GTX 980Ti finally launched - MSRP $649 - Reviews

[blackangst1]

official specs Fury:

4GB HBM Memory
4096 Cores @ 1050 Mhz
8.9 billion transistors
AIO water cooled
275W TDP
$649 USD
Release date is June 24th

I think the Fury pro will have the option.

http://wccftech.com/amd-radeon-fury-x-specs-fiji/

[UPDATE June 14th ] We’ve confirmed through our sources that Fury will also feature both liquid cooled and air cooled reference designs just like Fury X.

 

[happy medium]

http://wccftech.com/amd-radeon-fury-x-specs-fiji/

Wow that's gotta be one hellava air cooler to cool a overclocked Fury!

 

[Atomic Playboy]

My GTX 680 is finally giving up the ghost. It's been a trooper, but the I think the HDMI port on it is dead and considering that's the main option for streaming to our TV, I'm probably going to be in the market to replace it. I figure the 980ti will be a good buy for a few years of future-proofing (I got 3 years out of the 680, which is pretty solid). But I have some questions:

I currently have an i5-2500K @ 4.4 GHz on an old Asus p67 motherboard (and 16 GB DDR3). Will I be CPU-bound at Ultra settings in modern games like GTA V or Witcher 3, especially if I'm downsampling @1440P or 4K (yes, I know I wouldn't actually be able to run those games at ultra settings @4K)? If so, would upgrading just the CPU be enough, or would I need a complete platform upgrade? Could I make do with just a new motherboard/CPU and reuse everything else? Should I just wait for Skylake (as I've seen mentioned other places in the thread) or go with a current processor, ie Devil's Canyon?

I figure I've got a few weeks to get things sorted, but I'll probably need to jump on the 980ti before August (hopefully an factory-OC model with a non-reference cooler). I'm just torn on whether I should completely rebuild now or deal with being CPU-bound for a few months until I can put together a Skylake build from scratch.

 

[happy medium]

I currently have an i5-2500K @ 4.4 GHz

That cpu will be just fine for a gtx980ti @ 1440p @ ultra or 4k @ med/high settings..especially with GTA V and The witcher 3.


Wait for Skylake for a system upgrade but for now the gtx980ti would be fine, mabe a little cpu bound, mabe.

 

[Atomic Playboy]

Wait for Skylake for a system upgrade but for now the gtx980ti would be fine, mabe a little cpu bound, mabe.

And even if it is CPU-bound, I suppose it will still be a hell of a lot faster than a 680...

Damn, now I've got to start scouring Internet sites for non-reference 980tis coming out before they all get snatched up. Which will be slightly easier than explaining a $700 graphics card to the wife...

 

[happy medium]

And even if it is CPU-bound, I suppose it will still be a hell of a lot faster than a 680...

Damn, now I've got to start scouring Internet sites for non-reference 980tis coming out before they all get snatched up. Which will be slightly easier than explaining a $700 graphics card to the wife...

Should be 2x your gtx680.
Dude get the Gigabyte G1 gtx980ti,its a beast and about 688$ if you can find one!

http://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_G1_Gaming/

http://www.bhphotovideo.com/c/produ..._gv_n98tg1_gaming_6gd_geforce_gtx_980_ti.html

 

[Majcric]

My GTX 680 is finally giving up the ghost. It's been a trooper, but the I think the HDMI port on it is dead and considering that's the main option for streaming to our TV, I'm probably going to be in the market to replace it. I figure the 980ti will be a good buy for a few years of future-proofing (I got 3 years out of the 680, which is pretty solid). But I have some questions:

I currently have an i5-2500K @ 4.4 GHz on an old Asus p67 motherboard (and 16 GB DDR3). Will I be CPU-bound at Ultra settings in modern games like GTA V or Witcher 3, especially if I'm downsampling @1440P or 4K (yes, I know I wouldn't actually be able to run those games at ultra settings @4K)? If so, would upgrading just the CPU be enough, or would I need a complete platform upgrade? Could I make do with just a new motherboard/CPU and reuse everything else? Should I just wait for Skylake (as I've seen mentioned other places in the thread) or go with a current processor, ie Devil's Canyon?

