinked the gap from 46 % to 49 %. The biggest gain was seen in Far Cry 4 - 15 % FPS improvement. The other games showed very small FPS gains. However, i can achieve even greater SLI scaling results, since i found some testing places/methods were holding the cards back in some games. In the future, when i will make exclusive SLI GTX980 Ti benchmarks, i will choose different places for some games, and include some new games.

 

[Artas1984]

Over the past two days, i have maximized my SLI GTX980 Ti performance to the top - CPU 4.3 GHz OC, newer games replacing older games, new drivers and different testing places sure made a difference overall. But the most important fact - i found out the best anti aliasing methods to improve SLI performance in an overwhelming way! For instance, by using super sample AA, i've increased my SLI scaling in Rise of Tomb Raider from 10 % to 55 % with the game still being playable with single GTX980 Ti at 2X SSAA! Project Cars SLI scaling went from 18 % with no AA to a whooping 85 % with custom DS6X AA with the game still being playable with single GTX980 Ti at DS6X AA. That's what i am talking about!

Here is a summary

Alien Isolation: 76 %
Assassins Creed Syndicate (2X MSAA + FXAA): 68 %
Battlefield 1: 52 %
Call of Duty Black Ops 3: 51 %
Crysis 3 (2X MSAA): 55 %
Deus Ex Mankind Divided (2X MSAA): 83 %
Dragon Age Inquisition (4X MSAA + PPAA High): 82 %
Dying Light: 67 %
Far Cry 4 (4X MSAA): 93 %
Mad Max Fury Road: 76 %
Mass Effect Andromeda: 56 %
Metro Last Light Redux (2X SSAA): 82 %
Middle Earth Shadow of Mordor: 76 %
Mirrors Edge Catalyst: 72 %
Prey (2017): 60 %
Project Cars (DS6X + SMAA Ultra): 84 %
Rise of Tomb Raider (2X SSAA): 56 %
Sniper Elite 3 (2.25X SSAA + AA High): 93 %
Watch Dogs 2 (2X MSAA + SMAA High): 43 %
Witcher 3 Wild Hunt: 79 %

That's 70 % scaling for SLI GTX980 Ti! I would like nothing else than to get my hands on GTX1080 Ti now to see how would it compete.

Use this same method for SLI GTX970, and i predict SLI GTX970 performance to reach 85 - 90 %, by which single GTX1080 would be defeated.