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

[ZGR]

Why no frame time tests? It would have been far more effective than just saying instability is found with low VRAM. I don't care about 100+ fps if the frametimes are erratic. Wish AMD's CFX was a big issue again just to see more frametime graphs, LOL.

And I agree, the screenshots make Hyper look like a waste of time. Every time a "Texture Pack" is released by a Developer, it is underwhelming.

 

[antihelten]

To clear up some confusion. Both Gamers Nexus and ComputerBase managed to run the game at Hyper @ 1080P/1440P with 4GB VGAs (same as Overclock3D)

Actually Gamers Nexus didn't manage to run the game at those settings with 4GB GPUs, hence them reporting 0.00 FPS as the results.

 

[ShintaiDK]

Its a real shame if you got the horsepower and not the VRAM. Because this is the new standard.

 

[antihelten]

Its a real shame if you got the horsepower and not the VRAM. Because this is the new standard.

Absolutely, but will that really ever be the case though?

I mean we know that RX 480 (and possibly some versions of GP106) will come out with 4GB, but those don't have the horsepower, so that's not an issue.

1070/1080 obviously have the horsepower, but there is no indication that Nvidia (or AMD when they finally get around to release a competitor) will actually offer these with anything less than 8GB.

 

[richaron]

Its a real shame if you got the horsepower and not the VRAM. Because this is the new standard.

Lol way to ignore all the data in this thread.

If by "standard" you mean statistically and objectively insignificant edge case scenario to boost sales; you may have a point.

 

[kraatus77]

Its a real shame if you got the horsepower and not the VRAM. Because this is the new standard.

so standard is based on 1 game ?

that means Asc/dx12 is the standard because there are more than 1 game actually taking advantage of it. k den

 

[mohit9206]

Absolutely, but will that really ever be the case though?

I mean we know that RX 480 (and possibly some versions of GP106) will come out with 4GB, but those don't have the horsepower, so that's not an issue.

1070/1080 obviously have the horsepower, but there is no indication that Nvidia (or AMD when they finally get around to release a competitor) will actually offer these with anything less than 8GB.

How do you know rx480 does not have the horsepower? Do you know something we don't? Do you by extension also mean to say 390 also doesn't have horsepower to utilize more than 4gb? How do you know this?

 

[kondziowy]

https://www.youtube.com/watch?v=6JhDRtA8Vg4
Digitalfoundry analysis 970 vs 390.

Pls guys tell me, because it is getting far beyond ridiculous. Every gameworks AAA title there is day one patch that brings new options not available before that breaks Radeon performance. Are those settings really hidden from AMD before game release or what?
GOW, Doom, ME:Catalyst - the same story every time.

 

[dark zero]

https://www.youtube.com/watch?v=6JhDRtA8Vg4
Digitalfoundry analysis 970 vs 390.

Pls guys tell me, because it is getting far beyond ridiculous. Every gameworks AAA title there is day one patch that brings new options not available before that breaks Radeon performance. Are those settings really hidden from AMD before game release or what?
GOW, Doom, ME:Catalyst - the same story every time.

Even worse? Their console version is a dissaster of epic proportions. Only works well on the NEW generation of nVIDIA, even the older ones (Kepler and now Maxwell) gets bugged hard.

 

[PhonakV30]

https://www.youtube.com/watch?v=6JhDRtA8Vg4


How can possible ? Running GTX 970 on Hyper Settings ? Didn't we hear that Hyper Setting requires at least 6GB Vram? according to AMD and NVIDIA GPU vs Games

gtx 970 in this test is running in +/- ultra settings, not hyper because of the option "GPU memory restriction" which reduces details if you reach vram limit, this game on hyper uses 7,5gb vram

If It's true then this Test is not fair.So many complain in comments about Low fps on AMD Card.

 

[Sweepr]

https://www.youtube.com/watch?v=6JhDRtA8Vg4


How can possible ? Running GTX 970 on Hyper Settings ? Didn't we hear that Hyper Setting requires at least 6GB Vram? according to AMD and NVIDIA GPU vs Games



If It's true then this Test is not fair.So many complain in comments about Low fps on AMD Card.

Strange indeed, maybe this is happening:

By default, a protective function in the option menu is active, which turns off some details automatically when memory shortage. The function is switched off for all tests.

 

[IEC]

Its a real shame if you got the horsepower and not the VRAM. Because this is the new double standard.

Fixed. :shiftyeyes;

 

[antihelten]

How do you know rx480 does not have the horsepower? Do you know something we don't? Do you by extension also mean to say 390 also doesn't have horsepower to utilize more than 4gb? How do you know this?

Based on the benchmarks Hyper settings don't become an issue for 4GB cards unless you combine it with 4K (Fury X gets 77.4 FPS at 1440P and 4 FPS at 4K)

At 4K and hyper settings, you need 980 Ti/Fury X level horsepower to even achieve the bare minimum of a 30 FPS average (a 980 Ti gets 31 FPS at 4K). And off course a lot of people wouldn't really consider a 30 FPS average playable, and if you wanted something like 40-45 FPS you would probably need 1080 level horsepower.

So unless you think RX 480 will be as fast as 980 Ti/Fury X, we can conclude that it will run at less than 30 FPS at 4K plus Hyper settings, and as such it doesn't have the necessary horsepower and it doesn't matter if it has 4GB or 8GB.

And yes by extension this also means that in this game a 390 doesn't have the horsepower to make use of more than 4GB VRAM either.

 

[PhonakV30]

Allright I watched DigitalFoundry's Youtube:

R9 390 is faster than GTX 970 on Ultra/Medium Settings by margin up to 12FPS.
R9 390 is slower than GTX 970 on Hyper Setting by margin up to 30FPS.

yet many people Think Nvidia performs better and blame on AMD.