I figure I've got a few weeks to get things sorted, but I'll probably need to jump on the 980ti before August (hopefully an factory-OC model with a non-reference cooler). I'm just torn on whether I should completely rebuild now or deal with being CPU-bound for a few months until I can put together a Skylake build from scratch.

The 980ti is a massive upgrade from the 680. I just did it and I'm also using the i5 2500k. I'm seeing more than 2x performance in games. And if you're using the 2gb 680 there is a huge gain in IQ from games that were Vram starved and you didn't know it.

 

[tviceman]

To place things in a different perspective and how the market has changed - cut down
GF-110 cores were introduced as little as 289 dollars -- now, cut down bigger dies are launching, as value by some at 650 dollars.

Blame it on conditioning, or the competition's lack of having a clearly competitive product in all regards, or both. Both companies have been working hard to raise prices. One has succeeded while the other is trying hard to ride the coattails, and is trying again starting next week. I'm not buying a $650 graphics card anyways, nor do I care about how big the die, cutdown, bus size, ROP count, or type of ram it uses. If GP104 is 225mm2, only has 32 ROPs, and doesn't use HBM but doubles the performance of GM204 at the same TDP and is priced right for me I'm all in. All I care about is what kind of upgrade I get over what I currently have and the performance I'm getting for the dollar I am spending.

 

[Atomic Playboy]

The 980ti is a massive upgrade from the 680. I just did it and I'm also using the i5 2500k. I'm seeing more than 2x performance in games. And if you're using the 2gb 680 there is a huge gain in IQ from games that were Vram starved and you didn't know it.

Are you OCing the 2500K at all? Is it bottlenecking you in any way? And reference 980ti or one of the factory OCed models? That Gigabyte G1 model is so tempting; I'm half glad it's sold out everywhere so I have time to ponder my options.

 

[Majcric]

Not from a raw performance standpoint. However, I do think it would be better to have more cores/threads. i5 @4.5. My card is reference. I actually have a preference for reference cards.

Take a look through this thread :http://forums.anandtech.com/showthread.php?t=2434787

 

[davel]

The 980ti is a massive upgrade from the 680. I just did it and I'm also using the i5 2500k. I'm seeing more than 2x performance in games. And if you're using the 2gb 680 there is a huge gain in IQ from games that were Vram starved and you didn't know it.

I second that, I just upgraded to 980ti from a gtx 680, though i am using i5-3570k, it was a massive improvement and I am still @ 1080p

Playing GTA 5 with max settings with no stuttering is amazing.

 

[DEW73]

Wow that's gotta be one hellava air cooler to cool a overclocked Fury!

Isn't the Fury supposed to be more energy efficient/cooler running than the last generation AMD cards?

 

[DEW73]

That's the final nail in the Fury's coffin.

What was the first (or second) nail in Fury's coffin? Or does it just need one (final) nail?

 

[happy medium]

What was the first (or second) nail in Fury's coffin? Or does it just need one (final) nail?

The first one was the release of the 970/980
Second was the release of the 960.
third and final was the 980ti.

The Nano is the only chip that looks interesting in the AMD lineup, but 2 or 3 months away, it might be too late.

 

[xthetenth]

The first one was the release of the 970/980
Second was the release of the 960.
third and final was the 980ti.

The Nano is the only chip that looks interesting in the AMD lineup, but 2 or 3 months away, it might be too late.

You seem confused. The 970 is not at all competing with the Fury. The 980 and 960 are legitimately bad cards for their price, and only the 980 can be reasonably said to be potentially competing with the Fury, but that's more on the back of its wildly inflated price putting it up against a card it cannot compete with on performance.

The 980ti is actually competition, so you did manage to get one of three right, which is a plus. However, it is competition, and until benchmarks come out I'd hardly go calling it a nail in the coffin unless I was working from a perspective where any NV card at all is automatically better than any AMD card, in which case I'd be able to spend so much more time on things other than discussing hardware.

