Specs and Hardware
What’s 6 inches long and costs $650? The AMD Radeon R9 Nano!
The AMD Radeon Nano graphics card is unlike any product we have ever tested at PC Perspective. As I wrote and described to the best of my ability (without hardware in my hands) late last month, AMD is targeting a totally unique and different classification of hardware with this release. As a result, there is quite a bit of confusion, criticism, and concern about the Nano, and, to be upfront, not all of it is unwarranted.
After spending the past week with an R9 Nano here in the office, I am prepared to say this immediately: for users matching specific criteria, there is no other option that comes close to what AMD is putting on the table today. That specific demographic though is going to be pretty narrow, a fact that won’t necessarily hurt AMD simply due to the obvious production limitations of the Fiji and HBM architectures.
At $650, the R9 Nano comes with a flagship cost but it does so knowing full well that it will not compete in terms of raw performance against the likes of the GTX 980 Ti or AMD’s own Radeon R9 Fury X. However, much like Intel has done with the Ultrabook and ULV platforms, AMD is attempting to carve out a new market that is looking for dense, modest power GPUs in small form factors. Whether or not they have succeeded is what I am looking to determine today. Ride along with me as we journey on the roller coaster of a release that is the AMD Radeon R9 Nano.
AMD Radeon R9 Nano Specifications
I don’t want to spend a whole lot of time on this part of the review if only because we had another story posted just a couple of weeks ago where AMD released most of the technical information about the product. A quick recount of the information is provided here but for the full marketing detail about the card, check out my previous piece.
| R9 Nano | R9 Fury | R9 Fury X | GTX 980 Ti | TITAN X | GTX 980 | R9 290X | |
|---|---|---|---|---|---|---|---|
| GPU | Fiji XT | Fiji Pro | Fiji XT | GM200 | GM200 | GM204 | Hawaii XT |
| GPU Cores | 4096 | 3584 | 4096 | 2816 | 3072 | 2048 | 2816 |
| Rated Clock | up to 1000 MHz | 1000 MHz | 1050 MHz | 1000 MHz | 1000 MHz | 1126 MHz | 1000 MHz |
| Texture Units | 256 | 224 | 256 | 176 | 192 | 128 | 176 |
| ROP Units | 64 | 64 | 64 | 96 | 96 | 64 | 64 |
| Memory | 4GB | 4GB | 4GB | 6GB | 12GB | 4GB | 4GB |
| Memory Clock | 500 MHz | 500 MHz | 500 MHz | 7000 MHz | 7000 MHz | 7000 MHz | 5000 MHz |
| Memory Interface | 4096-bit (HBM) | 4096-bit (HBM) | 4096-bit (HBM) | 384-bit | 384-bit | 256-bit | 512-bit |
| Memory Bandwidth | 512 GB/s | 512 GB/s | 512 GB/s | 336 GB/s | 336 GB/s | 224 GB/s | 320 GB/s |
| TDP | 175 watts | 275 watts | 275 watts | 250 watts | 250 watts | 165 watts | 290 watts |
| Peak Compute | 8.19 TFLOPS | 7.20 TFLOPS | 8.60 TFLOPS | 5.63 TFLOPS | 6.14 TFLOPS | 4.61 TFLOPS | 5.63 TFLOPS |
| Transistor Count | 8.9B | 8.9B | 8.9B | 8.0B | 8.0B | 5.2B | 6.2B |
| Process Tech | 28nm | 28nm | 28nm | 28nm | 28nm | 28nm | 28nm |
| MSRP (current) | $649 | $549 | $649 | $649 | $999 | $499 | $329 |
AMD wasn’t fooling around, the Radeon R9 Nano graphics card does indeed include a full implementation of the Fiji GPU and HBM, including 4096 stream processors, 256 texture units and 64 ROPs. The GPU core clock is rated “up to” 1.0 GHz, nearly the same as the Fury X (1050 MHz), and the only difference that I can see in the specifications on paper is that the Nano is rated at 8.19 TFLOPS of theoretical compute performance while the Fury X is rated at 8.60 TFLOPS.
