Tuesday, August 25th 2015

AMD Radeon R9 Nano Core Configuration Detailed

AMD's upcoming mini-ITX friendly graphics card, the Radeon R9 Nano, which boasts of a typical board power of just 175W, is not a heavily stripped-down R9 Fury X, as was expected. The card will feature the full complement of GCN compute units physically present on the "Fiji" silicon, and in terms of specifications, is better loaded than even the R9 Fury. Specifications sheet of the R9 Nano leaked to the web, revealing that the card will feature all 4,096 stream processors physically present on the chip, along with 256 TMUs, and 64 ROPs. It will feature 4 GB of memory across the chip's 4096-bit HBM interface.

In terms of clock speeds, the R9 Nano isn't too far behind the R9 Fury X on paper - its core is clocked up to 1000 MHz, with its memory ticking at 500 MHz (512 GB/s). So how does it get down to 175W typical board power, from the 275W of the R9 Fury X? It's theorized that AMD could be using an aggressive power/temperature based clock-speed throttle. The resulting performance is 5-10% higher than the Radeon R9 290X, while never breaching a power target. Korean tech blog DGLee posted pictures of an R9 Nano taken apart. Its PCB is smaller than even that of the R9 Fury X, and makes do with a slimmer 4+2 phase VRM, than the 6+2 phase VRM found on the R9 Fury X.
Sources: VideoCardz, IYD.kr
Add your own comment

101 Comments on AMD Radeon R9 Nano Core Configuration Detailed

#26
uuuaaaaaa
commando55555what would be the point doesn't oc for shit any way.
I guess that would mean lower operating voltage at the same clock speed, lower power consumption and lower temperatures, so it could be close to Fury X's performance at such form factor. Also it may have some room to oc a little bit...
Posted on Reply
#27
commando55555
Sony Xperia SWhen will you understand that the card itself will be damn fast even without overclocking and with overlock you will achieve nothing but nonsense?
When will you understand that when buying a card oc is consideration. The fact it doesn't is nonsense the 980 ti is pretty dam fast but that 30% oc is pretty nice.
Posted on Reply
#28
Sony Xperia S
commando55555When will you understand that when buying a card oc is consideration.
No, it's just a wishful thinking from your side and no one guarantees you anything. Actually the opposite - you can void your warranty by changing the factory defauults. :D
commando55555The fact it doesn't is nonsense the 980 ti is pretty dam fast but that 30% oc is pretty nice.
Yeah, sure those magic 30%. I guess it woun't throttle even a tiny bit. :D
Posted on Reply
#29
Lionheart
Assimilator"Up to 1000MHz" = you'll see 1000MHz for a couple of milliseconds at best. Most of the time it will run at a far lower core clock.

Overclocking on this card is going to be nonexistent because it will probably catch fire, but that's okay because Fury/Fiji overclocks like s**t anyway, so why would you even bother.
Jesus christ every damn AMD article you're always negative, give it a rest.
Posted on Reply
#30
commando55555
Sony Xperia SNo, it's just a wishful thinking from your side and no one guarantees you anything. Actually the opposite - you can void your warranty by changing the factory defauults. :D


Yeah, sure those magic 30%. I guess it woun't throttle even a tiny bit. :D
I know not all chips will oc the same. But for example 980ti are good oc but it's going to very trial and error. The bottom line is some people like to oc for these people this card is a turd.
Posted on Reply
#31
Assimilator
LionheartJesus christ every damn AMD article you're always negative, give it a rest.
I'll stop being negative towards AMD, when AMD marketing stops lying to their customers and delivers the products they promise.
Posted on Reply
#32
RejZoR
It's probably going to be 800Mhz and 1000MHz max boost until it reaches the thermal limit. I can't see it running at 1GHz with this tiny cooler and same core configuration as Fury X. It just makes no sense even compared to vanilla Fury with its massive coolers.
Posted on Reply
#34
cokker
AssimilatorI'll stop being negative towards AMD, when AMD marketing stops lying to their customers and delivers the products they promise.
"GTX 970"

Posted on Reply
#35
5DVX0130
What's that I hear!?

Oh it's just the hype train making its rounds.


I do hope it’s a good card, for all our sakes, but till the benchmarks hit it’s just a pretty face.

But I do wonder what the temps/noise will be like. Seeing as the non X Fury got to 70C and it has a massive cooler.
Posted on Reply
#36
lilhasselhoffer
So....

What AMD has delivered is a very small card, that should reasonably approximate a much larger card sans overclocking. Honestly though, did anyone expect a card this small to somehow come in with insane overclocking? If they did, they were asking for something pretty crazy.

The card is interesting as a thought experiment. Move it to14nm, and HBM2 at 8 GB, and you'll have something truly amazing for the HTPC gaming crowd. As it stands now, meh. It'll be an expensive niche card that is great for someone needing a small form factor. Regular users won't allow the price premium, so it won't be something for the mainstream. I can see this card being amazing at 390 pricing, but it's not likely to be that reasonable.



As far as the AMD/NVIDIA debate, both companies are full of crap. AMD did do a rather hatchet job selling its current generation of GPU, but if you forgot the 970 debacle you're pure fanboy. Take everything with a grain of salt until it has been bench marked.
Posted on Reply
#37
SonicZap
Hawaii suddenly becomes very efficient with lower clock speeds - some users have reported their R9 290X power consumption being lowered by half with an underclock to 750 MHz or so. It's likely the same case here, dropping clock speeds and voltage makes Fiji a lot more efficient than the implementation in Fury X.