 

[crisium]

Allright I watched DigitalFoundry's Youtube:

R9 390 is faster than GTX 970 on Ultra/Medium Settings by margin up to 12FPS.
R9 390 is slower than GTX 970 on Hyper Setting by margin up to 30FPS.

yet many people Think Nvidia performs better and blame on AMD.

Wow, really? We know Hyper was a very, very late addition just days before release. Looks like they snuck it by AMD to avoid their optimizations. No VRAM to blame in this comparison - just "Hyper" optimizations that were likely not shared with AMD. Or all was fair and AMD had full time and [H]yper settings simply use the Maxwell magic.

 

[PhonakV30]

Wow, really? We know Hyper was a very, very late addition just days before release. Looks like they snuck it by AMD to avoid their optimizations. No VRAM to blame in this comparison - just "Hyper" optimizations that were likely not shared with AMD. Or all was fair and AMD had full time and [H]yper settings simply use the Maxwell magic.

According to Computerbase

By default, a protective function in the option menu is active, which turns off some details automatically when memory shortage. The function is switched off for all tests.

So I think Didgitalfoundry forgot to disable "Gpu Memory restriction" in their Test.they should test again.

 

[Bacon1]

https://www.youtube.com/watch?v=6JhDRtA8Vg4


How can possible ? Running GTX 970 on Hyper Settings ? Didn't we hear that Hyper Setting requires at least 6GB Vram? according to AMD and NVIDIA GPU vs Games

If It's true then this Test is not fair.So many complain in comments about Low fps on AMD Card.

Watching the video I noticed the exact same thing. 970 is NOT running in hyper, those are ultra shadows. They are soft shadows instead of hard ones.

https://youtu.be/6JhDRtA8Vg4?t=74

 

[Sweepr]

crisium if you don't mind I'll post your comment from the other thread here:

Hyper mode is a stutterfest if you uncheck "GPU Memory Restriction" on my Fury, even when I tested only at 1080p it is always below 30fps. Not realistically possible to play the real Hyper with 4GB.

While using the GMR it is possible to use Hyper, and it uses basically the full 4GB (more than Ultra which is usually around 3.5-3.9 at 1440P).

I would guess Hyper with GMR is a mix of Ultra and Hyper, dynamically scaling more Ultra when it is low on memory. Given that I'm already fairly close to 4GB on Ultra, it's probably much more Ultra than Hyper on a 4GB card.

In the Digital Foundry video it is likely that they played with Hyper GMR enabled, and the 970 was actually running nearly entirely Ultra due to the dynamic scaling. With a full 8GB, the 390 was likely running the full Hyper.

DF has quite the reputation for analyzing any tiny PS4 vs Xbone detail difference. It is quite embarrassing they did not realize this, and are now likely spreading misinformation.

I agree, there's something wrong with that DF test.

 

[antihelten]

crisium if you don't mind I'll post your comment from the other thread here:

I don't think the stuttering is necessarily a VRAM issue though, stuttering seems to be a widespread issue, including for people with more than 4GB of VRAM.

I have seen reports of people with both 980 Ti and 390/390X reporting stuttering as well.

I haven't seen anyone with Titan X report stuttering though, so I suppose it's possible that Mirror's edge actually requires 12GB of VRAM to run perfectly maxed out with no stuttering. That would of course also rule out the 1070 and 1080, but then again Hyper textures appear to make zero visual difference so who cares.

 

[Cookie Monster]

https://www.youtube.com/watch?v=6JhDRtA8Vg4
Digitalfoundry analysis 970 vs 390.

Pls guys tell me, because it is getting far beyond ridiculous. Every gameworks AAA title there is day one patch that brings new options not available before that breaks Radeon performance. Are those settings really hidden from AMD before game release or what?
GOW, Doom, ME:Catalyst - the same story every time.

I didn't know Doom was a gameworks title?

 

[antihelten]

I didn't know Doom was a gameworks title?

It isn't to my knowledge. Also GoW didn't underperform on AMD because of some last minute inclusion of GW features, it underperformed because the DX12 port was rushed and more or less broken on AMD. After patches to GoW it performed just fine on AMD.

And of course most importantly the Hyper settings in ME:Catalyst aren't gameworks stuff to my knowledge, they are just higher res versions of all the normal features of the Frostbite engine.

 

[Maverick177]

Wasn't Doom's ultra settings added to the last minute?

 

[Erenhardt]

Wasn't Doom's ultra settings added to the last minute?

Ultra was prometed by nvidia before release. Beta had no Ultra and run fine. After release the performance on ultra was a disaster. Ain't nothing gameworks about that /s

The story seems to be similar, like kondziowy noticed.

https://www.youtube.com/watch?v=6JhDRtA8Vg4
Digitalfoundry analysis 970 vs 390.

Pls guys tell me, because it is getting far beyond ridiculous. Every gameworks AAA title there is day one patch that brings new options not available before that breaks Radeon performance. Are those settings really hidden from AMD before game release or what?
GOW, Doom, ME:Catalyst - the same story every time.

 

[antihelten]

Wasn't Doom's ultra settings added to the last minute?

Not to my knowledge, they were just disabled for the multiplayer beta. More importantly though, the Doom nightmare settings are not related to gameworks either.

Ultra was prometed by nvidia before release. Beta had no Ultra and run fine. After release the performance on ultra was a disaster. Ain't nothing gameworks about that /s

Performance was a disaster until AMD released a hotfix driver (16.5.2.1)

 

[nenforcer]

This is the second DICE game in a row to drop support for AMD Mantle after Star Wars: Battlefront very disappointing after they championed it.