The memory system is also identical between the R9 Nano and the Fury X: 4096-bit wide high bandwidth memory bus, 4GB of capacity, 500 MHz memory clock rate and up to 512 GB/s of available memory bandwidth. Again, very impressive!
There is one very big difference that we have to point out though: The R9 Nano is rated with a 175 watt TDP, the Fury X at 275 watts. That is a difference in power consumption that we just haven’t seen in any other card to card variance when based on the same GPU.
That difference in power consumption is possible due to AMD “underclocking” the Fiji GPU: bringing the clock speed lower to meet a specific TDP target. And as we’ll see in an upcoming page of this story, the clock speed variance from game to game and even between resolutions is fairly substantial; I saw clocks as high as 1000 MHz and as low as 825 MHz.
It's important to note that even though AMD says the GPU is not thermally constrained, all aspects of the ASIC performance come into play during use. Efficiency is defined as the ability to run at a certain performance level and clock speed within a set TDP (which AMD has already defined at 175 watts). So if a particular game plays with a heavy GPU workload and attempts to draw more power than the 175 watts allowed, AMD's Fiji implementation will downclock until it arrives a voltage that hits ~175 watts draw. That will be different for all different kinds of software so its something we will pay particular attention to in our review.
For users that want to tweak things you will be able to adjust the power limit in the AMD Catalyst Control Center, thus decreasing efficiency, but pushing you closer to that 1.0 GHz frequency regardless of the GPU workload. This obviously takes AMD's new R9 Nano outside the range that it wanted to be able to claim for this product, and what it could guarantee in tight quarters that might be more thermally constrained, but gamers will have that flexibility if they wish, gaining as much as 10% in graphics performance based on AMD's quotes. Once you hit that 1000 MHz mark, however, you revert to the more standard overclocking models that we have already seen which aren't spectacular with Fiji GPUs.
The AMD Radeon R9 Nano – A Sexy Beast
From a design standpoint there is very little that was unknown about the R9 Nano hardware. This is a 6-in PCB design with a vapor chamber cooler that helps keep that sizeable Fiji GPU and HBM memory in the 75-85C range during gaming. Once again, our previous story has all the details on AMD’s position of the R9 Nano and its size, claiming that it can be used in cases that would otherwise not have access to cards in this performance class. Still, this is dead-sexy hardware so let’s take a look around our sample.
It’s hard to really even quantify with a single image, but the R9 Nano is incredibly tiny, measuring just 6-inches long. The fit and finish on the design is top notch, nearly matching the build quality of the Fury X in terms of the finish and rubber-coated plastic body. To be clear – the Nano is entirely plastic, none of that nickel plated gunmetal color to be found here. The single fan in the middle is responsible for moving around the necessary air to keep the Fiji GPU cool – this card is quiet but far from silent.
You won’t find any fancy LED lights on the card, which is a bit of a letdown considering how apt AMD is pushing this card for custom mods and windowed chassis. The Nano does support CrossFire but it’s using XDMA so no CrossFire connectors are needed.
There is no back plate on the R9 Nano which I am sure will aggravate some users immediately. I assume that this is not a cost saving measure and instead is to ensure that the card will fit onto some motherboards with very limited spacing between the primary PCIe slot and the CPU socket.
For power delivery the R9 Nano only requires a single 8-pin PCIe power connection, facing out the back of the card. This should help the Nano fit in more tightly-spaced cases than with the power connection up top, but I am sure there are cases where the opposite is true as well.
