AMD Nano Blacklist Situation

[Phynaz]

So back to the original subject, the dipweed in marketing that thought of this strategy will be looking for a job soon.

 

[Pariah]

Define significant. I personally don't think any game should be excluded but only if 10+ games are tested, using 5 games and 4 of those being GameWorks titles is not good review practice.

There is no way to come up with a cut off point as there is too much variability. You seem to have missed the outlier part of my comment. If the majority of games happened to be optimized for one platform, then that is tough luck for the other side. Dirt Showdown seemed to be a lone game that gave results not similar to any other game on the market. That's an outlier.

 

[Pariah]

Have you seen the benchmarks on Project Cars? That thing is ridiculously favoring Nvidia's arch.

No, I don't play that game. When perusing reviews at release to see if I wanted to get it, they all seemed to indicate that there were major performance issues for the PC version in general, not specific to either side.

 

[AnandThenMan]

There is no way to come up with a cut off point as there is too much variability.

I agree, but then you say this:

Dirt Showdown seemed to be a lone game that gave results not similar to any other game on the market. That's an outlier.

So if Showdown is the outlier, there had to be a cut off point so what is it?

 

[Pariah]

I agree, but then you say this:

So if Showdown is the outlier, there had to be a cut off point so what is it?

You would have to ask TR what there cutoff point is. I wasn't involved in the decision. If I was doing reviews, I wouldn't give an average, so this scenario would never be an issue to me.

 

[Phynaz]

Roy Tayor when he worked for Nvidia;

"The UK is the only place in the world where anyone talks about AMD or ATI", said Roy Taylor. Surprisingly enough, Taylor's spreadsheets outlined the graphics market share for the last year, but ATI's chips were not there. However, it is estimated that ATI, Intel and Nvidia managed to pump up 366 million graphics chips during last year. When asked about why ATI was removed from the charts, Taylor said that "no one cares".

Does anyone think he has a credibility issue?

 

[3DVagabond]

Well, to be fair, they need widespread adoption of HBM that way it'll get cheaper for them to manufacture as well, which allows more supplies.

AMD does it with everything. They post the source code for all of their rendering features. They gave away Mantle.

Do you think nobody would use HBM if there was a royalty attached? Seriously?

 

[boozzer]

I'm still looking for proof that the game was removed due to "poor Nvidia performance".

Maybe you can explain how they can be biased against AMD and still include Mantle in their benchmarks.

I am not nice like rs, hence super hard to troll as a target. I also don't waste time ^_^

Infraction issued for member callout.
-- stahlhart

 

[Ryan Smith]

I wonder if anandtech had to negotiate to get 1920x1080 results in there.

Huh? No. There's no negotiating of any kind about what's in our GPU reviews. I publish what I want to publish. In this case I didn't feel like 4K results would be especially useful, so I made sure to throw in 1080p as well.

 

[Erenhardt]

Huh? No. There's no negotiating of any kind about what's in our GPU reviews. I publish what I want to publish. In this case I didn't feel like 4K results would be especially useful, so I made sure to throw in 1080p as well.

Did nvidia send you a reviewers guide how to test AMD Fury Nano, the same thing they did when hawaii lunched?

Anyone benching minecraft or league of legends?!

 

[Phynaz]

I am not nice like rs, hence super hard to troll as a target. I also don't waste time ^_^

Or in other words, you don't have a rebuttal. Being proven wrong is fine, just don't make it worse by trying to deflect.

 

[boozzer]

Or in other words, you don't have a rebuttal. Being proven wrong is fine, just don't make it worse by trying to deflect.

it is ok. did anyone ever tell you that? it is ok

 

[Ryan Smith]

Did nvidia send you a reviewers guide how to test AMD Fury Nano, the same thing they did when hawaii lunched?

Anyone benching minecraft or league of legends?!

I honestly can't tell if this is serious or in jest... but either way the answer is no.

 

[3DVagabond]

I honestly can't tell if this is serious or in jest... but either way the answer is no.

Not to be a stirrer, but [H] is suggesting (rather strongly) that anyone who did get a card are glossing over the weaknesses and aren't reviewing objectively.

This is what I was thinking..... "The clearest bunch of ******** I have read today Roy. Thanks for that pathetic reply of trying to not look like AMD is fully cherry picking review sites for "good" Nano reviews.

The reason I think AMD is NOT sampling all its usual review sites is because of thinking like Roy Taylor is infecting the company. Roy wants AMD to duck and cover and only sample the journalists that they feel as though they have a sure chance of getting a "good" review with.

Just like they did with the "Paper Launch" preview, where they draw every conclusion about the card except for the benchmarks without ever testing it, Brent is critiquing the reviews before they are even written.

All I want to say is, when the "reviews" get published, look closely at what is reported, what is covered, and the comparisons made. I think how the clock speed operates, and what the real-world clock speeds are might be glossed over in some reviews, or even incorrect if not tested right.

