Fury Nano Discussion Thread

[RussianSensation]

Perf/watt will be forgotten now, perf/$ will be the new mojo from now on

Not this time. GTX980 owners dug their own perf/$$$ grave by ignoring price/performance for 8+ months before 980Ti launched. As a result, they can't say anything against the Nano's premium pricing without looking hypocritical. $550 GTX980 looks horrendous in this Nano review against its $300-350 R9 290X competitor:

Overall, I think I feel good about providing many gamers the advice to buy an after-market 290/290X/970/390 as a stop-gap 28nm card and then make the proper jump at next gen. 980 was and continues to be in no-man's land as far as price/performance goes, and unfortunately Fury and Fury X fall into the same categories. Whoever gets to 16nm HBM2 first is going to capture a lot of market share as I feel there is a lot of pent-up demand for a GPU upgrade come 2016. At least the Nano gets to claim the fastest miniITX card title in cases with limited spacing (<180mm length) where only the Nano or possibly Fury X would fit.

 

[Techhog]

Well, you could, since you have 1 8 ping power plug + 75W from PCIe, then the max the card can get is 225W, and not the 175W it is limited to now.
This would require a BIOS mod I bet.

Nope. The power limit isn't locked.

 

[Silverforce11]

Well, I've always valued perf/$ so I agree with RS fully. R290/X/970/390 are the bang for buck cards and have been this way for years (minus the 390 which was more recent).

Prior to these, it was the 7950, which was a beast of a card with 45-50% OC potential. But instead of praise, some of the same members on this forum who NOW value OC on Maxwell, blame AMD for setting the default clocks "too low, too conservative", making it look bad that it requires a user to OC it...

So yeah, I've been around long enough to see through the bias.

Nano was never intended as a perf/$ product. In fact, none of the Fiji SKUs are. If you want perf/$, there's plenty of choices.

 

[tential]

Well, I've always valued perf/$ so I agree with RS fully. R290/X/970/390 are the bang for buck cards and have been this way for years (minus the 390 which was more recent).

Prior to these, it was the 7950, which was a beast of a card with 45-50% OC potential. But instead of praise, some of the same members on this forum who NOW value OC on Maxwell, blame AMD for setting the default clocks "too low, too conservative", making it look bad that it requires a user to OC it...

So yeah, I've been around long enough to see through the bias.

Nano was never intended as a perf/$ product. In fact, none of the Fiji SKUs are. If you want perf/$, there's plenty of choices.

HD7950 owner
R9 290 Owner
Probably an Arctic Islands owner too if the perf/$ is right.

I've followed the advice, and built upon that advice.
I've gone from getting a 7950 new, to picking up a r9 290 on the cheap used.

So ya, I prerfer to get my perf/$ option, but I think the Nano is in a class of its own (obviously) with what it offers. It doesn't fit any of my current builds, but if I were making the amount of money I should be making, then sure lol. I'd have these crossfire in a far smaller/quieter rig and a 4K display (freesync as freesync gets to these 4K display sizes now that are usable in a living room), and have that be my mine HTPC.

I really think AMD had great intentions with this lineup. I'm thoroughly impressed with the idea behind it, even if the performance is currently lacking. I really don't think this will be the case with Arctic Islands.

 

[tential]

On a separate note, am I the only one who thinks we need a IQ analysis also in these reviews when you're dropping settings to LOW in the case of Crysis 3? Is this even charting a thing ANY OF US would ever do with an R9 Nano? Or any of those cards?

I don't know, but I don't think testing a card at settings no one will ever play at is useful but that's just my opinion.

This is why I do like HardOCP is they do have a Maximum Settings test.

 

[bystander36]

On a separate note, am I the only one who thinks we need a IQ analysis also in these reviews when you're dropping settings to LOW in the case of Crysis 3? Is this even charting a thing ANY OF US would ever do with an R9 Nano? Or any of those cards?

I don't know, but I don't think testing a card at settings no one will ever play at is useful but that's just my opinion.

This is why I do like HardOCP is they do have a Maximum Settings test.

I agree and I'd really love it if someone took what HardOCP does a little further. I'd like to see them choose the best settings which can get you near 80ish FPS, not just playable FPS (60 FPS is an obvious slot as well). I'll often turn settings down to get near 80 FPS, because I find that is the most enjoyable due to latency and smoothness. It really annoys me to see reviews comparing sub 40 FPS and lower, as most people buying these cards are not going to be playing with those settings.

 

[jpiniero]

People really need to lower their expectations for Arctic Islands (and Pascal for that matter).

 

[RussianSensation]

Obviously.
If they are the same there is no competitive advantage there anymore.

