[AMD_Robert] Concerning the AOTS image quality controversy

[AtenRa]

If the issue is now fixed, I am sure that all those sites will now update all their reviews with the fixed drivers, right?

Hardly, many dont have the time and others dont have the cards now.

We may see new results with custom GXT 1080 when available.

 

[Silverforce11]

I don't think that was the intention. Remember, when it first came out that using AC significantly hurt NV's performance everybody was surprised since NV had said that Maxwell could do it and they exposed AC in their drivers.

Also, AMD may have had a marketing agreement with Stardock, however NV had complete access to the AOTS source code and could optimize it however they wanted as long as it didn't hurt AMD's performance.

They added the option to disable it for hardware not capable so there's no performance degradation. That's going above & beyond.

We've been over this.

It's too bad Pascal didn't add functional hardware for it, so more devs can add performance enhancements for their games... I guess we'll have to wait til Volta for it.

Btw, this bug isn't even about AMD vs NV performance. It's the 1080 is made to look artificially better in DX12 than the 980Ti.

 

[Magee_MC]

Btw, this bug isn't even about AMD vs NV performance. It's the 1080 is made to look artificially better in DX12 than the 980Ti.

I would be careful making any definitive statement like that. Unless you've seen head to head benches of the press driver and the release driver we don't know that for a fact. It's possible that the 1080 was gaining a few points in benches because of the driver, but that doesn't necessarily mean that it was made to do that. It's possible that it was a genuine bug that landed in NV's favor. It's also possible that it gained almost nothing.

What makes me doubt that it was intentional was that it was in the press version. In order for it to make any sense that NV did it on purpose there had to be a meaningful gain, otherwise the entire concept is just stupid. If there was a measurable gain, then once the release drivers were out that meaningful difference would have been noticed. That would have led people to looking as to why it existed and blown the whole thing up.

NV does many things that I don't agree with, but the one thing that I have never thought of them as is stupid. This would have been an insanely stupid risk to take with a new launch. It's either enough of a boost to be worthwhile and noticible and thus easier to spot, or it's not enough to be noticed and hence not worth the risk.

 

[Silverforce11]

The fact that it's in the press version means nobody would mention it at all cos the press are blind/deaf to these issues. They don't even do side-by-side IQ comparisons anymore.

If it was in a normal driver for gamers, many would notice the problem. I mean if AMD didn't use the press drivers, none of us would have known about the 1080 skipping shaders.

 

[renderstate]

Magee, you are wasting your time. Logic and reason have no power here. AMD can do no harm while NVIDIA is guilty by definition until proven innocent. And if you do prove they have done nothing malicious they won't you listen or believe to you anyway.

 

[littleg]

I would be careful making any definitive statement like that. Unless you've seen head to head benches of the press driver and the release driver we don't know that for a fact. It's possible that the 1080 was gaining a few points in benches because of the driver, but that doesn't necessarily mean that it was made to do that. It's possible that it was a genuine bug that landed in NV's favor. It's also possible that it gained almost nothing.

What makes me doubt that it was intentional was that it was in the press version. In order for it to make any sense that NV did it on purpose there had to be a meaningful gain, otherwise the entire concept is just stupid. If there was a measurable gain, then once the release drivers were out that meaningful difference would have been noticed. That would have led people to looking as to why it existed and blown the whole thing up.

NV does many things that I don't agree with, but the one thing that I have never thought of them as is stupid. This would have been an insanely stupid risk to take with a new launch. It's either enough of a boost to be worthwhile and noticible and thus easier to spot, or it's not enough to be noticed and hence not worth the risk.

You really can't see any gains to be had from sending out a higher performance driver to the people who write articles about your product, post benchmarks that everyone looks at and references and inform peoples purchasing intentions?

 

[3DVagabond]

Magee, you are wasting your time. Logic and reason have no power here. AMD can do no harm while NVIDIA is guilty by definition until proven innocent. And if you do prove they have done nothing malicious they won't you listen or believe to you anyway.

AMD is not the topic. It's nVidia reducing IQ to gain an advantage in review benches.

 

[Minkoff]

Magee, you are wasting your time. Logic and reason have no power here. AMD can do no harm while NVIDIA is guilty by definition until proven innocent. And if you do prove they have done nothing malicious they won't you listen or believe to you anyway.

(OT)
It's interesting you say that, but it's kinda funny. When you try to steer a thread your way, sometimes it turns around and bites you.

These are some of your comments in this section. I don't care to re-read them all but...

...in an application everyone knows heavily favors it showing numbers that make no sense..well, we immediately know you are in trouble.

- Oxide has said numerous times, that they work with all vendors. Are they guilty, without proven innocent? Or do you mean, that DX12 favors AMD?
You appear to be additionally certain, that AMD is in trouble? How you got that I have no idea.

With regards to future proof HW I'd be more concerned about Polaris apparently not having nothing new to accelerate VR applications

- Again, we do not have any data, architecture insights, software insights but you seem certain they got nothing new.

GCN still doesn't support like conservative rasterization.

- and again...

...Also why did AMD use a driver that they were not even supposed to have?

