GTX 980Ti vs R9 Fury X in Small Form Factor Application

[Ducman69]

Linus, generally considered an NVidia fanboi, actually gives props to the Fury X as the better solution for SFF gaming, for those so inclined to go with a mITX build.

https://www.youtube.com/watch?v=ZbONfPkwa_s

He usually hates on AMD, so it was almost like pulling teeth where he had to admit the numbers don't lie and give them the win this time.

But he does still prefer the 980Ti in a large case with lots of cooling.

 

[[L]imey]

Be interested to see the 980ti hybrid vs the fury X in a small form factor shootout (using the aio space for the gpu in both scenarios.)

 

[Ducman69]

I'd assume the hybrid would no longer be thermally limited and thus faster, but the hybrid would be more expensive and not quite as compact as the Fury X of course since it now not only has the water lines like the Fury but is bigger to boot.

Fury X vs 980 Ti Hybrid:

LENGTH (mm): 194 vs 268 (32% difference)
WIDTH (mm): 102 vs 112 (9% difference)

 

[5150Joker]

Linus, generally considered an NVidia fanboi, actually gives props to the Fury X as the better solution for SFF gaming, for those so inclined to go with a mITX build.

https://www.youtube.com/watch?v=ZbONfPkwa_s

He usually hates on AMD, so it was almost like pulling teeth where he had to admit the numbers don't lie and give them the win this time.

But he does still prefer the 980Ti in a large case with lots of cooling.

Did you bother watching the video you're linking? Specifically: https://youtu.be/ZbONfPkwa_s?t=232 And all he ran was Crysis 3..gimme a break.

Not exactly what I'd want in my SFF. Fury X was pointless on launch and it's even more pointless with AMD releasing their air cooled Fury at $100 less with 95% of the performance. At least the Titan X had a cushion of 3 months before 980 Ti came out and is still the top reference card, Fury X can't even claim that much.

 

[jwcalla]

What the heck did I just watch? They only figured out at the end that they put the fan on backwards and then.................................. end of video.

Anyway... the Fury X wouldn't work in my SG07 for a couple of reasons... no room for the cooler, and the case uses a positive pressure design.

What you actually want in a mITX case is a card with a blower fan that sucks directly from outside and blows it right back out.

 

[TaintedSquirrel]

Did you bother watching the video you're linking? Specifically: https://youtu.be/ZbONfPkwa_s?t=232 And all he ran was Crysis 3..gimme a break.

Not exactly what I'd want in my SFF. Fury X was pointless on launch and it's even more pointless with AMD releasing their air cooled Fury at $100 less with 95% of the performance. At least the Titan X had a cushion of 3 months before 980 Ti came out and is still the top reference card, Fury X can't even claim that much.

Did YOU even watch the video?
He said the Fury X is actually better for SFF since it can maintain full boost clocks while the 980 Ti overheats.

 

[Squizzles]

Did YOU even watch the video?
He said the Fury X is actually better for SFF since it can maintain full boost clocks while the 980 Ti overheats.

As you most likely know, it all depends on what case you put it in. Heck even in a ATX case, if I cram the card up next to another one with little airflow for the blower, the same effect will be made. Even though a lot of SFF cases can fit a big card...not all are made to handle the cool air required by a high power card and its cooler.

 

[5150Joker]

Did YOU even watch the video?
He said the Fury X is actually better for SFF since it can maintain full boost clocks while the 980 Ti overheats.

87 celsius CPU temps..really great for SFF..I mean...really really great.

 

[TaintedSquirrel]

87 celsius CPU temps..really great for SFF..I mean...really really great.

In a burn test. With the exhaust fan pointing backwards.
He even makes a point to emphasize that a cooler running GPU is more important to gaming than the CPU, which is why the extra boost clock got him more performance in the end. If you can only choose to put either the CPU or GPU under water, the GPU will give you more benefit. I think he said his 980 Ti stopped at 1070 boost.

Fury X was running 67C in his bench while the 980 Ti went 80C+ and throttled. Inverse of the CPU benchmark.
He had to include the CPU chart to meet his Nvidia fanboy quota. Notice he didn't show the GPU temperature?

 

[kasakka]

The 980 Ti runs hot in general so I would not cram one in a SFF without a watercooler.In that sense the Fury X is a lot better choice, though the radiator poses its own set of placement problems and where to put the pipes. I recently installed an AIO cooler on my MSI 980 Ti Gaming card and it pretty much halved idle temps and dropped load temps around 20 degrees even with higher overclocks.

Are there any small form factor cases specifically built with watercooled (AIO or custom loop) coolers in mind? Because more often than not it's trying to cram a radiator in a slot intended for a fan.