Something we do, that I know not everyone does, is test cards after the card as "warmed up" while gaming. Typically 15+ minutes in from gaming the GPU will have warmed up and clock speeds will be different from just testing the card right out of the gate in a short benchmark. A short benchmark is going to show higher results compared to running the card for thirty minutes and then playing a game. The clocks will be different, the results will be different. We test cards after they have gone through this warm up period, which takes extra time, but it is worth it cause you get real-world results.

I think some points will not be made about how Fury X (also a small card) wiill fit into the same cases as the Nano. Granted, you need a space for the rad and fan, but most SFF cases, even ITX, have room for this. I have seen Fury X ITX builds on youtube and the Internet. It is possible. I want to see this talked about. mATX cases for sure can house Fury X.

Finally, the price being the same as the Fury X, I have a feeling this will be downplayed. I also have a feeling availability might be downplayed.

We shall see, but it will be interesting to see if things are left out, downplayed, swept under the rug, and just not discussed when it comes to Nano in the reviews to come. These are all important factors.

"Fair" reviews would keep these topics in-mind and discuss them in reviews.

I understand that they think they do the best reviews, as I'm sure most sites do, but not only did Kyle question the integrity of any site that does get a card, Brent basically said they are going to use sites that don't do thorough informative reviews. And that's assuming he meant they won't review those details because of ignorance or laziness, if not out and out collusion.


On Topic:
With AMD's actions and what some posters here have insinuated when questioning you, I think that maybe reviewers have to look at themselves for why what is happening is. In many peoples minds the tech press is not objective and follows the "support" trail of these IHV's rather than putting the reader's best interests first. Whether this is actually happening or not could never be proven one way or the other, but just the backlash over it should raise a big red flag for the sites.

Free hardware, free trips, GTX460 FTW situations, etc... all have to be looked at too and kept to a minimum. Because it is real easy to become beholding to a supplier's generosity, or be taken by their competitive tactics.

 

[Goatsecks]

I understand that they think they do the best reviews, as I'm sure most sites do, but not only did Kyle question the integrity of any site that does get a card, Brent basically said they are going to use sites that don't do thorough informative reviews. And that's assuming he meant they won't review those details because of ignorance or laziness, if not out and out collusion.

The quote from brent is spot on.

It feels that AMDs new line-up is tripping over its own feat and no one seems to be talking about this:

If fury-X is a 4k card, then why does the 390x need 8gb of memory? Currently, I do not believe that any single GPU (from AMD or Nvidia), will give a reasonable 4K experience. Consequently, a multi-gpu setup is preferable. In the case of crossfire it would be crazy to choose crossfire fury-x over 390x, since GPU memory is not pooled. Yet all the review sites push the whole Fury lineup as 4K cards.

Echoing Brents comments: (imo) at 4K it seems that the whole fury line up is surplus to requirement. :\

 

[Glo.]

Why are you quoting and replying to my post to Russian? Did we just find somebody who has a second account?

You are jumping to conclusions too fast. And I was just pointing that to you. That is your business, but dugs your credibility, also. Especially when you jump to conclusions about people who you are responding to.

 

[3DVagabond]

The quote from brent is spot on.

It feels that AMDs new line-up is tripping over its own feat and no one seems to be talking about this:

If fury-X is a 4k card, then why does the 390x need 8gb of memory? Currently, I do not believe that any single GPU (from AMD or Nvidia), will give a reasonable 4K experience. Consequently, a multi-gpu setup is preferable. In the case of crossfire it would be crazy to choose crossfire fury-x over 390x, since GPU memory is not pooled. Yet all the review sites push the whole Fury lineup as 4K cards.

Echoing Brents comments: (imo) at 4K it seems that the whole fury line up is surplus to requirement. :\

I don't see how that has anything to do with what I posted. One of us is totally misunderstanding each other.

On what you've said though, I haven't seen a single review, on H or anywhere else that backs up what you just posted.

The 390's have 8gig of RAM purely to differentiate and add perceived value over the 290's.

 

[Flapdrol1337]

Blocking TR is not cool.

When nvidia contacted TR on 4K 7990 issues they made an item on how nvidia was a bit agressive a few weeks before the 290x review.
They were the only site with a positive r9 285 review.
When they tested freesync they said it worked as well as gsync and nvidia should start supporting it asap.
They also posted a result without project cars after the fury x review, TPU did this too, blocking them is also uncalled for.

Pcper just ran with the 7990 4k story, when pcper got freesync they were trying real hard to find smoothness issues at low refreshrate (only real issue was ghosting imo, now fixed). But pcper got their nano just fine.

 

[Goatsecks]

I don't see how that has anything to do with what I posted. One of us is totally misunderstanding each other.

Sorry my fault, I should have quoted the Brent quote, I was just being lazy.

On what you've said though, I haven't seen a single review, on H or anywhere else that backs up what you just posted.

The 390's have 8gig of RAM purely to differentiate and add perceived value over the 290's.

Similarly how Brent suggests that the Fury-x is stepping on the toes of the Nano I was suggesting that the 390x is stepping on the toes of the Fury-x as a 4K solution (I understand the motivation for sticking 8gb of memory into the 390's, but it has a knock on effect of making them a rather good choice for 4K when implemented in a crossfire setup).

Weather you think the Fury lineup (or the maxwell lineup) is suitable as a single gpu 4K solution is purely subjective. I do not think any results from AMD or Nvidia are particularly compelling at this resolution (for a single GPU!), I have this opinion from perusing the review sites.

Sorry for the confusion!

 

[3DVagabond]

Sorry my fault, I should have quoted the Brent quote, I was just being lazy.



Similarly how Brent suggests that the Fury-x is stepping on the toes of the Nano I was suggesting that the 390x is stepping on the toes of the Fury-x as a 4K solution (I understand the motivation for sticking 8gb of memory into the 390's, but it has a knock on effect of making them a rather good choice for 4K when implemented in a crossfire setup).

Weather you think the Fury lineup (or the maxwell lineup) is suitable as a single gpu 4K solution is purely subjective. I do not think any results from AMD or Nvidia are particularly compelling at this resolution (for a single GPU!), I have this opinion from perusing the review sites.

Sorry for the confusion!

It's cool.

No I don't think that a single GPU is ideal for 4K. Unless you don't mind No AA or FXAA (I don't see the point of all those pixels just to blur them together.) and maybe high instead of ultra shadows.

Dual Fury works typically as good or better than 980 ti @ 4K, even with less VRAM (Which hasn't shown to be a problem.). Then theirs always the promise of DX12 and pooling memory and other dual GPU goodness. Assuming the cards are capable enough in DX12 to take advantage. I'm in the camp of thinking that GCN will prove superior to Maxwell in DX12. YMMV.

 

[LTC8K6]

I don't see how that has anything to do with what I posted. One of us is totally misunderstanding each other.

On what you've said though, I haven't seen a single review, on H or anywhere else that backs up what you just posted.

The 390's have 8gig of RAM purely to differentiate and add perceived value over the 290's.

The 290X had 8gb of ram, though. Released November 2014.

 

[3DVagabond]

The 290X had 8gb of ram, though. Released November 2014.

Those were custom versions. They were 4Gig standard. It's all marketing gymnastics anyway.

 

[Goatsecks]

It's cool.

No I don't think that a single GPU is ideal for 4K. Unless you don't mind No AA or FXAA (I don't see the point of all those pixels just to blur them together.) and maybe high instead of ultra shadows.

Dual Fury works typically as good or better than 980 ti @ 4K, even with less VRAM (Which hasn't shown to be a problem.). Then theirs always the promise of DX12 and pooling memory and other dual GPU goodness. Assuming the cards are capable enough in DX12 to take advantage. I'm in the camp of thinking that GCN will prove superior to Maxwell in DX12. YMMV.

Pooled memory :wub:

I just can't get my head around the Fury-X being a 4K card, even in multi-gpu setups, due to the memory limitation. A fairly quick and ubiquitous uptake of DX12 and, critically, memory pooling is just to large a gamble to opt for fury cards.

Back on topic, as innovative as the fury lineup is, it is also very confused/confusing.

I commend Ryan for including 1080p results. 1080p results should have been included in the Fury-X reviews. Yet this was not the case for the majority of sites. :hmm:

 

[Attic]

Pooled memory :wub:

I just can't get my head around the Fury-X being a 4K card, even in multi-gpu setups, due to the memory limitation. A fairly quick and ubiquitous uptake of DX12 and, critically, memory pooling is just to large a gamble to opt for fury cards.

Back on topic, as innovative as the fury lineup is, it is also very confused/confusing.

I commend Ryan for including 1080p results. 1080p results should have been included in the Fury-X reviews. Yet this was not the case for the majority of sites. :hmm:

Good points. AMD's lineup is a bit schitzo. AMD is perhaps being a bit of a baby over their own failures or confusions with the Fury and Hawaii Rebrands. [H] honestly comes across as a bit over the top in this entire thing, but Kyle has always been straight up and Brent is solid as well. [H] along with Anandtech remain my only two trusted review sites, but i'm definately having a longer look at [H] after the Nano thing.

Beyond all this, the Fury Nano looks great as a 1440P or under card and priced at $499. I think AMD is trying to jam a square into a circle peg and then getting a bit upset that folks realize and say "hey guys, that doesn't fit where you are trying to place it". The Nano is still the most efficient card and is very powerful, big props there, AMD just IMO was trying to control the narrative toooooo much with the release.

 

[Headfoot]

Pooled memory :wub:

I just can't get my head around the Fury-X being a 4K card, even in multi-gpu setups, due to the memory limitation. A fairly quick and ubiquitous uptake of DX12 and, critically, memory pooling is just to large a gamble to opt for fury cards.

Back on topic, as innovative as the fury lineup is, it is also very confused/confusing.

I commend Ryan for including 1080p results. 1080p results should have been included in the Fury-X reviews. Yet this was not the case for the majority of sites. :hmm:

The easiest way to get your head around Fury 4k performance is to read Fury 4k performance reviews, which show it doing quite well. Not so much at 1440 and definitely not so much at 1080....