AMD Radeon R9 Nano Review by TPU...Not

[btarunr]

@btarunr , will TPU do a review after release?

I'm not sure. Getting a sample after NDA isn't difficult. But it has to be worth the effort.

 

[RejZoR]

Because many TPU reviews are far too nice. Ratings over 8 or 9 for so-so products. I've looked at them and seen many noting some glaring issues with a product, then that product getting a TPU recommendation/etc. It's like...seriously TPU? You state it has some major issues then you highly recommend it? I thought they might be getting some pay for good reviews but I've just seen it too many times that it is clear some of the people are just too lenient here.

The content of the reviews are good, but the TPU rating system and such is often crap. If a product has a major issue, it should reflect greatly in the score and also not garner a recommendation.

Well, the fact that ASUS Strix (and MSI Gaming) cards are so freaking massive that they won't fit into all cases is a prety big problem and yet they didn't deduct score because of it. Why? Because this is so freaking good card regardless of that. I mean they managed to make R9-390X a passive graphic card in idle and harsly audible in load. Same for GTX 980. Granted, I had to buy a new case because of it (coming from a miniATX case), but I still think it was worth it. Some would disagree however since buying a new case for graphic card isn't exactly something people often have in mind.

Dunno, you have to understand what you read and apply it to your specific scenario to get a context. Otherwise score will be meaningless for you like you said.

 

[Pill Monster]

AMD DX11 drivers are not as good as Nvidia but having both i know Mantle destroys Nvidia DX11 performance.
Ashes has nothing to do with Mantle, what it does show is problems Nvidia have with DX12 and the fact that despite all the noise about having better DX12 support than AMD they don't have Hardware ASync while AMD do. and that actually matters.

You sir, are an expert at strawman babblespeak.

Translation:

AMD DX11 drivers are not as good as Nvidia
No shit.


but having both i know Mantle destroys Nvidia DX11 performance.
A 290X gets more fps in BF4 than a GTX970. All other things being equal.


Ashes has nothing to do with Mantle,
Ashes supports Mantle, now called DX12.




what it does show is problems Nvidia have with DX12
Neither AMD or Nvidia fully support DX12, any feature supported will be GPU dependent.





despite all the noise about having better DX12 support than AMD they don't have Hardware ASync while AMD do. and that actually matters
When Hallock & Co claimed DX12 support, they lied. But all that doesn't matter now.....

 

[qubit]

@btarunr Looks like AMD are in a death spiral to bankruptcy or being bought out. I'll bet this thing massively underperforms so they're afraid to let reviewers get their hands on it as they know they'll get a beating.

They've been overpromising and underdelivering for years now so I wouldn't expect this product to be any different.

 

[cadaveca]

@btarunr Looks like AMD are in a death spiral to bankruptcy or being bought out. I'll bet this thing massively underperforms so they're afraid to let reviewers get their hands on it as they know they'll get a beating.

They've been overpromising and underdelivering for years now so I wouldn't expect this product to be any different.

That's the thing... who cares if it "underperforms"? Given that it takes a fully enabled core and smashes into a tiny space, some concessions must be given, and price is not one of them. It will be interesting to see if heat is the only factor that holds it back, or if it is limited by power consumption. Either way, since it does fit a lot of power in a tiny space, a price premium makes sense. There's really nothing for AMD to fear here.

It's just a shortage of cards, and a paper launch, just like the Fury and FuryX. No big deal. It only makes sense that it's be a bit lower than the Fury at least... and optimized for 4k (hence the added shaders). If it ends up being more than that, AMD has a clear-cut winner that NVidia cannot answer. If not, it will still be a good card.

 

[RejZoR]

If you'll be able to buy it. If they can't supply few ten samples to reviewers, how will they to thousands of buyers!?

 

[qubit]

That's the thing... who cares if it "underperforms"? Given that it takes a fully enabled core and smashes into a tiny space, some concessions must be given, and price is not one of them. It will be interesting to see if heat is the only factor that holds it back, or if it is limited by power consumption. Either way, since it does fit a lot of power in a tiny space, a price premium makes sense. There's really nothing for AMD to fear here.

It's just a shortage of cards, and a paper launch, just like the Fury and FuryX. No big deal. It only makes sense that it's be a bit lower than the Fury at least... and optimized for 4k (hence the added shaders). If it ends up being more than that, AMD has a clear-cut winner that NVidia cannot answer. If not, it will still be a good card.

Ok cad, that sounds plausible and I sincerely hope you're right. Still, for bta to put out an official announcement like this doesn't bode well and we've seen many paper launches in the past, but reviewers still got cards. Dunno, gotta wait and see.

 

[john_]

Why not TPU? How about our lack of a social media presence, and lack of youtube presence? That's enough reason...

There was also an editorial that every tech site on the planet copied and pasted as news. That editorial was talking about a company and it's inability to publish a WHQL driver for 6 months. All that buzz started from here. From that editorial. I am not kidding. Everybody copied THAT editorial.
Just a thought.

Ah! yes, there was also that other editorial with the not so kind tittle about a company not trusting it's own CPUs, but choosing CPUs from the competition for it's little Quantum project.
Just another thought.

But I don't think that AMD's practices will help. They only guaranty more editorials in the future.

 

[Tsukiyomi91]

I was wondering... did that Sony dude got silenced by the moderators or he just had enough that right-thinking forumers is considered too much to handle, with his knowledge he has lost the battle...?

 

[btarunr]

I was wondering... did that Sony dude got silenced by the moderators or he just had enough that right-thinking forumers is considered too much to handle, with his knowledge he has lost the battle...?

Because their posts insult nobody else but themselves.

 

[Tsukiyomi91]

@btarunr right... didn't like him from the start since all he blabbers is AMD AMD AMD... like a typical simpleton... bet his electricity bills goes double of mine on a yearly basis with that "massive heater" he has in his room/house...

 

[entropy13]

In terms of video cards reviews Tech Report and TechPowerUp are both tied at the top because they really complement each other.

While TR focuses only on a few games, they go thorough with it with the frame latencies/times and what-have-you.

For TPU, it's the sheer amount of video cards benchmarked at the same time, AND the number of benchmarks as well. Then there are the power consumption and noise values too.

So together, you can say you have a "complete picture".


Both are also bashed for being "pro-Nvidia" even after bta said that the Titan-Z isn't worth it at all, while Damage over at Tech Report aren't amused with Nvidia over G-Sync...lol

 

[uuuaaaaaa]

What a shame... TPU has some of the finest unbiased GPU reviews... In the mean time guru3d is getting a Nano.

 

[Tsukiyomi91]

Titan Z was an exception because of it's price tag, in which it didn't get much sales despite being a completely new beast that has 2 full-blown chip sandwiched together on the same PCB, super-massive VRAM & a really broad pipeline (bus interface). AMD's case however was the unwanted price tag ($650) for something that's not as fast as the GTX980 & uses a very new tech on uncharted waters, aptly named "HBM". Sad to say however, that new tech isn't as fast compared to the current tech, GDDR5 chips.

G3D's testing methods aren't as good as TPU's IMO.... so no surprise to see AMD or anyone owning it sends them for testing. All I know is that the bench won't be realistic (or unbiased) like TPU & all the fanboys who are in a cringe will rejoice coz the Green Team's base 980 or ITX 970 got it's arse kicked in a botched benchmark where the settings used aren't as uniform as TPU's.

 

[ypsylon]

At reported 650$ price tag it's so mad that only completely mental AMD fans will get it. I was interested in getting this, but simply NO. AMD behaves like many game developers. Gives us money first for pre-orders and subscriptions and we will perhaps give you software in semi-working order. After release you can kiss our corporate back side.

Prefer half the price Maxwell and with Pascal round the corner Nano can be swan song from AMD, because Pascal really appears to be beast of the architecture from few leaks/presentations released so far.

 

[remixedcat]

This sounds just like how amped dealt with me... not enough to give me, one of thier biggest and comprehensive testers, review units after the whole rta15 fiasco (even for just a couple weeks), yet give out to soccer mom type "reviewers" that don't even bench them, and then when someone does, in fact bench them they are nothing special and in fact, have WORSE range then thier competition... I guess amped hired some amd people to manage thier company since this seems so the same in every way, shape and form and I track both companies religiously, as I am an IT consultant.

 

[john_]

Sad to say however, that new tech isn't as fast compared to the current tech, GDDR5 chips.

It's faster. But memory bandwidth is not everything. You need a powerful GPU and perfectly optimized drivers to take advantage of the extra bandwidth.

G3D's testing methods aren't as good as TPU's IMO.... so no surprise to see AMD or anyone owning it sends them for testing. All I know is that the bench won't be realistic (or unbiased) like TPU & all the fanboys who are in a cringe will rejoice coz the Green Team's base 980 or ITX 970 got it's arse kicked in a botched benchmark where the settings used aren't as uniform as TPU's.

I think Guru3D got that card because of Afterburner software. They probably have strong contacts in the industry that will guaranty them to have an advantage over other sites. That being said, most people post benchmarks from TPU and Anandtech, rather than Guru3D.