 

[Dayaks]

I cannot imagine a sff case having temp issues if you get the right one. If designed right they take in fresh air from the side of the case for the GPU.

 

[Ducman69]

Did you bother watching the video you're linking? Specifically: https://youtu.be/ZbONfPkwa_s?t=232 And all he ran was Crysis 3..gimme a break.

Not exactly what I'd want in my SFF. Fury X was pointless on launch and it's even more pointless with AMD releasing their air cooled Fury at $100 less with 95% of the performance. At least the Titan X had a cushion of 3 months before 980 Ti came out and is still the top reference card, Fury X can't even claim that much.

Can't tell if you're trolling or serious, lol!

Linus is a hardcore self-admitted NVidia fan, so he had to give the NVidia as much positive spin as possible, but this isn't rocket science here. Crysis 3 was used because cranked up its a good game to test the thermals for a CPU and GPU combo since it uses both pretty good. We don't need to compare 20 games, because that was already done on an open bench test by everyone and their mother... although not as many on the new 15.7 drivers (supposedly they give a bit of a performance edge to AMD, just as the NVidia drivers a little while back gave a nice performance bump).

All you need to know is, how is fitment of the Fury X and 980Ti in a SFF, how are the thermals, and do the thermals differ enough to even really affect the bottom line or is it so close its a non-issue.

This answered that:
1) Yes, the Fury X fits better with its notably smaller card, even with the water lines since the lines are reasonably small and flexy.
2) Yes, GPU thermals matter most, and the Fury X was far superior in this regard, with its watercooling cooling taking care of not just the GPU, but the VRM and VRAM to dump all that heat outside of the case.
3) Yes, the GPU thermals were a big enough delta between the two that it affected the bottom line in performance in AMD's favor.

Yes, there is the 980 Ti hybrid, but its still a considerably larger card, especially needing to bend the tubes out of the top of the card, and it still needing air-cooling breathing room, and that's a noticable price bump. So for a SFF build, its hard to argue against the Fury X, and I don't see how an aircooled Fury negates the advantage of water on a SFF that the Fury X provides.

I find it hilarious that even when the most NVidia loving "youtube celebrity" delivers the message as softly as possible that the Fury X is better in a SFF build, you still go into convulsions like someone just shot you in the neck with a poison truth dart! C'mon, if even Linus of all people can give the Fury X props in a particular niche, you can't possibly hit a fanboi level higher than that.

 

[5150Joker]

In a burn test. With the exhaust fan pointing backwards.
He even makes a point to emphasize that a cooler running GPU is more important to gaming than the CPU, which is why the extra boost clock got him more performance in the end. If you can only choose to put either the CPU or GPU under water, the GPU will give you more benefit. I think he said his 980 Ti stopped at 1070 boost.

Fury X was running 67C in his bench while the 980 Ti went 80C+ and throttled. Inverse of the CPU benchmark.
He had to include the CPU chart to meet his Nvidia fanboy quota. Notice he didn't show the GPU temperature?

All he had to do was move the GPU temp target to 91C and increase fan speed on the blower and problem solved. Yet with Fury X shit inside would have baked.

 

[TaintedSquirrel]

All he had to do was move the GPU temp target to 91C and increase fan speed on the blower and problem solved. Yet with Fury X shit inside would have baked.

And yet if this were an AMD card running like a leafblower at 91C then you'd be on the other side of the issue saying "Why should users have to deal with temps in excess of 90C+ and unbearable noise when Nvidia offers a more efficient solution?"

I could probably dig through your post history and find that exact sentiment in different terms. If Fury X were an Nvidia SFF card with factory watercooler don't even try to tell me you wouldn't be praising it up and down right now.

There's really no point saying why that's not a fair comparison when there are literally thousands of Nvidia fanboys who spent an entire year arguing against AMD's shitty blower. And now here we are, the tables have turned... and so has the debate. It's very odd the things people will say without recognizing their own hypocrisy.

 

[5150Joker]

And yet if this were an AMD card running like a leafblower at 91C then you'd be on the other side of the issue saying "Why should users have to deal with temps in excess of 90C+ and unbearable noise when Nvidia offers a more efficient solution?"

I could probably dig through your post history and find that exact sentiment in different terms. If Fury X were an Nvidia SFF card with factory watercooler don't even try to tell me you wouldn't be praising it up and down right now.

There's really no point saying why that's not a fair comparison when there are literally thousands of Nvidia fanboys who spent an entire year arguing against AMD's shitty blower. And now here we are, the tables have turned... and so has the debate. It's very odd the things people will say without recognizing their own hypocrisy.