 

[happy medium]

The 970 is not at all competing with the Fury. The 980 and 960 are legitimately bad cards for their price

Please check market share for the past 10 months and tell me what cards sold more and took market share. I'll give you a hint, gtx960,970,980. the first 2 nails, and the gtx980ti is the final nail.

The real 4th and final nail is the rebadge of the entire 280/290x lineup.
They are killing themselves now.

 

[tviceman]

You seem confused. The 970 is not at all competing with the Fury. The 980 and 960 are legitimately bad cards for their price

Which is crazy because they've sold in the millions. And it's even more crazy to think that Nvidia has way more pricing room to drop prices than AMD since GM204 is smaller than Hawaii and Fury and uses less power and has less vram than the new 390/390x. If AMD actually forces a price drop on the 980, then it will end up hurting AMD more, IMO.

 

[happy medium]

If AMD actually forces a price drop on the 980, then it will end up hurting AMD more, IMO

exactly!

 

[xthetenth]

Please check market share for the past 10 months and tell me what cards sold more and took market share. I'll give you a hint, gtx960,970,980. the first 2 nails, and the gtx980ti is the final nail.

The real 4th and final nail is the rebadge of the entire 280/290x lineup.
They are killing themselves now.

I too believe all video card purchases are fungible, and that if AMD released a really good APU they'd cannibalize 980 Ti sales because a good product selling well affects all market segments.

I mean if you were wrong there'd be people selling 970s to buy 980 Tis because the two cards are in different performance categories, and that most certainly isn't a thing that's happening. /s

Tell you what, if the Fury performs well I'll send you a picture of my case with my 970 in it and a picture with the Fury that replaced it in there.

Which is crazy because they've sold in the millions. And it's even more crazy to think that Nvidia has way more pricing room to drop prices than AMD since GM204 is smaller than Hawaii and Fury and uses less power and has less vram than the new 390/390x. If AMD actually forces a price drop on the 980, then it will end up hurting AMD more, IMO.

The market including a large number of people who are willing to buy a bad card because it's under $200 and they are easily swayed by marketing means literally nothing when discussing the potential sales of a product over two and a half times that price. This is not a difficult concept.

The 390's positioning in the market is (surprise, surprise) totally irrelevant to the Fury.

 

[tviceman]

The market including a large number of people who are willing to buy a bad card because it's under $200 and they are easily swayed by marketing means literally nothing when discussing the potential sales of a product over two and a half times that price. This is not a difficult concept.

The 390's positioning in the market is (surprise, surprise) totally irrelevant to the Fury.

Look how wrong you are. If a $550 product is 5 times faster than a $200 product, there will be plenty of people unwilling to spend $200. The 390's position is entirely relevant to Fury because otherwise prices would make absolutely zero sense in the market. If Fury forces a price drop on the 980, the 980 will make the 390x look really, really stupid at $429. That will then force a drop on the 390x, which will force a drop on the 390....

Suddenly the dots connect.

 

[RussianSensation]

The first one was the release of the 970/980
Second was the release of the 960.
third and final was the 980ti.

The Nano is the only chip that looks interesting in the AMD lineup, but 2 or 3 months away, it might be too late.

1. 970 doesn't compete with $550+ cards
2. 980 has no hope of competing with Fury PRO or X.
3. 960 has nothing to do with this pricing segment. For brand objective users, it was irrelevant for gaming from day 1 as 280X and 290 are better gaming cards.
4. We still don't know how the pair or Fury PRO/X cards perform to claim that 980Ti nailed them.
5. You completely ignored dual-Fiji card launching in the fall. NV has no response to that right for now.

Wow that's gotta be one hellava air cooler to cool a overclocked Fury!

> 11,500 posts and you still don't know that you can have an air cooler than can easily tame 350W of GPU power in overclocked states while running < 75C?
https://www.youtube.com/watch?v=rwRWi3Bz0qc

They are killing themselves now.

More doom and gloom from armchair financial advisors. Just what we needed.

Suddenly the dots connect.

Yes they do. If the entire AMD stack was 50% faster and cost 50% less than each NV's product, AMD would just gain market share back close to 50% and not much more. Considering the market bought $200 960 2GB that's 50% slower than a $250 R9 290 4GB, how are you not connecting the dots?

