Radeon R9 Nano Small Form Factor Overclocking Review @ [H}

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,635
Radeon R9 Nano Small Form Factor Overclocking Review - We take the AMD Radeon R9 Nano in our Corsair 250D small form factor case and find out if the Nano has the potential to overclock and match the AMD Radeon R9 Fury X card's performance. We will also find out what it takes to deliver a consistent, non throttled, clock speed while gaming.
 
I have a serious doubt, but first of all thanks hardOCP for make this review possible, Now the doubt:

Can be really those Nano results cross compared between reviews? we are watching an increased performance but it was due the overclock? or due power tune +50 plus fans at 100%?

If I understand correctly, that only fact changes drastically the way clocks behave, then affect the performance, in my opinion and don't take it wrong. this review was wrongly made, if you really wanted to compare overclocking potential you had to run first the nano at stock clock but with powertune +50 and fans at 100%, then compare it with the overclock.

in my opinion that performance increase are because clocks are in average higher and not due specific overclock, the reviewer just copied and pasted the results from past review and compared straight to the nano with power tune +50 and fans at 100% which in my opinion is wrong.
 
I'm really impressed by this card for it's size. This sentence sums up my thoughts on OCing sometimes: "In our opinion the performance differences aren't worth the potential shortened lifespan or breaking of the video card to push it to those extremes to achieve that extra 10% performance, and remember what we said about the sound/noise profile."
On my desktop I watercool to get around most of those negatives.

My SFF build is going to wait for the next set of cards. If I was buying now I'd most likely buy the nano. It's price/performance curve is steep though.
 
Last edited:
macristina said:
I have a serious doubt, but first of all thanks hardOCP for make this review possible, Now the doubt:

Can be really those Nano results cross compared between reviews? we are watching an increased performance but it was due the overclock? or due power tune +50 plus fans at 100%?

If I understand correctly, that only fact changes drastically the way clocks behave, then affect the performance, in my opinion and don't take it wrong. this review was wrongly made, if you really wanted to compare overclocking potential you had to run first the nano at stock clock but with powertune +50 and fans at 100%, then compare it with the overclock.

in my opinion that performance increase are because clocks are in average higher and not due specific overclock, the reviewer just copied and pasted the results from past review and compared straight to the nano with power tune +50 and fans at 100% which in my opinion is wrong.
Click to expand...

The overclocking review is a continuation of the Radeon R9 Nano Small Form Factor Competition Review - http://www.hardocp.com/article/2015/11/24/radeon_r9_nano_small_form_factor_competition_review the data was collected at the same time of the review, the hardware is the same, the drivers are the same, the games are the same, yes it is a valid comparison.
 
Brent_Justice said:
The overclocking review is a continuation of the Radeon R9 Nano Small Form Factor Competition Review - http://www.hardocp.com/article/2015/11/24/radeon_r9_nano_small_form_factor_competition_review the data was collected at the same time of the review, the hardware is the same, the drivers are the same, the games are the same, yes it is a valid comparison.
Click to expand...

I understand, but I ask again, are we watching that increased performance but it was due the overclock? or due power tune +50 plus fans at 100%?. that could make a whole difference between even try to overclock or not if its only a matter of move the power tune slider, because in the past review you tested the card with stock fan curve and no modification to power tune.
 
macristina said:
I understand, but I ask again, are we watching that increased performance but it was due the overclock? or due power tune +50 plus fans at 100%?. that could make a whole difference between even try to overclock or not if its only a matter of move the power tune slider, because in the past review you tested the card with stock fan curve and no modification to power tune.
Click to expand...

The review is comparing stock Nano (no modifications) to our overclock we achieved, which is 1050MHz/525MHz. How we achieved that overclock is detailed quite well on page 3. Page 4 shows you the overclocked clock speed over time and what it really looks like while overclocked.
 
macristina said:
I understand, but I ask again, are we watching that increased performance but it was due the overclock? or due power tune +50 plus fans at 100%?. that could make a whole difference between even try to overclock or not if its only a matter of move the power tune slider, because in the past review you tested the card with stock fan curve and no modification to power tune.
Click to expand...
PowerTune is changing the factory Pstates of the card through the drivers. That would be considered an overclocking feature and wouldn't be representative of how the card performs out of the box.
 
Last edited:
Armenius said:
PowerTune is changing the factory Pstates of the card through the drivers. That would be considered an overclocking feature and wouldn't be representative of how the card performs out of the box.
Click to expand...

Neither would manually overclocking the gpu clock speed and memory as we have done.