There are NVIDIA 3rd party hybrids and I don't really care about them. Point is this test was completely and utterly pointless. Fury X is shit and a dud, especially vs 980 Ti at that price point. And then the Fury at $550 just makes it entirely pointless.

 

[TaintedSquirrel]

There are NVIDIA 3rd party hybrids and I don't really care about them. Point is this test was completely and utterly pointless. Fury X is shit and a dud, especially vs 980 Ti at that price point. And then the Fury at $550 just makes it entirely pointless.

Well it's Linus, he does a lot of pointless videos (and mostly misleading/inaccurate).
If someone were spending this kind of dough on a high-end small build they're probably better off with a 980 Ti Hybrid or putting their own watercooling solution on a 980 Ti, closed or open loop. The extra OCing headroom will cover the extra cost in the end.

The point is, this test speaks for itself, but only in the specific environment he tested. It's not really the end-all be-all solution... Or even a good solution. Just one of many possible solutions, one in which the Fury X pulls ahead.

 

[Ducman69]

There are NVIDIA 3rd party hybrids and I don't really care about them. Point is this test was completely and utterly pointless. Fury X is shit and a dud, especially vs 980 Ti at that price point. And then the Fury at $550 just makes it entirely pointless.

Let me explain this more slowly for you...

The Fury is not watercooled and is not as small.

The 980Ti was an inferior choice in a SFF case because its not watercooled and is not as small, limiting how small of a case you can fit it in, and causing thermal throttling.

The Fury X is unique in that not only is it a watercooled GPU, but it watercools the VRM and VRAM to boot, dumping all of that heat outside the chassis, which is important when the chassis is very small, and its unique HBM memory allows it to use an extremely small board which is a full 1/3rd shorter than the 980Ti.

The combination of an extremely compact card with an excellent full watercooling solution gives it an edge in a high-end SFF niche market. It doesn't mean the FuryX is better than the 980Ti for all builds. That's a good thing to have options, so I'm not sure why you're so perturbed by that.

 

[5150Joker]

Let me explain this more slowly for you...

The Fury is not watercooled and is not as small.

The 980Ti was an inferior choice in a SFF case because its not watercooled and is not as small, limiting how small of a case you can fit it in, and causing thermal throttling.

The Fury X is unique in that not only is it a watercooled GPU, but it watercools the VRM and VRAM to boot, dumping all of that heat outside the chassis, which is important when the chassis is very small, and its unique HBM memory allows it to use an extremely small board which is a full 1/3rd shorter than the 980Ti.

The combination of an extremely compact card with an excellent full watercooling solution gives it an edge in a high-end SFF niche market. It doesn't mean the FuryX is better than the 980Ti for all builds. That's a good thing to have options, so I'm not sure why you're so perturbed by that.

Doesn't give it an edge in shit, the test wasn't the least bit valid. But thanks for playing anyway.

 

[Ducman69]

Doesn't give it an edge in shit, the test wasn't the least bit valid. But thanks for playing anyway.

LOL! So the fact that the card is considerably smaller with a superior cooling solution for a SFF application is void again... how? *pops popcorn*

 

[sonsonate]

LOL! So the fact that the card is considerably smaller with a superior cooling solution for a SFF application is void again... how? *pops popcorn*

To be fair, the 980Ti Hybrid exists, but I can't understand how someone could possibly be so obtuse. I guess the block button exists for moments like this.

 

[5150Joker]

LOL! So the fact that the card is considerably smaller with a superior cooling solution for a SFF application is void again... how? *pops popcorn*

Because it's a worst case scenario and he didn't adjust any of the temperature sliders for the 980 Ti. We all know AIO runs cooler for the GPU, that isn't a revelation in itself. However, trying to make the claim that the Fury X is somehow better for SFF based on that limited and flawed test is laughable.

 

[Ducman69]

Because it's a worst case scenario and he didn't adjust any of the temperature sliders for the 980 Ti. We all know AIO runs cooler for the GPU, that isn't a revelation in itself. However, trying to make the claim that the Fury X is somehow better for SFF based on that limited and flawed test is laughable.

So the only thing you don't like is that he wasn't running the 980Ti overclocked at very high temperatures? Can you link an article that shows the 980Ti remains stable overclocked over long periods of time when its cooking itself? And why would you even want that... it makes no sense.

Even on an open bench, the 980Ti is considerably noisier than the Fury X, so why would you want its fans blasting on turbo mode in a constricted SFF case running beyond normal design temps?

 

[jwcalla]

The combination of an extremely compact card with an excellent full watercooling solution gives it an edge in a high-end SFF niche market.

It depends on the SFF case. Even the one tested by Linus doesn't seem appropriate. It looks like a positive-pressure case so unless you enjoy dusting off your motherboard every two weeks that setup is not going to fly.