I'm not in the market for buying Nano, but I'm interested to see how efficient it is. It might even be more efficient than Maxwell considering that Fury (non-X) is already very close to Maxwell in power efficiency.
Posted on Reply
#38
the54thvoid
Super Intoxicated Moderator
The only niche competition it has, the itx 970's came out in late 2014. It's easily the fastest itx card.
As for clocks, that's how the professional cards hit power limits, low clocks.
Given the stock Fiji chip being used, I guess some firmware is involved, unless there is PCB hardware for power limiting. If its hardware, no over clocking but if its firmware, it'll be flash happy. But risky....
Posted on Reply
#39
64K
Price and performance need to be seen but maybe there won't be scarcity. Fury have started showing up where I buy my hardware but still no Fury X unless you want to be gouged at Amazon. I don't know why anyone would pay $1,000 for a Fury X.

If the Nano performs close to a 980 then efficiency is definitely good.
Posted on Reply
#40
FordGT90Concept
"I go fast!1!11!1!"
btarunrSo how does it get down to 175W typical board power, from the 275W of the R9 Fury X? It's theorized that AMD could be using an aggressive power/temperature based clock-speed throttle.
That's concerning. :(
Posted on Reply
#41
Aquinus
Resident Wat-man
There is an "11" footnote marker on the "Compute Units" section of the specs. I would really like to know that footnote actually says. Maybe the CUs are gimped in some way to keep the GPU within TDP and VRM limits? How interesting would it be if CUs disabled themselves if power draw or heat become too big of a problem? CU level power-gating in addition to clock scaling. It could be a blueprint for power saving features going forward on AMD GPUs.
Posted on Reply
#42
Sony Xperia S
SonicZapHawaii suddenly becomes very efficient with lower clock speeds - some users have reported their R9 290X power consumption being lowered by half with an underclock to 750 MHz or so. It's likely the same case here, dropping clock speeds and voltage makes Fiji a lot more efficient than the implementation in Fury X.
This means that Fiji and Hawaii offer the best or optimal characteristics performance to power ratio in a different part of the curve compared to the ones they are being sold with.

It means that overclocking for more performance scales rather poorly.
Posted on Reply
#43
newtekie1
Semi-Retired Folder
This thing is going to run so hot, and throttle constantly. Though I'm guessing it will remain at a high enough speed just long enough to get through the benchmarks used in most reviews, so the reviews show way higher performance than you actually get, which is typical for AMD.
Posted on Reply
#44
uuuaaaaaa
newtekie1This thing is going to run so hot, and throttle constantly. Though I'm guessing it will remain at a high enough speed just long enough to get through the benchmarks used in most reviews, so the reviews show way higher performance than you actually get, which is typical for AMD.
Afaik this only happened with the reference models based on Hawaii...
Posted on Reply
#45
newtekie1
Semi-Retired Folder
uuuaaaaaaAfaik this only happened with the reference models based on Hawaii...
I haven't tried a whole lot of the aftermarket cards, but my Sapphire Tri-X's throttled when I had them in crossfire(well the top card throttled).
Posted on Reply
#46
uuuaaaaaa
newtekie1I haven't tried a whole lot of the aftermarket cards, but my Sapphire Tri-X's throttled when I had them in crossfire(well the top card throttled).
Maybe it needed more room to breathe. Which cards did you have?
Posted on Reply
#47
GhostRyder
AssimilatorI'll stop being negative towards AMD, when AMD marketing stops lying to their customers and delivers the products they promise.
This coming from the guy who owns a GTX 970...
cokker"GTX 970"

Yes this sums up my thoughts nicely^
AquinusThere is an "11" footnote marker on the "Compute Units" section of the specs. I would really like to know that footnote actually says. Maybe the CUs are gimped in some way to keep the GPU within TDP and VRM limits? How interesting would it be if CUs disabled themselves if power draw or heat become too big of a problem? CU level power-gating in addition to clock scaling. It could be a blueprint for power saving features going forward on AMD GPUs.
It will be interesting to see how this is handled because there are so many questions left open because of this announcement. I am curious how well this cooler handles things honestly even with that TDP and if it really is going to throttle. Its a great mystery that I only think will be resolved once we see them all around :).
Posted on Reply
#48
john_
AssimilatorI'll stop being negative towards AMD, when AMD marketing stops lying to their customers and delivers the products they promise.
So you support Nvidia because it says the truth?
I haven't check if TPU have a subforum with funny stories. Your can quote me and post your answer there.
Posted on Reply
#49
64K
GhostRyderThis coming from the guy who owns a GTX 970...
I had a GTX 970 before and it was a very nice card. I was an early adopter and got the card before the truth came out. I did think it strange that I got it for $360 when the 980 was $550 with only a little better performance. The previous generation was $400 for the 670 and $500 for the 680. I expected similar pricing.

But yeah, Nvidia definitely told some lies and AMD lies sometimes and Publishers lie sometimes. It's a bit of a shady hobby we're in.
Posted on Reply
#50
LightningJR
Honesty if this card has a competitive price/performance ratio it'll be a really great card for consumers. I'll never understand AMDs reasoning behind using a full Fury X core in it, I swear AMD is trying to kill themselves... I really can't wait for the review of this card, if anything AMD is an extremely interesting company. :P
Posted on Reply
Add your own comment
Nov 21st, 2024 10:15 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts