[UPDATE 3 - Sapphire's Reference Costs $649 USD] AMD Reveals R9 Nano Benchmarks Ahead of Launch

[HKZeroFive]

AMD has a few days before the official launch of the Radeon R9 Nano released a few internally created benchmark values. The compare the mini graphics card with the older Radeon R9 290X and the current flagship Radeon R9 Fury X The Mini-PCs for imaginary Radeon R9 Nano achieves a higher frame rate than the Radeon R9 290X and to work almost twice as efficient. the Radeon R9 Fury X is struck.

The readings are from the Hot Chips 27, an annual technology conference, which is currently being held in Cupertino, California. Concretely compares AMD Radeon three-R9-cards in the game Far Cry 4, but without going into the graphics settings. However, the achieved frame rates speak clearly for UHD-resolution (3840 x 2160 pixels) with ultra high details. The Radeon R9 290X provides just over 30 who Radeon R9 Nano about 33 and the Radeon R9 Fury X around 42 fps.

When power AMD does not mention numbers, but only relative values: Starting from a Radeon R9 290X as a base to the Radeon R9 Fury X represent about one-third more frames per second per watt and the Radeon R9 Nano by a factor of 1.9 more efficient expected. Carefully folded in absolute numbers, is likely AMD's upcoming mini graphics card require at least FC4 benchmark shown less than 150 watts of energy.

According to the manufacturer, the typical power consumption is 175 watts, installed is a 8-pole connector. The theory provides 150 watts, the PCIe-Slot again 75 Watt. The Radeon R9 Nano measures 15 cm and uses a chip with full Fiji 44 Compute Units. The clock is likely in view of the performance lie at about 850 MHz.

As always, I'll remain skeptical of these numbers since they're from AMD directly, but it gives us a rough idea of what the performance of the Nano will be like. I quite like the fact that it has a small form factor, perfect for mini-ATX builds, but is better than the 290X (although that was to be expected). But I'll be waiting until Greenland comes out .

The source is in German, so have fun reading it. Google Translate to the rescue!

UPDATE

Claims that the R9 Nano is 30% faster than the GTX 970 mITX. Not surprising, they should really be comparing it against the 980, but you know, advertising. However, it gives us a rough idea of the performance.

Update Sauce

Update 2

Bunch of advertisements AMD has just released, with claims of a 75 degree target operating temperature and is 42dBA loud. Remember, these are claims directly from AMD, so they may be biased. As always, wait for actual review to come out.

[spoiler=AMD's R9 Nano Performance Claims]

Update 2 Sauce

Update 3

Sapphire is first to launch the R9 Nano as a reference card. Rumoured release on September 7th with a price of $649 USD.

Sauce: wccftech

[ZetZet]

Not benchmarks. Pictures made in paint.

[ANewFace]

If it's from AMD then these should be taken with a grain of salt. As we saw their benchmarks had the Fury X crushing the 980 ti, but in reality the 980 ti beats it

[79 Potatoes]

I'll wait till it comes out and for reviewers for real life power consumption too.

[Misanthrope]

I believe it mainly because they left out Nvidia cards this time.

[ICantThinkOfAnyGoodName]

I hope that AMD doesn't lie like they did with the Fury X.

[Tomsen]

I'm excited to see AMDs R9 nano in action.

Its efficiency should be quite good, as obviously it will run at lower clockspeed (heard around +800Mhz should be the sweetspot, in regards to efficiency).

[marldorthegreat]

I hope that AMD doesn't lie like they did with the Fury X.

They didnt lie. There benchmarks were a bit biased. Showing an area of the game where the AMD card did better.

[othertomperson]

If it's from AMD then these should be taken with a grain of salt. As we saw their benchmarks had the Fury X crushing the 980 ti, but in reality the 980 ti beats it

 

They're "boasting" a massive 2 fps more than a 290X. I think we can probably trust their honesty this time.

[Doobeedoo]

Seems to fit the spot.

[Megahurt]

Fury nano will be a great card, let's hope it won't be overpriced.

[Briggsy]

Looks like the 175W soft-limit is real. If those numbers are even close to accurate, the Nano is basically AMD's answer to the 980 in dx11, and almost the 980ti for dx12. Will be an interesting card if the numbers are even true given the inflated results of their FuryX pre-release benchmarks.

[othertomperson]

Looks like the 175W soft-limit is real. If those numbers are even close to accurate, the Nano is basically AMD's answer to the 980 in dx11, and almost the 980ti for dx12. Will be an interesting card if the numbers are even true given the inflated results of their FuryX pre-release benchmarks.