- You have answer to the question of the thread, yet you tried to steer away. The above implies that AMD is guilty of stealing secret drivers?

Yep... Let's better get back on topic and stop with the accusations, shall we?
End (OT)

 

[selni]

AMD is not the topic. It's nVidia reducing IQ to gain an advantage in review benches.

Or you know, prerelease drivers having a bug. Would hardly be the first time.

 

[3DVagabond]

Or you know, prerelease drivers having a bug. Would hardly be the first time.

That's always possible. Just like it's always possible there are more cheats hidden in these drivers just needing to be exposed.

 

[Leadbox]

Magee, you are wasting your time. Logic and reason have no power here. AMD can do no harm while NVIDIA is guilty by definition until proven innocent. And if you do prove they have done nothing malicious they won't you listen or believe to you anyway.

There's nothing logical or reasonable about downplaying the issue and deflecting by asking pointless questions like where AMD got said driver from. A genuine consumer should be happy this came out. If this never got caught you'd be blissfully ignorant of the true nature of Pascal's DX12 capability.

 

[96Firebird]

Guys, stop downplaying this huge issue on this revolutionary game that has sold millions!

 

[3DVagabond]

This has potential to be a huge issue. Let's not just attempt to sweep it under the rug. It needs to be checked out.

 

[Minkoff]

Guys, stop downplaying this huge issue on this revolutionary game that has sold millions!

If it is used in benchmarks, it has been seen by millions. If there is a problem with something, is should be discussed and answers found.

And what has the amount of sold units, have to do with quality?

 

[Yakk]

So wait this .19 driver was released with the 1070, not the 1080?

Nevermind AMD, it could possibly make the 1070 look better than nvidia's own 980 cards.

 

[Leadbox]

Guys, stop downplaying this huge issue on this revolutionary game that has sold millions!

Issue is bigger than the game and you know it but yes, lets all make like it ain't nothing while we wait for that driver that enables async compute :awe:

 

[96Firebird]

Sorry what was I thinking, I should just read reviews and not play the actual games with my video card...

 

[Minkoff]

Sorry what was I thinking, I should just read reviews and not play the actual games with my video card...

(OT)
When you went and bought that EVGA 970, did you look up for reviews and benchmarks or did you blindly went and buy it?
After that, do you upgrade drivers to improve performance and visuals, or are you on stock drivers yet?

That would answer alot...

 

[3DVagabond]

Issue is bigger than the game and you know it but yes, lets all make like it ain't nothing while we wait for that driver that enables async compute :awe:

We'll just add it to the long list of nVidia transgressions until someone gets to the bottom of it.

 

[Leadbox]

Sorry what was I thinking, I should just read reviews and not play the actual games with my video card...

non sequitur much :thumbsup:

 

[showb1z]

Sorry what was I thinking, I should just read reviews and not play the actual games with my video card...

Ok, so benchmarks don't matter, got it. Why are we all wasting our times on this forum again?

btw, were you as happy as I was when it turned out Nvidia had lied about the 970 specs? At least I got a free upgrade to a better card out of it.

 

[96Firebird]

(OT)
When you went and bought that EVGA 970, did you look up for reviews and benchmarks or did you blindly went and buy it?
After that, do you upgrade drivers to improve performance and visuals, or are you on stock drivers yet?

That would answer alot...

Since you asked, I went off the reviews from the launch night and sold my 770 before the price tanked.

Been a pretty good buy for me, and I haven't upgraded my drivers since sometime in 2015...

Any other questions?

 

[Minkoff]

Since you asked, I went off the reviews from the launch night and sold my 770 before the price tanked.

Been a pretty good buy for me, and I haven't upgraded my drivers since sometime in 2015...

Any other questions?

So, you make purchasing decision based on reviews and not actual gaming done by you? And you watch benchmarks from games, that you do not have or play.
I'm certain, if there was a negative part in the reviews, you would have held your switch from 770 to 970.

So, you do not play a lot of new and different games, as most of them have first day drivers? Or you don't care about visuals and performance...or...

No, no further questions, as I honestly do not care...plus it's way offtopic.

 

[positivedoppler]

This has potential to be a huge issue. Let's not just attempt to sweep it under the rug. It needs to be checked out.

I'm a little late catching up to this so can someone fill me in on what the last 4 pages of debate is over.

Is the issue a debate over rather or not the drivers provided by Nvidia to the press artificially inflated frame rates by sacrificing image quality.

or

Is the issue a debate over if Nvidia purposely try to pull a fast one by intentionally providing rigged drivers hoping nobody would ever notice?

 

[ShintaiDK]

I'm a little late catching up to this so can someone fill me in on what the last 4 pages of debate is over.

Is the issue a debate over rather or not the drivers provided by Nvidia to the press artificially inflated frame rates by sacrificing image quality.

or

Is the issue a debate over if Nvidia purposely try to pull a fast one by intentionally providing rigged drivers hoping nobody would ever notice?

AMD selected the drivers. Its a driver not used by any reviews. And its a driver they selected despite of the retail driver being out.

So, ask AMD