390X coming soon few weeks

cJLKhSk.gif
 
I can sleep easy now knowing that 980Ti is just as good.

Edit: The difference will be which card can OC better.
 
Oh, now we have performance only tied with titan x.

We had higher & slower and were just missing the tied performance rumor, I think.
 
Running a tech site must be great times right now. Throw up "leaks" and say "confirmed!" then you get tons of clicks and traffic. I see people do this with Star Citizen too as they have a very vocal fanbase that falls for clickbaits as well.
Less than a week and we will know for sure, poor tech sites will have to come up with other clickbait articles after that....pascal Q1 launch rumors perhaps?


Note that the author does not say he ever tested the real card or values from someone testing the real card. Likely meaning the numbers are extrapolated and made up using various statistics based on leaked stats. WCCFTech had an article like this too with Fury XT ahead of Titan X if I remember correctly.
So still FUD with no real benchmarks
 
Last edited:
that's 3DMark.

IF AMD matches TitanX, why is that a dissapointment?
Because AMD would rather use HBM to lower their power consumption, increase cost, and offer no extra performance rather than optimize their architecture. What a joke, Fiji XT needs to be 5~10% faster than the Titan X for anyone to care... Or they need to price this beast at $500. If Fiji XT is $500 I will buy two of them.
 
I can sleep easy now knowing that 980Ti is just as good.

Edit: The difference will be which card can OC better.

I got a couple of 980 Ti cards just this week and they overclock like mad.
I got essentially a 37% overclock [1400 MHz] (1000 base clock,1075 boost stock) on stock voltage using the reference cooler.:D
 
I got a couple of 980 Ti cards just this week and they overclock like mad.
I got essentially a 37% overclock [1400 MHz] (1000 base clock,1075 boost stock) on stock voltage using the reference cooler.:D
Nvidia should start shipping GPUs with <1000 cores that way people can run them at 4000 MHz and think they got a 300% OC.
 
54-55% scaling over 290x

we'll find out soon enough who's been releasing bs articles

Yea beats the 980ti and Titanx
No idea if its real or what driver or even if it is a XT card.
This with new drivers pretty impressive if accurate this early.
 
Because AMD would rather use HBM to lower their power consumption, increase cost, and offer no extra performance rather than optimize their architecture. What a joke, Fiji XT needs to be 5~10% faster than the Titan X for anyone to care... Or they need to price this beast at $500. If Fiji XT is $500 I will buy two of them.

Meh.

IMHO, if it ties the 980ti, it should be priced as the 980ti.

If it ties the Titan X it should be priced as the Titan X.

I don't understand why everyone has so much higher expectations from AMD than they do from Nvidia.
 
Zarathustra[H];1041655667 said:
IMHO, if it ties the 980ti, it should be priced as the 980ti.
If it ties the Titan X it should be priced as the Titan X.
They will sell precisely zero cards if they do that.
There is absolutely no way to justify paying $650 for an AMD card when you can get the exact same performance from Nvidia at the same price with [potentially] less power usage and more OC headroom.
 
That's a huge disappointment.

Wut? How is that a disappointment...especially if the claims of it performing better than the Titan X at 4K are true?

Can't really call it one way or the other until we know overclocking headroom and pricing, though.

They will sell precisely zero cards if they do that.
There is absolutely no way to justify paying $650 for an AMD card when you can get the exact same performance from Nvidia at the same price with [potentially] less power usage and more OC headroom.

Well, the AMD fanbois will still buy them.

But yes, if they tie in performance, but are inferior in other categories (cooler quality/noise, power usage, etc.), then pricing them the same is a terribad idea.
 
Zarathustra[H];1041655667 said:
I don't understand why everyone has so much higher expectations from AMD than they do from Nvidia.

I wouldn't say I have higher expectations... more of demands if they want my money. I generally feel nvidia offers a superior experience for more money.

AMD often convinces me to purchase their products with similar performance at a fraction of the cost. In exchange I deal with awful and outdated drivers, video cards that run at 90 degrees, and manufactures that aren't EVGA.

Until AMD proves to me that they can give me powerful hardware aaaand excellent support they are going to have to keep selling me discounted performance.

5970+5870, 6990, 590, r9 290 * 2 is my most recent upgrade path so no one can accuse me of being a fan boy lol.
 
They will sell precisely zero cards if they do that.
There is absolutely no way to justify paying $650 for an AMD card when you can get the exact same performance from Nvidia at the same price with [potentially] less power usage and more OC headroom.

So you are the genius at sales and marketing?
Zero cards huh?

If they price it at twice the cost of Titan, people will buy it no matter the cost no matter what the benchmarks.

Why?
Because people believe expensive means it's better, like Nvidia.
 
They will sell precisely zero cards if they do that.
There is absolutely no way to justify paying $650 for an AMD card when you can get the exact same performance from Nvidia at the same price with [potentially] less power usage and more OC headroom.

Really? And the watercooling and low noise level accounts for nothing?
 
Wut? How is that a disappointment...especially if the claims of it performing better than the Titan X at 4K are true?

Can't really call it one way or the other until we know overclocking headroom and pricing, though.



Well, the AMD fanbois will still buy them.

But yes, if they tie in performance, but are inferior in other categories (cooler quality/noise, power usage, etc.), then pricing them the same is a terribad idea.

If they are priced the same, and the performance is similar, why wouldn't you choose Nvidia at that point? Worth it for me to get away from AMD driver support.
 
Really? And the watercooling and low noise level accounts for nothing?

IMHO no. It's not a feature if it's a requirement for their card to run properly. If a 980ti gives me good temps and low noise without a dedicated aio loop then I don't see why I would be excited about a card that needs all this extra cooling to function.
 