Display connectivity on the R9 Nano consists of a set of three DisplayPort connections and a single HDMI port. There are no DVI connections – kind of a bummer. Also, the HDMI port here is still HDMI 1.4a, not HDMI 2.0, which is one of the biggest drawbacks to Fiji over NVIDIA’s Maxwell architecture. This is to an even greater degree in the case of the R9 Nano as it fits perfectly into the stereotype of an HTPC card, with previously unseen graphics performance, though it won’t be able to output to 4K 60 Hz TVs without an active adapter. An adapter, mind you, that I have yet to touch or see or price or validate.









You amd shill! im kidding!
You amd shill! im kidding! Ryan this was a great review i see a nano in my mitx build thank you.
With so many ITX cases
With so many ITX cases designed around the ability to hold full sized cards, this is like a niche within a niche, for true SFF cases.
This so much, generally if
This so much, generally if your case has 2 pci slots then you can fit a decent sized card in. Take for example the Silverstone sugo 13 at 11.8L can fit a reference 980ti inside, how many smaller cases cases can even hold the nano?
LIAN LI PC-TU200B Black
LIAN LI PC-TU200B Black Aluminum Mini-ITX Tower Computer Case
OMG this is so
OMG this is so disappoint.
I’ll just get the asus gtx970 mini for $335, and OVERCLOCK IT, it will be just about the exact same thing using much less power.
Plus we’ll have all the in game recording and instant replay and all those other driver settings goodies from nvidia.
amd just can’t win, I can’t do it i’m so frustrated
Seems like a bit of a waste
Seems like a bit of a waste to using a full chip, would it not have been a better idea to use a cut down chip to achieved the lower heat and power targets. Not only would a cut down chip be cheaper in theory but the yeilds would probably be better. Might cut into chip allocation of the R9 fury though.
it is a waste of a full chip
it is a waste of a full chip from a performance standpoint, but charging the same as a card with an elaborate cooling solution is good for the bottom line. At least AMD trying to make a profit this time rather than giving away stuff for free(sync).
+1
+1
Generally, it’s more power
Generally, it’s more power efficient to run more shaders at a lower clockspeed (notice how the drop in performance for the Nano vs the Fury X is smaller than the drop in power consumption, and also consider discrete laptop GPUs that run at 30% lower clock speeds but half the power consumption of their desktop counterparts).
I think the yield issues are more to do with HBM and the interposer than the GPU itself, which, while large, is built on a very mature process.
I think it’s a power/ASIC
I think it’s a power/ASIC issue, so they for instance they weren’t getting a bunch of great chips that clock high and stable, but all the internals worked, so they decided to not cut shaders and rop’s down but clock it down and cover their backsides on their recent extreme power usage failures.
Yep.
I’d like to see some clock
I’d like to see some clock speed numbers at different power targets. +10% +20% etc. would be interesting know how much extra it would take to average around 950mhz and so forth. I’m guessing the sweet spot isn’t +50%
I’m sure the sweet spot is
I'm sure the sweet spot is 0%, as that is where AMD put it. Remember that as voltage increases, power increases by a factor squared. I can try to run a set of test on that today if I get time though!
would be nice to have manual
would be nice to have manual control of the voltage too 🙂
I do suspect that you’ll be able to get a decent percentage of the extra performance without ramping things up to 11.
Just a guess, but I think there will be little trade off to setting the powerlimit somewhere between 10% and 20%.
Keeping consumption <= 200w should not require a huge bump in voltage. just as an aside, do you know if AMD is going to let partners produce Nanos (Nani?) with custom coolers? Or is it reference only?
rumors are theirs a chance
rumors are theirs a chance the partners would be able to produce this.
Gotta love the name,
Gotta love the name, Nanosaurus. Kinda badass adorable.
Other than that, I like that you guys actually tested the card in a fitting case.
Good god, the first and
Good god, the first and second graph. The accountant in me is crying, way to much noise to actually see what is going on.
I know…couldn’t find a
I know…couldn't find a better way to display it, hence the average MHz graph following it.
But in reality, the mess of a graph demonstrates how "all over" the frequency behaves during gaming.