 

You think so? To me it's just making me question all the harder what the point of the 390X was. It seems like the Nano is going to perform the same but with about half the TDP. Not to mention that if they had called this the 390X they would have avoided all the criticism of rebadging the 290X to begin with.

 

AMD has such a problem with marketing it's phenomenal.

[Humbug]

You think so? To me it's just making me question all the harder what the point of the 390X was. It seems like the Nano is going to perform the same but with about half the TDP. Not to mention that if they had called this the 390X they would have avoided all the criticism of rebadging the 290X to begin with.

AMD has such a problem with marketing it's phenomenal.

They are having supply issues with HBM. They are selling out whatever limited of quantities of Fiji GPUs that they can deliver. It's causing them to lose market share.

Ironically at this stage it's a good think that they at least have the other GPUs to sell.

[othertomperson]

They are having supply issues with HBM. They are selling out whatever limited of quantities of Fiji GPUs that they can deliver. It's causing them to lose market share.

Ironically at this stage it's a good think that they at least have the other GPUs to sell.

 

Because no one saw that coming at all

 

Considering how little difference HBM even makes they should have just gone with GDDR5 until they could mass produce them. Owait that's what Nvidia did.

 

It's like... you look at what Intel do, and you look at what Nvidia do and then you look at what AMD are doing and think one of these isn't in profit. WHY MIGHT THAT BE.

[Briggsy]

Based on tech specs its just a more power efficient, ITX equivalent of the R9 290X/390.

Good to see AMD working towards power efficiency, which has been Nvidia's forte all the while.

Which has been Nvidia's focus since Kepler you mean. Nvidia cards before Kepler were electric grills.

[Jerakl]

Which has been Nvidia's focus since Kepler you mean. Nvidia cards before Kepler were electric grills.

*cough cough* Fermi *cough cough*

[Popingheads]

Because no one saw that coming at all

Considering how little difference HBM even makes they should have just gone with GDDR5 until they could mass produce them. Owait that's what Nvidia did.

It's like... you look at what Intel do, and you look at what Nvidia do and then you look at what AMD are doing and think one of these isn't in profit. WHY MIGHT THAT BE.

How do we know it makes "little difference"? Losing the huge amount of memory bandwidth going back to GDDR5 could cost the card a few more FPS worth of performance, in which case the card wouldn't be almost on par with the 980 Ti it would be behind it all the time, which is a big deal.

Plus the power draw of the card would increase by about 50 watts at least, in which case everyone would just be complaining about that.

Either way they aren't going to lose market share because of a flagship card not many people buy anyway, the mid range cards will sell more, and lucky for AMD their cards do have a better price to performance ratio than Nvidia cards below the $300 price point.

Also they get experience in using the new technology ahead of time. It isn't unusual for companies to introduce new tech on their low volume flagship products and then have it trickle down into other ones. What they learned from making Fiji will hopefully speed up development of the next generation mid/high end products that are likely going to be using HBM too.

[ZetZet]

They're "boasting" a massive 2 fps more than a 290X. I think we can probably trust their honesty this time.

Sick graph design, they should have made it 30-45 instead of 0-45, fail in marketing.

[carolkarine]

*cough cough* Fermi *cough cough*

You mean thermi

[CTR640]

After the false "overclocker's dream" about the Puny X, I don't believe AMD anymore.

They can lie all they want and in the end, they only got themself to blame if people picking nVidia over AMD.

Their words are like poop, you poop out of your ass and flush the shit down the toilet aaaaaaaand, it's gone.

[spartaman64]

Because no one saw that coming at all

 

Considering how little difference HBM even makes they should have just gone with GDDR5 until they could mass produce them. Owait that's what Nvidia did.

 

It's like... you look at what Intel do, and you look at what Nvidia do and then you look at what AMD are doing and think one of these isn't in profit. WHY MIGHT THAT BE.

if nvidia got hbm everyone would be applauding them for being an industry leader even if they have yield problems. and if you look at the gpu line up amd has nvidia beat in almost every match except the fury x and gtx 980 ti. the r9 390 beats the 970 r9 380 beats 960 but nooo amd cant compete with nvidia even though they offer better choices

[BuckGup]

If the price is right

[TheMissxu]

It's pretty much a downclocked Fury X on a really small PCB. I wonder what board partners can do with this... Custom PCB (with extra power) and custom coolers should make this into an overclockers dream/reality.

 

Oh and AMD has a chance to redeem themselves with pricing, which should make custom versions of this card amazing.

 

I'm guessing $499-599 for pricing, if it's anwhere between $450-550 for the custom cards I've been talking about, AMD has won...

[Ohlyver]

any ideas of the price point ?