IMHO no. It's not a feature if it's a requirement for their card to run properly. If a 980ti gives me good temps and low noise without a dedicated aio loop then I don't see why I would be excited about a card that needs all this extra cooling to function.
EVGA is selling an AIO version of the 980 Ti for about $100 extra. Is it a requirement? No, it's a feature. AMD is doing the same thing but they're selling a water-cooled version as a separate model, "Fury Nano" if you believe the rumors.

The 980 Ti's blower is terrible and it's a 250W+ card. Whatever cooler AMD uses on the Fury X / Pro, I hope it's not a blower.
 
there is both an air-cooled and a water cooled version of Fury.

Its a feature aimed at SFF builders.
 
AMD often convinces me to purchase their products with similar performance at a fraction of the cost. In exchange I deal with awful and outdated drivers, video cards that run at 90 degrees, and manufactures that aren't EVGA.

I've only ever had one EVGA card (my current Titan actually). It's a good card, but there are many makers of good cards. It doesn't really stand out to me.

My best overclocked ever was a OEM board designed Gigabyte 470GTX which was a phenomenal beast, beating out 480 perfomance by a wide margin.

Also, I don't understand why people keep complaining bout AMD drivers. it's a prejudice more than anything else at this point.

AMD's drivers are just fine in single GPU configurations. I tend to like Nvidias driver configuration layout a little more, but that is minor.

Things DO change when you go multi-GPU though, but lets be honest here, neither AMD nor Nvidia are perfect in this regard.

They both suck in multi GPU. Nvidia just happens to suck slightly less.

And as far as the 90 degree thing goes, did you forget about Fermi? :p Let Me remind you. :p
 
Zarathustra[H];1041655722 said:
They both suck in multi GPU. Nvidia just happens to suck slightly less.

Unless you're looking at FCAT results...

...but no one does that anymore.
 
IMHO no. It's not a feature if it's a requirement for their card to run properly. If a 980ti gives me good temps and low noise without a dedicated aio loop then I don't see why I would be excited about a card that needs all this extra cooling to function.

You miss the entire point. Only reason it's being waterecooled is to fit in smallest form factor. They will have aircooled version of the card that too. It's not there cuz card won't function without it, it's there to purely reduce the size and appeal to people that want performance in a small size
 
Last edited:
there is both an air-cooled and a water cooled version of Fury.

Thats been mentioned god knows how many times in this thread yet people see a pic of the aio version and automatically assume its the only version. You could have it at the top of every page and people would still ramble about it "needing" an aio. AMD would be completely stupid to release only an aio version as not everyone have cases that can accommodate it, and even fewer would be willing to change their case just for a new gpu. 295x2 was already a niche market and AIO was literally the only way it could be cooled and remain relatively quiet, powercolor had the devil 13 air cooled version and it didn't impress with its noise levels and throttling issues.
 
Last edited:
You miss the entire point. Only reason it's being waterecooled is to fit in smallest form factor. They will have airfields version of the card that too. It's not there cuz card won't function without it, it's there to purely reduce the size and appeal to people that want performance in a small size

Airfield?

airfield-clipart-stock-vector-airport-terminal-and-liner-fun-cartoon-map-elements-89404003.jpg
 
Holy shit, I was right. Fury X sits right in between 980ti and Titan X. You guyz were all yelling at me that that would never happen since the difference between them was so slim.

VINDICATED! :) (lucky guess)
 
Holy shit, I was right. Fury X sits right in between 980ti and Titan X. You guyz were all yelling at me that that would never happen since the difference between them was so slim.

VINDICATED! :) (lucky guess)

stay-on-target.gif


Stay on target!

Wait for it.

Only a few more days until we have REAL information.

Nothing is confirmed yet. These are still just rumors from unreliable sources.

All we know right now is that there will be a new AMD launch with HBM in the near future. Nothing else.

You can claim "vindicated" once there are charts on the [H] main page :p
 
Holy shit, I was right. Fury X sits right in between 980ti and Titan X. You guyz were all yelling at me that that would never happen since the difference between them was so slim.

VINDICATED! :) (lucky guess)
If you are going by the 3dscore it more like matches Titan x. We will have to wait and see the in game performance
 
People place too much weighting on aggregate performance differences and how that translates into actual usage. With basically comparing two completely different architectures anything within 5% (even 10%) might as well be a wash due to different strengths (work load dependent) and clock speeds. Not to mention the perceptible difference is basically negligible in most scenarios.

You can really see this play out in HardOCP reviews, game settings are essentially either identical across cards in roughly the same class (or even within +/- 1-2 segments), or with some specific work loads swinging either way.

It will likely end up being the most probable scenario in that Fiji is in the same performance class as GM200.
 
People place too much weighting on aggregate performance differences and how that translates into actual usage. With basically comparing two completely different architectures anything within 5% (even 10%) might as well be a wash due to different strengths (work load dependent) and clock speeds. Not to mention the perceptible difference is basically negligible in most scenarios.

You can really see this play out in HardOCP reviews, game settings are essentially either identical across cards in roughly the same class (or even within +/- 1-2 segments), or with some specific work loads swinging either way.

It will likely end up being the most probable scenario in that Fiji is in the same performance class as GM200.

Partially agree and partially disagree.

Settings will most likely be the same and the differences in most cases will not be noticeable.

That being said, it doesn't take much to go from acceptable to unacceptable, in my opinion.

The second my frame rate chart (in something like MSI Afterburner) is anything but a flat line at 60fps in a multiplayer FPS using Adaptive V-Sync, the GPU is dead to me :p
 
Back
Top