Agreed and thank you for the
Agreed and thank you for the review.
Everyone talks form factor.
Everyone talks form factor. Who the F cares. Most smaller builds can fit a bigger GPU. This card makes no sense.
Better to get a bigger GPU for less money and less noise.
Most but not all cases. We
Most but not all cases. We had the CM Elite 110, the Lian Lin Q33 has even tighter requirements. Just because many or most cases can fit larger cards doesn't mean we can't look to other options.
Ryan:
Yet, you reviewed in a
Ryan:
Yet, you reviewed in a case that could fit a Fury X… so why?
Not long ago you questioned the Nano’s existence on a podcast and now all of a sudden you think it has a place at $650, when you still admit that most cases don’t require such a small GPU.
You say because most cases can fit a larger card it doesn’t mean you can’t look at other options, but again why? Is it power savings? No. Is it cost? No. So what is it?
Think Ryan was very clear in
Think Ryan was very clear in the review, if your case accommodates a bigger card then the Nano is not the right choice.
Within the niche it fits in (pun intended) its good choice (extreme SFF and custom setups for instance).
Wouldn’t have sucked for them
Wouldn’t have sucked for them to lower the price to $600. It would’ve been a major difference in the psychological and marketing departments, and brought in some goodwill rather than hostile incredulity.
Maybe. But again, I don’t
Maybe. But again, I don't think they will have issues selling through all they can make.
no kidding resale fury’s are
no kidding resale fury’s are 200 over retail atm
My view on this card is very
My view on this card is very similar to its slightly bigger brother.
I like what AMD are trying to achieve, I will wait for HBM v2 as jumping onto new memory tech first is rarely a good idea. I watercool my PCs in a custom loop so I want a full cover waterblock versions (but without the silly AIO system of its bigger brother).
With this performance on 28nm
With this performance on 28nm and HBM1 and at just 175W, the next-gen card should be spectacular.
It would be interesting to
It would be interesting to see an Ashes of the Singularity comparison between the Nano, Fury and FuryX. Please Make it so Ryan.
It takes some guts to go with
It takes some guts to go with a gold recommendation just on the performance of this card when AMD chose not to send this card to other reviewers. I applaud your courage. I expect you’ll incur suspicion for it, but it isn’t deserved.
The AMD Radeon R9 Nano will
The AMD Radeon R9 Nano will have to be tested again once the gaming software/engines and graphics APIs like Vulkan/DX12 begin to become more utilized. There also needs to be a benchmark specifically designed to test the hardware’s asynchronous compute ability. The AMD GCN ACE units will come into their own with the newer graphics APIs, so not only is it just a matter of retesting AMD’s driver updates, but testing again with the newer graphics API’s and games that will take advantage of asynchronous compute. The entire gaming ecosystem software stack from gaming engines to to graphics APIs are turning over completely with the newest going completely into using the GPU hardware’s asynchronous compute ability, especially for VR gaming. This year testing will have to be redone more often as the new graphics APIs start to come online, especially with the Steam OS based systems and Vulkan adding coming into the gaming market.
Edit: adding coming
To :
Edit: adding coming
To : coming
Its definitely going to be a
Its definitely going to be a weird transition going from DX11 to DX12 benchmarks. At first this worried me, but then I realized that the first year or so of DX12 games will likely give the option to switch between dx11 and dx12 (and possibly vulkan too), which should really make benchmarks look interesting as the old and new API’s are compared. Even though we probably won’t see any more relevant GPU releases this year besides the dual Fiji, reviewers are going to be all over every dx12 release anyway, doing benchmarks and comparisons. Luckily there will only be one or two DX12 titles (Cars and ARK) with any gameworks, so the comparisons should be good. I think I even heard a rumor about DX12 project cars getting insane performance on both sides with DX12.
Interesting times ahead.
I should mention that I meant
I should mention that I meant DX12 “Patches” for Cars and ARK, not launch API support.
I’m looking forward to seeing
I’m looking forward to seeing what customized SFF cases that people will cram these cards into. SFF cases are one thing and nice to see, but what I really want to see is cases that you wouldn’t expect to see a PC in it, much less a top-tier GPU in it too. A gutted and then fully modded game console that has a Nano in it? or how about a case in the same style that Sebastien reviewed last year in the link below?
https://pcper.com/reviews/Cases-and-Cooling/Perfect-Home-Theater-Ultra-Low-Profile-Aluminum-HTPC-Case-Review
(disclaimer – the nano is probably too thick to fit in this particular case, but it’s this style of case i think would be cool to see a full PC with Nano in it)
Great review!
Coil whine, high price for
Coil whine, high price for what it is, no HDMI 2.0, no overclocking headroom, no fun. But it gets a Gold award anyway. Okay then.
You seem to have missed the
You seem to have missed the part where this is a SFF product
You could make most of those and similar complaints about a gaming laptop (vs a gaming desktop), but that would be utterly missing the point, as you have done here.
So its form factor means we
So its form factor means we should disregard all of these failings?
Sorry, don’t agree.
Admittedly, HDMI 1.4 is a
Admittedly, HDMI 1.4 is a problem, but only a tiny one. There are very few TVs worth using as a monitor (OLED and Panasonic plasmas), and none of those are 4k under $5000. I’m more bothered by the lack of VGA/DVI-I then the lack of HDMI 2.0 here. Coil whine is the only real issue with this card, and it’s not enough of an issue on its own to withhold an award.
Overclocking is only ever relevant on the largest form factors, which this is not. If you really want to overclock it, you still can, if you can tolerate the noise. Performance per dollar is also not comparable across different form factors- if it was, there would be no such thing as a laptop or smartphone.
Salty much.
Salty much.
The coil whine is a concern
The coil whine is a concern but there are ways to mitigate it, if retail samples still have that problem.
The only real flaw is having hdmi 1.4 instead of 2.0. I don’t know why AMD wouldn’t support hdmi 2.0 for this card. If I were hooking this up to a 4k TV I would prefer to use hdmi. However, I do prefer DP for hooking up to a 4k monitor- and that is what I will actually use (maybe, I’m on the fence about whether this is the time to upgrade my SFF gaming computer, but this is the first time I have been convinced that it is feasible without getting terrible performance).
Coil whine-
Not sure if
Coil whine-
Not sure if everyone cares about that,lot of gtx 970s had it yet the buyers kept it.
No hdmi 2.0- Im sure you can get a converter
-Not true,Check out guru3d’s review they overclocked it so much with considerable gains.
-It gets the gold award for being an execeptional product unmatched it’s category.Esecially given the fact that ‘Real ITX’ cases requires an ITX card and cant have full sized ones.
I’ll give you the price not being appealing however they are only going to make so much of it that they’ll sell it all in the niche.
It’s not for everyone
I agree 100% this is not a
I agree 100% this is not a Gold award product.
Very disappointed in this site. I give Ryan credit for being honest and writing a good review, but how can a review have so many negatives and yet still come out with Gold award.
Yeah its great that they packed so much power into a tiny device, who cares.
Exactly, and remember the
Exactly, and remember the Fury X got a Silver.
PCPer’s reviews make no sense just like the Nano.
“Yeah its great that they
“Yeah its great that they packed so much power into a tiny device, who cares.”
A few people care, and for them, this is a gold-worthy product. For everyone else, it’s mostly irrelivant, but Ryan wrote earlier that if you can fit a larger card, don’t bother with this one.
A product doesn’t need to be great for everyone in order to be a good product, it only needs to be good for it’s target audience. This card’s audience may be small, but it does exist, and for those users, it’s nearly perfect.
If Ryan didn’t write a nice
If Ryan didn’t write a nice ‘fair’ review, he would of been AMD blacklisted…..