I don't see the problem. This is how you overclock AMD GPUs, and is how we overclock every single video card since forever.

You cannot realize the potential of your overclock unless you increase PowerTune. It is exactly the same as increasing the PowerTarget on NVIDIA GPUs. If you don't do this, you are ultimately power capped and changing the clock speed will not result in a performance increase, or will throttle the GPU.

This is the correct way to overclock to achieve the best performance out of a video card.
 
Brent_Justice said:
I don't see the problem.
Click to expand...

there is no problem, you guys tested exactly how it should be done (and how you always do it): out-of-the-box vs overclocked. what macristina is suggesting would be of no value. there's no need for some quasi in-between overclocked case.

thanks for the review.
 
Too bad the fans have to be at 100% =( Even 75ish would probably be more bearable, thanks for the review guys!
 
Yikes. That is about the best air cooling scenario you could get without doing an open air bench.

The card needed the fan cranked to 100% and power tune up 50% just to get it to overclock to Fury X levels and its the same chip as the Fury X. Sounds to me like that's confirmation the Fury X really did need the AIO watercooler to keep it cool and give enthusiasts a little headroom to play with.

Fury Nano is still a neat little card for a very specific niche.
 
macristina said:
I understand, but I ask again, are we watching that increased performance but it was due the overclock? or due power tune +50 plus fans at 100%?. that could make a whole difference between even try to overclock or not if its only a matter of move the power tune slider, because in the past review you tested the card with stock fan curve and no modification to power tune.
Click to expand...

The Nano clocks at up to 1ghz, It is possible that increasing the powertune would make the nano get closer to the 1ghz mark on average but not surpass it Just like it did with a bigger case with better airflow. So the increased performance comes from overclocking at 1050 mhz.
 
Any chance you could put an aftermarket AOI on it and give it another try. Interested to know if the slight improvements and binning of the chips would actually give you superior performance to the Fury X or not when both are liquid cooled.
 
Wasn't expecting much success with overclocking, and the way [H] got around to do the overclock seems like too many hurdles. And running the fan 100% on a GPU that would more than likely sit beside my TV as a HTPC if I were to use it in a build? No, thank you. Still a neat, little card. It's crazy just how niche this little guy is.

Here's my crazy hypothesis:

When MS and Sony went to AMD to create the GPUs for their next generation consoles, they wanted something very powerful but in a very small package. AMD went ahead and started development on the little guy, however, at some point both MS and Sony decided to make their consoles more price friendly and opted for a lower cost GPU.

At this point, AMD has devoted too much time into this tech and repurposed it for a niche DIY release.

That all sounds crazy and pulled straight out of my ass: but imo, it still makes more sense than AMD deciding to devote so many resources on such a niche product.
 
Taldren said:
Any chance you could put an aftermarket AOI on it and give it another try. Interested to know if the slight improvements and binning of the chips would actually give you superior performance to the Fury X or not when both are liquid cooled.
Click to expand...

I'm a bit curious about how Nano would perform under water vs. a fury X as well, but I think that might be a bit much to ask. If I were buying, and I knew I was gonna put the card in a custom loop, my inclination would be to buy a nano. But maybe that's wrong? Does fury X have more/better power regulation? The nano full cover blocks are so nice looking though....
 
Methos said:
I'm a bit curious about how Nano would perform under water vs. a fury X as well, but I think that might be a bit much to ask. If I were buying, and I knew I was gonna put the card in a custom loop, my inclination would be to buy a nano. But maybe that's wrong? Does fury X have more/better power regulation? The nano full cover blocks are so nice looking though....
Click to expand...

Probably very similar to identical to a FuryX which is already water cooled. Plus the FuryX has an additional 8 pin pcie power plug allowing for more power. Looks like the best way to get more performance out of the nano is just turn up PowerTune and maybe a more aggressive fan profile for cooling and skip the voltage and OClock percent. Card is not great for aggressive Overclocking.

For off the shelf components best suited for the Nano, I say look at HardOCP first review, that is about as good as it gets. Of course there is always room for improvement with custom configurations.
 
Perhaps you could under volt it and run the fan a little slower while still maintaining clocks. It seems they have the Fury cards pretty close to the limit to start but imagine you could tweak a little out of it.

For those curious of the AIO on it. When I put an AIO in my 980 IIRC I was able to OC 3% faster. Apples to apples the power usage drops around 1% for 20C. Neglible gains. I saw similar when I took my 5960x sub zero. I would imagine the same here. It's more for a reduction in noise than anything.
 