At reported 650$ price tag it's so mad that only completely mental AMD fans will get it. I was interested in getting this, but simply NO. AMD behaves like many game developers. Gives us money first for pre-orders and subscriptions and we will perhaps give you software in semi-working order. After release you can kiss our corporate back side.

Prefer half the price Maxwell and with Pascal round the corner Nano can be swan song from AMD, because Pascal really appears to be beast of the architecture from few leaks/presentations released so far.

Nano is like Titan X. If there are people spending $350 extra on a Titan X for those 12GBs of memory and the - how much? 10%? - extra performance, then there will be probably enough out there to spend $300 more for that 10-20-30% extra that Nano offers over GTX 970 ITX cards. Of course Nvidia only have to come out with a GTX 980 ITX.

 

[btarunr]

I think Guru3D got that card because of Afterburner software. They probably have strong contacts in the industry that will guaranty them to have an advantage over other sites. That being said, most people post benchmarks from TPU and Anandtech, rather than Guru3D.

W1zzard's OCTool SDK powers several AIB OC software too, including Sapphire's (much bigger Radeon volume than MSI).

 

[Sony Xperia S]

W1zzard's OCTool SDK powers several AIB OC software too, including Sapphire's (much bigger Radeon volume than MSI).

I hope you guys will use the occasion to pay more attention on the drawbacks in GPU-Z too.

It was recently when I kindly ask to turn some work on improving the design of this application:

It will be nice when added Boost frequency field instead of the blank faded Shader field for AMD cards.
AMD cards come with Boost frequencies as well.

Also, it would be nice if you change somehow the Computing part - probably it would be better if you fade the CUDA and PhysX for AMD cards, and rearrange the order with putting the features present on ALL cards first (i.e first DirectCompute, then OpenCL and OpenGL, and then all the others like Mantle, Vulkan, TressFX, PhysX, CUDA).

Thank you!

http://www.techpowerup.com/forums/threads/gpu-z-feature-request.212644/

 

[john_]

W1zzard's OCTool SDK powers several AIB OC software too, including Sapphire's (much bigger Radeon volume than MSI).

Thanks, didn't knew that. Then I can see one more reason for W1zzard to get a card in advance, for testing purposes at least if not benchmarking.

 

[dorsetknob]

today the Postman delivered My Nano for me to review !!!! then i woke UP with an errection and realised it was just a wet Dream

 

[btarunr]

I hope you guys will use the occasion to pay more attention on the drawbacks in GPU-Z too.

It was recently when I kindly ask to turn some work on improving the design of this application:



http://www.techpowerup.com/forums/threads/gpu-z-feature-request.212644/

View attachment 67822

There's a forum dedicated to GPU-Z feedback. Post it there.

 

[Tatty_Two]

I hope you guys will use the occasion to pay more attention on the drawbacks in GPU-Z too.

It was recently when I kindly ask to turn some work on improving the design of this application:



http://www.techpowerup.com/forums/threads/gpu-z-feature-request.212644/

View attachment 67822

You make the assumption that because he has not stopped the hundreds of things he is doing to attend to your personal request that he 1. Is not going to at least look at the feasibility and value of the suggestion 2. Is it is even possible. All I am saying here is sometimes things like this go into a "to do list" usually that list is dictated by priorities, if the list has several hundred things on it then it may not happen for some time, damn I suggested to my work in 2012 a simplified system for some of the data capture and management of information we use on a daily basis, they introduced it last week! Then again of course it may not ever happen, it's not often you find anything totally free being used by millions of people changed because just one or two think it should be, I suppose I would have to ask myself why it appears to matter so much to you when it does not to me? At a guess I would think that I have probably owned far more AMD graphics card than many.

 

[Assimilator]

First [H]ardOCP, now TPU... at this point I'm wondering if ANY review sites will be getting Nanos. I can only see one of three reasons for this, from least to most likely:

• AMD is so badly in the poop, finances-wise, that they can no longer afford to send free samples to reviewers.
  
• AMD has gone full retard and decided to shoot themselves in the foot, with a bazooka, by only sampling cards to sites that they believe will give positive reviews.
• Fiji yields are far, far worse than anyone has anticipated. Fewer Fijis = much fewer low-voltage Fijis suitable for Nano = massive supply problems.
• Nano doesn't perform anywhere near as well as AMD has claimed.

Any which way you look at it though, this is a massive issue for AMD, and the root problem is that they just aren't honest with the press and their customers. Every time they delay a launch or release a product that doesn't perform as claimed, they lose more and more trust. Soon they won't have any left, and that will be the end.

 

[remixedcat]

so they'd rather save a few thousand bucks than possibly be out hundreds of thousands, if not, millions from unconfident customers not purchasing said products at all?