Mirror's Edge Catalyst 'Hyper' Settings Made Possible by 8GB VRAM

[Bacon1]

Yes, and I also bothered to check drivers first:





Not only this but many Geforce VGAs are (factory) overclocked in GamersNexus. There goes your conspiracy theory.

Right because their test was from beta. That didn't stop you from using their charts in your OP though. So are you saying that you also posted invalid data in the OP?

Are you going to respond to the rest of my post?

 

[Adul]

Also, lets look at the real story here. A low clocked 980Ti is almost on par with an OC'd GTX 1080? Why is that even possible? Put a normal OC on the GTX 980Ti to even 1400mhz and it's above an OC'd GTX 1080. Talk about a failure.

Its a different story at 4k resolutions, there it is outpaced. So I would chalk this up to some other limitation besides the cards.

 

[Tumaras]

The problem with the 390/390x (and I'm an owner) are the minimum framerates. It doesn't really matter if it can peak higher if it's frequently dipping under 40 for minimum framerate. That and that the 390/x can literally raise room temp 10 degrees, and that's not exaggerating.

Thankfully the 480 should solve the heat problem, and hopefully it can improve on minimum framerates. This is more example of why the 4GB $199 version is more marketing than something most gamers would want. The extra $29 for the 8GB is well worth it if you go 480.

 

[Mercennarius]

The problem with the 390/390x (and I'm an owner) are the minimum framerates. It doesn't really matter if it can peak higher if it's frequently dipping under 40 for minimum framerate. That and that the 390/x can literally raise room temp 10 degrees, and that's not exaggerating.

Thankfully the 480 should solve the heat problem, and hopefully it can improve on minimum framerates. This is more example of why the 4GB $199 version is more marketing than something most gamers would want. The extra $29 for the 8GB is well worth it if you go 480.

I mean...the 390X is approaching 3 years old architecture wise. Its honestly holding its own quite well considering....an overclocked 390X can still game pretty good and will likely still be faster than 480X.

 

[tential]

The problem with the 390/390x (and I'm an owner) are the minimum framerates. It doesn't really matter if it can peak higher if it's frequently dipping under 40 for minimum framerate. That and that the 390/x can literally raise room temp 10 degrees, and that's not exaggerating.

Thankfully the 480 should solve the heat problem, and hopefully it can improve on minimum framerates. This is more example of why the 4GB $199 version is more marketing than something most gamers would want. The extra $29 for the 8GB is well worth it if you go 480.

Most people purchasing a $200 gpu are not expecting to be able to play at "hyper" settings. Jesus I mean people spent 350+ on the gtx 970 for a 1080p card. This is an entry level 1080p card. Does the gtx 960 run ultra settings today?

Crazy how the goal posts shift where now amds 200 gpu is supposed to be able to handle the craziest of settings.

 

[antihelten]

Something weird is going on with those hyper benchmarks. Normally when you're VRAM limited your framerate tanks, but you shouldn't experience that the game outright refuses to run.

I suppose it is possible that the game has a built in VRAM check when you turn on hyper settings, but that means that it's effectively impossible to say exactly how much VRAM is necessary. You quite obviously don't need 8GB for 4K + hyper settings, seeing as the 6GB 980 Ti performs exactly as you would expect (i.e. 5% slower than 1070 and 25% slower than 1080).

 

[Yakk]

Something weird is going on with those hyper benchmarks. Normally when you're VRAM limited your framerate tanks, but you shouldn't experience that the game outright refuses to run.

I suppose it is possible that the game has a built in VRAM check when you turn on hyper settings, but that means that it's effectively impossible to say exactly how much VRAM is necessary. You quite obviously don't need 8GB for 4K + hyper settings, seeing as the 6GB 980 Ti performs exactly as you would expect (i.e. 5% slower than 1070 and 25% slower than 1080).

Very true, the VRam "brick wall" should tank you to single digit fps. 6gb shouldn't be enough per nvidia, but 8gb seems exaggerated per logic.

Something weird going on?

Edit: unless it's uncompressed textures?

 

[CakeMonster]

So, can we expect no visual difference like what is usual with these kinds of modes, or have they actually managed to give us something actually noticeable and better?

 

[amenx]

So, can we expect no visual difference like what is usual with these kinds of modes, or have they actually managed to give us something actually noticeable and better?

Its a 'marketing mode', to entice more people to buy new gen 8gb cards.

 

[Thinker_145]

Whatever it is the fact is we can't do anything about it. Those who already have 4GB cards it's okay you can still enjoy games for a long time. But anybody buying a 4GB card in 2016 is just being stupid even for 1080p.

 

[96Firebird]

Something weird is going on with those hyper benchmarks. Normally when you're VRAM limited your framerate tanks, but you shouldn't experience that the game outright refuses to run.

I suppose it is possible that the game has a built in VRAM check when you turn on hyper settings, but that means that it's effectively impossible to say exactly how much VRAM is necessary. You quite obviously don't need 8GB for 4K + hyper settings, seeing as the 6GB 980 Ti performs exactly as you would expect (i.e. 5% slower than 1070 and 25% slower than 1080).

Very true, the VRam "brick wall" should tank you to single digit fps. 6gb shouldn't be enough per nvidia, but 8gb seems exaggerated per logic.

Something weird going on?

Edit: unless it's uncompressed textures?

It ran, just had stability problems.

From the review, right under the chart that shows the 970 and Fury X at 0:

"The Fury X and GTX 970 refused to run with any stability – we think that's a VRAM limitation – and so the chart only shows the 390X, 1080, 980 Ti, and 1070."

If you really do run out of VRAM, the game will essentially be a slideshow. This happened to me when I tried maxing out Far Cry 4 with AA @ 4K using DSR, just to see what would happen. It was difficult to even get to the menu to quit...

 

[tential]

So they think it's VRAM aka not confirmed. For all we know, the GTX 980 and R9 390/x could be fine. Maybe even the GTX 970 with it's VRAM situation gets sorted out. All of this cheering as a game is rolling out is silly. Who cares?

Let's see if there is any driver response from Nvidia/AMD.

 

[tential]

No it wouldn't, even if you add +20% to Geforce GTX 980 Ti @ 1228 MHz at 1080P Hyper / 4K High:

And I'm pretty sure that's not the best example of Geforce GTX 1080 overclocking out there.

It's above the boost clock, hence it's OCed. 1873 is an OC. So a 980Ti at a normal OC would beat an OC'd GTX 1080. Pretty amazing what that shows. It's odd too that Nvidia doesn't recommend Hyper on the GTX 980Ti when it's clear the GTX 980Ti is clearly able to run Hyper. I guess they have to downplay the GTX 980Ti as much as possible to move GTX 1080s....

 

[selni]

It's above the boost clock, hence it's OCed. 1873 is an OC. So a 980Ti at a normal OC would beat an OC'd GTX 1080. Pretty amazing what that shows. It's odd too that Nvidia doesn't recommend Hyper on the GTX 980Ti when it's clear the GTX 980Ti is clearly able to run Hyper. I guess they have to downplay the GTX 980Ti as much as possible to move GTX 1080s....

Advertised boost clocks are a little shady in that they're expected averages in "typical" use. They're not minimums or maximums - individual cards may do better/worse and are definitely not limited to never boost past the advertised number. The line between gpu boost and overclocking is really blurry.

 

[boozzer]

I am curious, how many of you guys buys into this?

 

[StrangerGuy]

So, can we expect no visual difference like what is usual with these kinds of modes, or have they actually managed to give us something actually noticeable and better?

Glad to see at least somebody who gets the point.

 

[Azix]

in other news, I'm seeing over 50-60fps on ultra (high folliage) in witcher 3 with a 290x.

Eventually the nvidia cancer in the game will be cured and perf will go up.

Waiting or a comparison of this hyper setting with ultra.

 

[digitaldurandal]

Instead of just in words, I would like to see in screenshots. It seems to me that looking at Ultra vs Very High already is difficult to discern in most modern games. Is this going to be an even smaller boost to fidelity for an even larger performance drop relatively speaking?

That said, it is nice that they give the option for those who want it.

 

[iiiankiii]

I'm going to assume this is GameWorks in full effect.

 

[Yakk]

I'm going to assume this is GameWorks in full effect.

Isn't the Frostbyte engine one of, if not the most efficient and optimized game engine today? I am very curious at how much gameworks stuff/garbage had be thrown into this Hyper option to slow down even this game engine so much.

 

[alcoholbob]

Also, lets look at the real story here. A low clocked 980Ti is almost on par with an OC'd GTX 1080? Why is that even possible? Put a normal OC on the GTX 980Ti to even 1400mhz and it's above an OC'd GTX 1080. Talk about a failure.

Not quite, a 1500MHz 980Ti, if gains were linear, would be about 95 fps on to the 1080's 102 fps.

So OC'ed 1080 at 1987 MHz is still about 7-8% faster than a 1500MHz 980 Ti.

 

[RussianSensation]

Isn't the Frostbyte engine one of, if not the most efficient and optimized game engine today? I am very curious at how much gameworks stuff/garbage had be thrown into this Hyper option to slow down even this game engine so much.

The game looks significantly worse than Starwars Battlefront (perhaps other than the lighting), but runs like a complete dog.

7970Ghz/280X easily got > 60 FPS at 1080p
R9 390 got > 60 FPS at 1440p
http://www.techspot.com/review/1096-star-wars-battlefront-benchmarks/

970 beating Fury X in Frostbite? Looks legit.

Looks like another AAA title ruined by GameWorks/NV Partnerships. Anyone keeping track of GameWorks success vs. fail/gimp ratio? I guess on the positive side it makes it very easy to know which games to buy for $5 or less when they have the GWs stamp on them.

 

[tential]

Not quite, a 1500MHz 980Ti, if gains were linear, would be about 95 fps on to the 1080's 102 fps.

So OC'ed 1080 at 1987 MHz is still about 7-8% faster than a 1500MHz 980 Ti.

100% quite. The 1080 is oced at 1837 and the 980ti oced is faster than that.

Quite a shame that an achievable oc on the 980ti is as fast as an oced 1080.

 

[Sweepr]

100% quite. The 1080 is oced at 1837 and the 980ti oced is faster than that.

Except... that's a Geforce GTX 1070. Hate to burst your bubble.

 

[tential]

Except... that's a Geforce GTX 1070. Hate to burst your bubble.

5% slower than a speed that isn't sustainable on the founders edition gpu is way too close. Last generation it was the 970 vs the 780ti.now it's the 1080 vs the 980ti. Quite the regression.

In a typical case the 1080fe would be slower than the 980ti in this test.