If someone is building a small miniITX case that can fit FuryX/980Ti, well that's the way to go, but if someone wants the smallest miniITX case or is form factor constrained, the Nano is the only game in town. It literally has no competition which makes it a halo miniITX product at the moment. In the case of the Titan or Titan X, after-market 780Ti and 980Ti beat those cards maxed out. Even though the Nano is a bad deal, at least some business case can be made for it. It's actually possible to build a miniITX system that can only fit a Nano or GTX970Mini and nothing faster. In the case of $1000 cards like the Titan or Titan X, they literally have a worthless marketing/e-peen premium for gaming as they offer nothing worth discussing compared to after-market 780Ti/980Ti cards. The Titan X could have made sense in a Tri-SLI or Quad-SLI setup but unfortunately NV's Tri-SLI and Quad-SLI performance and frame times are terrible, so that doesn't work either.

And, the Nano's cooler for the card is actually good, unlike the Titan / Titan X cooler which is basically a jet engine when they are overclocked.

That means in its respective niche category, the Nano is actually a far better executed product than the Titan series is. It's cool, it's quiet, it's easily the fastest smallest card and it doesn't even cost $1000. The original Titan failed to be the fastest card that generation, and aged horribly. The newest Titan X has a crappy cooler for its price tag and its extra VRAM is useless for games. So essentially all the Titan series is a worse executed GTX980Ti with a crappier cooler, worse overclocking performance and worse VRM/mosfet components but costs $350 more. Essentially that means add another $100+ for a better cooler or water-block just to get to the level of a great GTX980Ti. What a rip-off that is. That means it's actually possible to almost cross-fire dual Nanos for $1300 that will wipe the floor with a $1100 water-cooled Titan X.

And when the Nano is max overclocked, it's actually quieter than a stock Titan X.

But how did the online PC gaming community treat the Titan/Titan X (many people worship the overpriced Titan line and buy them at launch) and how did they treat the Nano? The same for sites like HardOCP? So let me get this straight-- some Titan/Titan X owner can justify for spending $1000 on an overpriced tool like the Titan X for his special use case, but he can't fathom how a Nano which is another overpriced tool for its own special use case can make sense? :sneaky:

Also, the very same people who bought GTX980 and defended its $550 price are now bashing the Nano's price/perf. despite it offering the fastest performance in its class? Oh, how amusing to read. News at 11: all that tells us is people who buy NV will only buy NV and will only focus on metrics where NV is on top. This is another case where even if the Nano was $199, GTX980/980Ti/Titan X owners wouldn't buy it anyway so who really cares? :awe:

Well, I've always valued perf/$ so I agree with RS fully. R290/X/970/390 are the bang for buck cards and have been this way for years (minus the 390 which was more recent).

Exactly. I am price/performance guy so neither the Nano, the Fury, nor the Fury X are appealing to me at current prices against better price/performance products like the 290/290X/390/GTX980Ti. But when people who bought $500-650 GTX780s, $500-550 GTX980s or $1000 Titan Xs are constantly bashing the Nano's price/performance prior to launch and now, it's basically The pot calling the kettle black.

I recall certain forum members across AT, TPU, TechReport, HardOCP spouting how they could just buy a GTX970 mini and overclock it to Nano speeds. Now all you hear is silence when real world numbers came out. You basically need a max overclocked 980 to even try to do that.

I guess now all the perf/watt loyalists will start recommending the Nano over GTX980Ti, right?

"Forgetting about the Nano's tiny size and focusing squarely on price and performance, it's the $650 GTX 980 Ti that we're interested in. This comparison sees the Nano trail the GTX 980 Ti by a 13% margin at 1600p and 10% at 4K, though the Nano consumed 16% less power on average, so it looks to be the more efficient graphics card." ~ TechSpot

I personally couldn't care less about perf/watt -- only as its relationship pertains to getting max performance in 250-300W TDP flagship cards -- but it's funny how perf/watt no longer is an important metric at all, esp. since some of the MOST vocal supporters of perf/watt kept claiming for years that AMD will never match NV in perf/watt this generation without a node shrink.

 

[Silverforce11]

Why does AT have AMD Radeon R9 390X at 60.8? There IS NO REFERENCE 390/X. Nobody can buy a reference 390/X.

What is going on there @Ryan?

 

[RussianSensation]

Why does AT have AMD Radeon R9 390X at 60.8? There IS NO REFERENCE 390/X. Nobody can buy a reference 390/X.

What is going on there @Ryan?

Club3D R9 390X 8GB royalQueen OC (Underclocked to 1050MHz)

Sounds like either this card has a very poor cooler, or a very aggressive BIOS fan profile. Maybe this particular card's BIOS is set up to crank the fan speed to 70-100% at max load? It runs at only 67C:

Club3D engineers did a terrible job in finding the balance of noise to performance ratio on their videocard as they could have probably allowed the temps to rise to 78-80C and ensured the card runs far cooler. I couldn't find any 390X Club3D cards on Newegg or Amazon though.

Newegg has 4 Nano SKUs in stock and 1 OOS.

BTW, AT's system with the Nano uses more power than an HD7970 system but the Nano registers at 44.9 db(A) @ 73C while the reference HD7970 comes in at 53.5 db(A) @ 76C. More evidence that a blower cannot compete in terms of noise levels or performance against a properly designed after-market open air cooler.