I think it testifies to [H]'s thoroughness, professionalism, and responsiveness to its readership that they took a card that AMD stiffed them on and did such an in-depth series of reviews on it. Only AMD die-hards and ultra-pure SFF fanatics would even buy this thing, and yet they have spent all this time turning over every rock to give the world the whole picture.

It's all treading water until Pascal/Arctic Islands at this point (unless nVidia gives us some kind of Maxwell-based surprise for Christmas), but [H] puts in the time. Kudos.
 
Commander Suzdal said:
I think it testifies to [H]'s thoroughness, professionalism, and responsiveness to its readership that they took a card that AMD stiffed them on and did such an in-depth series of reviews on it. Only AMD die-hards and ultra-pure SFF fanatics would even buy this thing, and yet they have spent all this time turning over every rock to give the world the whole picture.

It's all treading water until Pascal/Arctic Islands at this point (unless nVidia gives us some kind of Maxwell-based surprise for Christmas), but [H] puts in the time. Kudos.
Click to expand...

That is why I have been such a long time reader. They go in depth in ways no,other reviewer will. I just they would bring back Hard Console!
 
..thnx for this, easy to understand, thankfully we don't get to hear that fan @100%����! Seems an awful lot of zigzagging in those graphs - is this the norm for most video cards or..?
 
Commander Suzdal said:
I think it testifies to [H]'s thoroughness, professionalism, and responsiveness to its readership that they took a card that AMD stiffed them on and did such an in-depth series of reviews on it. Only AMD die-hards and ultra-pure SFF fanatics would even buy this thing, and yet they have spent all this time turning over every rock to give the world the whole picture.

It's all treading water until Pascal/Arctic Islands at this point (unless nVidia gives us some kind of Maxwell-based surprise for Christmas), but [H] puts in the time. Kudos.
Click to expand...

I bought one and having some fun with it ;). I don't recommend one unless you put it in a very small case that has good cooling, bigger cases (really slightly maybe) other cards come into play and probably better choice in the end.
 
noko said:
I bought one and having some fun with it ;). I don't recommend one unless you put it in a very small case that has good cooling, bigger cases (really slightly maybe) other cards come into play and probably better choice in the end.
Click to expand...

Well, you know, if you have the means and you like to tinker with unique products, I can see the appeal of the Nano. Glad you're having fun with it!
 
this thread came just as i needed it, hope you guys have seen my Nano help post http://hardforum.com/showthread.php?t=1883424

Since the stock fan was loud and just awful, i did this:

KTGMFcZ.jpg


and here are the results Running Valley ExtremeHD bench

1.
1080/580/+50 PL (artifacts at 580mem)
base temp 29c
max temp during bench 84c
score 3345
mhz dropping to 950-975 when 80c + for most of the benchmark

2.
1000/560/+50PL/ minus 48 mV
base temp 30c
max temp during bench 78c
score 3219
constant 1000mhz

3.
1000/560/+50PL/ minus 72 mV
base temp 31c
max temp during bench 75c
score 3222
constant 1000mhz

So its not worth overclocking this card, its way better to just undervolt it and keep the core at stock while increasing the mem. really. and 1080/570 is the max stable, if i increase voltage, i end up with worse results and a hot card.

if i use the stock fan, my ears bleed from the noise but its a bit cooler than the 140mm noctua.

here are some more pics!

http://imgur.com/a/k2Sfn
 
SeymourGore said:
Wasn't expecting much success with overclocking, and the way [H] got around to do the overclock seems like too many hurdles. And running the fan 100% on a GPU that would more than likely sit beside my TV as a HTPC if I were to use it in a build? No, thank you. Still a neat, little card. It's crazy just how niche this little guy is.

Here's my crazy hypothesis:

When MS and Sony went to AMD to create the GPUs for their next generation consoles, they wanted something very powerful but in a very small package. AMD went ahead and started development on the little guy, however, at some point both MS and Sony decided to make their consoles more price friendly and opted for a lower cost GPU.

At this point, AMD has devoted too much time into this tech and repurposed it for a niche DIY release.

That all sounds crazy and pulled straight out of my ass: but imo, it still makes more sense than AMD deciding to devote so many resources on such a niche product.
Click to expand...

Given that there's no difference between the GPU used for the Nano and the GPU used for the Fury X, you are correct; that sounds crazy.
 
Nano is power starved due to single 8 pin. Even if you threw on a waterblock to eliminate thermal throttling, it'll still throttle due to insufficient power unless you did a bios mod to exceed the normal 8 pin limit of 150W. (which is fine as long as you don't go too nuts with it).
 
You must log in or register to reply here.
Back
Top