Then figure in the fact that a lot of SFF cases probably couldn't even fit that huge radiator and that many have separate vents for blower-fan cards, so a blower card should run just as cool in such a case as in the big ATX cases.


The Fury X card (PCB) is small, but the entire unit is not at all small.

 

[Ducman69]

It depends on the SFF case. Even the one tested by Linus doesn't seem appropriate. It looks like a positive-pressure case so unless you enjoy dusting off your motherboard every two weeks that setup is not going to fly.

Then figure in the fact that a lot of SFF cases probably couldn't even fit that huge radiator and that many have separate vents for blower-fan cards, so a blower card should run just as cool in such a case as in the big ATX cases.


The Fury X card (PCB) is small, but the entire unit is not at all small.

Its considerably smaller (and less expensive) than a 980Ti hybrid, with thinner more flexible hoses that come out of the very short card for easy packaging in most SFF cases.

The AIO on the Fury X runs quieter and doesn't vent any heat into the rest of the case, which is an obvious advantage. The 980Ti is considerably noisier and the blower would require a SFF case design with an appropriately placed separate intake for the GPU, a design limitation not otherwise required.

Aftermarket quieter and more efficient air-cooled coolers like the tri-fan design 980Tis (or regular Fury) just blow the heat everywhere inside the case, which again isn't an option in most air constricted SFF cases.

Its just about using the right tool for the job, and it seems painfully obvious that the Fury X is the smallest high performance watercooled card on the market, almost begging to be used in SFF designs.

 

[Dayaks]

I've done 250D builds which is labeled as SFF. The nVidia blowers suck in air through a filtered side panel and exhaust it out the back.

Now, if I was doing a SFF build today I'd probably do a Fury X. The idea of exhausting the heat from the GPU and CPU out the side tickles me in the right places. IIRC the 250D has room for 2x 120mm fans. One for the GPU and one for the CPU.

My only concern with the Fury X is thermals with the rubber shroud removed show it having quite high temps. Some reviews didn't make sense as the GPU was cooler than the liquid. Others didn't make sense because the whole god damn GPU was at 100C... I haven't looked in a week or two though.

 

[jwcalla]

Again, it depends on the case. The Fury X wouldn't even work in my mITX case, and I doubt it's an appropriate solution for even the case Linus tested.

I'm not saying a non-blower fan card is appropriate either.

 

[Ducman69]

If you look at 2:49 mark, it fit in the chassis with room to spare, whereas the 980Ti was barely jammed into the spot.

Non-blower aircooled cards are usually a disaster in SFF cases since the heat just gets swirled around in the confined space, and again the blowers while a little better aren't just noisy as seen, but not very efficient at keeping the card cool either (there's a reason no one buys reference blower cards):

Fact is, the 980Ti needs boost to shine, but it needs to be in an appropriately sized good airflow case to achieve that. That's why 980Ti SLI is often so disappointing, its really just about breathing room. Blower designs are frankly obsolete, and I find it silly that people argue that a larger card with a blower is a better solution in SFF applications to an AIO, especially when they have an example from an NVidia fanboi no less flat out forcing himself to admit it on tape.

The bad parts of the review are actually in the Fury X's favor BTW, not the other way around, as they were showing temps with the GPU heat pumped into the chassis the wrong way around, which they later corrected... something that should have been a no brainer that you always exhaust radiators out of the chassis, not into them. Even with such a case of pure stupidity, the Fury X didn't throttle and bested the 980Ti.

Ideally, I'd actually go for a SFF case that has a 240mm radiator mount, then you could fit a 120mm CPU AIO and FuryX side by side venting out.

 

[Ducman69]

Speaking of, has anyone seen a Fury X build with an NCASE? Would make a great gaming HTPC where small size AND low noise are both primary considerations.

 

[durquavian]

I've done 250D builds which is labeled as SFF. The nVidia blowers suck in air through a filtered side panel and exhaust it out the back.

Now, if I was doing a SFF build today I'd probably do a Fury X. The idea of exhausting the heat from the GPU and CPU out the side tickles me in the right places. IIRC the 250D has room for 2x 120mm fans. One for the GPU and one for the CPU.

My only concern with the Fury X is thermals with the rubber shroud removed show it having quite high temps. Some reviews didn't make sense as the GPU was cooler than the liquid. Others didn't make sense because the whole god damn GPU was at 100C... I haven't looked in a week or two though.

Those temps were burn runs, heavy stress tests. Gaming runs never got anywhere near 100C. Most showing 75C max with gaming runs.

 

[Zuul]

waah waah

hahahahaha delicious tears for breakfast, thanks!