What a have!….This nonsense needs to stop! gift card for a favourable review..LMAO!
Nice&informative review
Nice&informative review Ryan,Thank you!
Finally, Thanks AMD.)
Just
Finally, Thanks AMD.)
Just bought one, and it coming tommorow.) So in this weekend, I will try to crossfire my Fury X with the Nano, just to see, if is worth to buy an X more.)
Thanks PCPER, “Ryan” for a good review.!
Nice day to all.)
I’m not sure this makes much
I’m not sure this makes much sense, even for a make-it-as-small-as-possible ITX case.
The Dan A4-SFX (http://www.dan-cases.com/) is the smallest self-contained (i.e. no external power bricks to lug around) ITX case I’ve seen at 7.25L, and that’s with a full length GPU. But switching to a shorter GPU would save barely any space, due to the combined length of the motherboard and SFX PSU. And any case using external bricks and an internal DC-DC PSU would end up having to use two power bricks at a minimum, and those things aren’t tiny at high wattages.
I guess if you have a super-custom one-of-a-kind case that only takes short GPUs, then this is the only game in town. But you could buy a larger, better performing card AND a new case if you just wanted something small.
Except you can do this with a
Except you can do this with a Fury Nano.
https://s-media-cache-ak0.pinimg.com/originals/7c/ec/40/7cec406cfa52d80f465a28ac4a672459.gif
The GTX 980 would have helped
The GTX 980 would have helped in this review.
Another aspect would have been case thermal…
Its a nice little card. But at $650 its appeal is so, so limited, I can only guess to match its supply.
And its hard to find even micro atx cases that cant fit a GTX 980…. On paper that deliver better thermal, about equal performance, lower power? And its $170 cheaper.
The 980 wasn’t included
The 980 wasn’t included because it is irrelevant since it is not a mini-itx card.
He included information about the thermals, go to the page marked as:
Noise Testing, SFF System Build, Temperatures
and at the bottom Ryans states that on the open test bench the card was at about 75c and in the Cooler Master Elite 110 chassis the card got to about 83. No graphs included.
The supply of fiji chips is only part of it. I am inclined to believe Ryan’s assessment that much like Intel charging a premium for their chips that go in the ultra-books(good performance with low power usage), AMD is doing the same thing here, charging a premium because you get very good performance in such a small form factor.
Again, comparing it to the 980 is irrelevant. This is about the SFF that it will fit in and how much performance it has in that SFF. It’s not about putting it in a case that any other card can fit in.
Cases like the one that Ryan used in this article that will not fit a regular sized gfx card. Things like the idea of gutting a console and putting a PC inside of it including a high-end GPU. Imagine turning on an xbox one and it could actually run games at 4k and not just 720 or 900p. Keep in mind, and Ryan talks about it too, that the trend lately is to go smaller and smaller.
TL;DR – People, you are paying a premium to get very good performance in a very small form factor so the card can be in places that no other card could be in.
The size difference is
The size difference is actually very minimal, its just 4″.
This is why the large majority of SSF case can accept 10″ cards.
Also the GTX 980 is not dumping large amount of heat inside the case, the nano is. And this will affect your CPU cooling (more fan noise)
Most mini-itx cases that fit
Most mini-itx cases that fit full size cards are joke. They are only a few inches shorter than micro-atx cases that support SLI
To me all those cases are the ones that don’t really make any sense. Why pay premium for only a couple inches and less upgradeablity?
If you are going ITX, you have to go all the way small with small form factor psu.
Ryan, you didn’t mention how
Ryan, you didn’t mention how much hotter the card got after you gave it +50% power target.
Ah, true! It didn’t get much
Ah, true! It didn't get much hotter (75C up to 83-84C) but it did get louder! I'll update tomorrow!
84C on the test bench? That
84C on the test bench? That should hit close to 90C in the CM case, or the fan would bleed your ear off.