You now how many people buy Honda Civic and Toyota Corolla? Best selling cars in their class but are basically garbage compared to VW Golf or Mazda 3. Sales do not always mean the product was worth buying. Only a person who doesn't care about car safety or a person 100% ignorant about all things automotive would buy a Toyota Corolla with rear drum breaks in 2015 but boy millions of people do.

Brand name/marketing would still ensure 980Ti could outsell an AMD card 20-30% faster and you know it. Just be man enough to admit it. AMD destroyed NV for 6 months with HD5850/5870 and market share still didn't go to 75% like it is now for NV. AMD has the perf/watt crown with HD4000, 5000 and 6000 series and it took NV 6-9 months to roll-out Kepler top-to-bottom. Heck, the fact that most PC gamers ignored bitcoin mining that made AMD cards free and they made thousands of dollars is enough proof brand name, marketing and awful knowledge of tech for the average PC gamer is what drivers sales. This is the only generation in the last 20 years I can think of where people are buying 750/750Ti and foregoing the R9 270/270X which are 30-45% faster for $20 more. This is the only generation I can think of where people are buying a $200 960 with gimped VRAM, a card that's SO overpriced, even $360-400 960 SLI can't beat an after-market 290. That's truly pathetic that 960 still sells well. Marketing and consumer ignorance FTW.

NV even managed to sell GeForce 5 and 7, both far inferior to ATI's offerings of that time. NV is like Apple. They have great/excellent products but even if they release turds, they still sell well. Cards like FX5200, GTX550/550Ti, 650/650Ti and 960 are pure trash and sell well. That's evidence in itself the types of gamers buying these cards are clueless/brand driven.

Therefore, it's a foregone conclusion that when it comes to market share and sales 980Ti would beat AMD's Fury even if AMD beat 980Ti in every metric. When I was working in Asia for 2 years, in some countries, you can't even buy AMD cards. You'd have to order them to that country's store from China. Why? The store owner would tell me no one here buys NV. The same country where people think Coca Cola, Heinz Ketchup, Levi's jeans, NIKE shoes, Apple products and Toyota/Lexus are the best products in the world. So many people are brand brainwashed in 3rd world countries, it's mind blowing. Just like Audi sells very poorly in the US but Audi owns BMW/Mercedes in China is 100% proof that brand name and marketing dictates most premium/non-essential consumer purchases in today's world.

I am pretty sure if Fiji was an NV product, you'd be saying how outdated 980Ti is for using that junky reference blower, outdated GDDR5 tech and requiring after-market versions just to keep up with a reference Fury X at high-rez gaming.

 

[moonbogg]

People who are concerned about 2500K bottlenecking 980ti, I wouldn't be too concerned about that. I have the same CPU when it comes to most games, and I got two 980ti's, and if there is a bottleneck, its certainly not a big one because everything hauls insane ass. You'll definitely be GREAT with a single 980ti and 2500K. Friggin GREAT!

 

[tviceman]

Yes they do. If the entire AMD stack was 50% faster and cost 50% less than each NV's product, AMD would just gain market share back close to 50% and not much more. Considering the market bought $200 960 2GB that's 50% slower than a $250 R9 290 4GB, how are you not connecting the dots?

There are a plethora of reasons for why AMD is not selling well and you know this. Just like GTX 480 and 470, r9 290x and 290 didn't get the best reviews, despite having great outright performance. Just like gtx480 and 470, the prices dropped relatively quickly. Unlike the GTX480 and 470, the 290x and 290 were not replaced with better products relatively quickly and instead the competition trounced on them with an amazingly reviewed slew of new products. We've all said "first impressions have a lasting effect" on these forums and this is no exception. Do I think AMD can sell at price parity against Nvidia at this point in time? Probably not quite, not unless they have tricks up their sleeves..... like HBM which sounds aMaZ!ng! Also, to use your analogy, who wants to buy a car from a company that is financially in bad shape, switching out CEO's like tampons, and has rumors swirling several times a year of buyouts or bankruptcy? Same thing applies to the computer segment. Having a negative stigma like that following AMD around, no matter how much or little it's warranted, hurts.

Therefore, it's a foregone conclusion that when it comes to market share and sales 980Ti would beat AMD's Fury even if AMD beat 980Ti in every metric.

I disagree. If Fury X smacks Titan X by 10% in games and overclocks as well as GM200, then it will sell extremely well for a high end card and force a price drop on the 980 TI. If Fury X ties Titan X (which means it's only about 3% faster than the 980 TI) or doesn't have much OC headroom, then I don't see the status quo changing at all. Nvidia executed the 980 TI launch very well. It reviewed well, it's pricing was received well, and it got plenty of eager enthusiasts to forget about Fiji. If Fury X had come out ahead of 980 TI, then Nvidia would not have garnered nearly as much acclaim for the GTX 980 TI.

 

[happy medium]

. 970 doesn't compete with $550+ cards

No its been out for almost a year and has been selling like hotcakes. Taking market share.

980 has no hope of competing with Fury PRO or X.

It don't have to it will be priced 100$ less and has been out for almost a year and has already takin its market share from AMD.

960 has nothing to do with this pricing segment. For brand objective users, it was irrelevant for gaming from day 1

Oh I guess your gonna say this card had nothing to do with cutting AMD's market share?
irrelevant for gaming? so we only make money off gaming cards now?

We still don't know how the pair or Fury PRO/X cards perform to claim that 980Ti nailed them.

Who said anything about "A PAIR" of cards? Seems AMD went back to the drawing board when the gtx980ti released @ 650$. How many post do you have (and wall of %$ texts ), and you cant see what that means?

11,500 posts and you still don't know that you can have an air cooler than can easily tame 350W of GPU power in overclocked states

WHo said in an overclocked state the Fury was only gonna be 350watts?
You have some links you wanna share. ANd my guess would be well over 400 watts of freight train like noise.

More doom and gloom from armchair financial advisors. Just what we needed.

So I guess you think its a good thing they rebranded there 390/380 lineup with more expensive cards at basically the same performance/wattage? Don't have to be a financial advisor to give that answer do ya?

You now how many people buy Honda Civic and Toyota Corolla? Best selling cars in their class but are basically garbage

Save your spin and walls of text jibberish for the other suckers.
The AMD name sucks and will continue to suck until their products/software/marketing are at the level of the #1 brand Nvidia. That's a fact and no wall of text or %/$ BS will change that.
Market share: Nvidia78%, AMD 22%, the numbers speak for themselves

 

[xthetenth]

Look how wrong you are. If a $550 product is 5 times faster than a $200 product, there will be plenty of people unwilling to spend $200. The 390's position is entirely relevant to Fury because otherwise prices would make absolutely zero sense in the market. If Fury forces a price drop on the 980, the 980 will make the 390x look really, really stupid at $429. That will then force a drop on the 390x, which will force a drop on the 390....

If that situation had ever existed in the history of graphics cards, maybe but very likely not. There's plenty of people out there who simply don't have $500 to spend and definitely don't considering that we live in reality where you're lucky to be paying twice the price for 170% of the performance.

Next you're somehow trying to say that the Fury pushing the 980 down and eventually pushing the 390s down is evidence that the 980 damaged the Fury's positioning? How in the world is a card being priced to cause the market to shift to reflect it a sign that the card is anything but well priced? It seems you consider scoring points on the red team to be a valid refutation that the Fury is well positioned, even when it includes a tacit admission that the Fury is well positioned!

Suddenly the dots connect.

You should try being actually capable of a coherent argument before being smug.

Save your spin and walls of text jibberish for the other suckers.
The AMD name sucks and will continue to suck until their products/software/marketing are at the level of the #1 brand Nvidia. That's a fact and no wall of text or %/$ BS will change that.
Market share: Nvidia78%, AMD 22%, the numbers speak for themselves

Good on you for putting your money where your mouth is and supporting a tautological argument.

You shouldn't be cheering how good your preferred brand is at overcharging its customers.