AMD drops it all, who will buy the new Fury?

[RussianSensation]

I used Ati/AMD cards exclusively for over a decade. Not any more. You've burned me enough. Until they can demonstrate an ability to deliver what they promise and release stable drivers on a consistent/frequent basis, I am steering clear of AMD.

I've said rock solid drivers with GeForce 6, 8, HD4000, 6000 and 7000 series. You must be talking about CF situations then because AMD's drivers have mopped the floor with NV's when it comes to GCN vs. Fermi/Kepler generation. If anything, AMD's driver team has been way more rock solid than NV's in the last 3 years when it comes to single GPUs.

This ^100000x. AMD has, in the past, clocked their high end halo part beyond the sweet spot of perf/w to be more competitive at stock speeds. Aftermarket 980 TI's are getting 30% overclocks vs. reference 980 speeds with decent (but not ultra-quiet) noise levels and temps. It is a tough order to keep up with that, but at least AMD strapped Fury with the cooling capable of easily handling it.

Disclaimer: I'm not interested in Fury or 980 TI.

True, for brand agnostic/objective gamers, 980Ti OC vs. Fiji OC is a key comparison. For all those other gamers who for years ignored HD7950/7970/7970Ghz overclocking though and constantly linked us after-market 680 reviews vs. a reference HD7000 card, it's an interesting position how now a reference Fiji is a fail unless it outperforms a max overclocked after-market 980Ti card.

You are missing some factors - we would also need to see an after-market Fury X (will this exist?) vs. an after-market 980Ti. Secondly, you aren't at all compensating for Fury X exhausting 300W of power out of the case while running cool and quiet. An after-market 980Ti cannot claim that. This gets worse for SLI vs. CF comparisons.

Right now NV charges $750+ for EVGA 980Ti Hybrid for the same setup AMD is bringing for $650.

Another thing to keep in mind is TPU sometimes doesn't use AA in some of their game testing. Sites like Sweclockers do. The end result is 980Ti isn't even 40% faster than a 290X at 4K. That means if Fiji scales well, it's possible it might end up 7-8% faster than 980Ti at stock which means it doesn't need to overclock as much as the 980Ti.

They key advantage of a 980Ti is that it doesn't use that much more power when overclocked. If both of them are very similar in performance in overclocked states but Fiji OC uses 450W of power or something, then it could tilt things more in 980Ti's favour.

----

As far as the OP's question goes, I am currently thinking about upgrading. Most likely will wait to 14nm/16nm GPUs or if I do decide to upgrade, will wait much later than now. In the summer time I play sports and travel so the card would just be idling most of the time.

 

[tviceman]

True, for brand agnostic/objective gamers, 980Ti OC vs. Fiji OC is a key comparison. For all those other gamers who for years ignored HD7950/7970/7970Ghz overclocking though and constantly linked us after-market 680 reviews vs. a reference HD7000 card, it's an interesting position how now a reference Fiji is a fail unless it outperforms a max overclocked after-market 980Ti card.

Agreed. Reference Fury X should be compared to reference 980 TI, but the most interesting comparisons will be Fury X manually OC'd vs. reasonably priced aftermarket 980 TI's that are manually OC'd. If OC'd Fury X can keep up with OC'd 980 TI's like the G1, but ends up drawing 200 more watts, then it's just another factor to consider when balancing noise levels, power draw, price, etc.

You are missing some factors - we would also need to see an after-market Fury X (will this exist?) vs. an after-market 980Ti. Secondly, you aren't at all compensating for Fury X exhausting 300W of power out of the case while running cool and quiet. An after-market 980Ti cannot claim that. This gets worse for SLI vs. CF comparisons.

I don't think there will be after market Fury X's with AIO water. How can the design be improved upon to make it worthwhile? Also, if we're talking about pluses and minuses to multi-GPU setups, then I think it's important to consider vram amounts. 6gb will be more useful at 4k in multi-GPU setups than 4gb. Other than that, I agree with the heat management issues.

Right now NV charges $750+ for EVGA 980Ti Hybrid for the same setup AMD is bringing for $650.

$100 more for the Hybrid 980 TI isn't worth the price premium vs. going with the Fury X, IMO.

 

[stuff_me_good]

I for one swore not to buy any 28nm GPU anymore, but lately I've been worrying that maybe while right now fiji seems like really late for the game but maybe TSMC and GloFo are as useless as usual and the 14-16nm finfets get postponed yet another 6 months or so. In that regard fiji launch could be spot on in the long run and that is the reason why I've been eyeing fury nano after the presentation.

 

[Azix]

I'll be trying for a 290x, then wait till 2017 or something. These prices aren't making sense right now

 

[sze5003]

I'll be trying for a 290x, then wait till 2017 or something. These prices aren't making sense right now

Depends how much you are willing to spend for all the stuff set to max without issues. For all that, prices are making sense looking at top level gear.

 

[slayernine]

I told myself a long time ago that I wouldn't ever buy a >$500 video card again. I'm still leery to buy a >$300 video card. That said I believe the Nano would be the only one I'd consider. Provided the heat/noise/power/performance is reasonable.

Even though the Fury X gives me a boner. I just can't see myself paying $650 for a video card.

In Canada it's going to end up retailing for around $800+. I think I'm going to buy one, I might hate myself for wasting money but it'll be awesome.

 

[bryanW1995]

I should buy something since my gtx 480 in backup rig just gave up the ghost, though I'm hard-pressed to justify spending even $200 on that, much less $500+. I might move the 7970 down to the backup rig and wait to see how cheap 970 or 290 gets here in a couple of weeks.

If Nvidia has to answer:

GTX 980 @ $400
GTX 980 Ti @ $550
GTX 980 Ultra @ $650 (Based on full GM200).

But, I don't see Nvidia having to respond. They got name brand and word of mouth. Fury is a new card that isn't even out yet. The only people even remotely interested are still but a fraction of buyers.

Where AMD needed to impress (R9 300 series, where MOST buyers are) completely felt unimpressive and a tad misleading during their presentation.

Nvidia has nothing to worry about.

Halo effect is kind of a big deal, you should research that some time. I do agree that the knowledgeable buyers like us are unlikely to jump onto a 380 or 390 though.

I disagree with your naming strategy, we all know that the new $650 card would be GTX 980TI +.

Nvidia would be at a bit of a spot if they drop prices right now because they just dropped prices and introduced the Titan. How upset would existing owner be if they dropped prices this soon. Nvidia should just ride it out or offer more free games.

We're still a month away from Fury Pro, and several months away from Nano. And the rest of AMD's lineup looks to be so-so at best. I think that NV will be ok if they drop prices on the top end again in, say, 6-8 weeks, especially if they do it slowly.

This is nothing like 4870 fiasco when NV slapped down huge price drops just a few days after launch...remember that NV beat the competition by 9 months this time instead of 9 days to launch.

I agree, but only if they have similar 'potential'. If you have 10% less shaders plus the Pro appears to be lower-binned vs. the X, then the $100 difference in savings might not be worth-it if you plan to OC it to max.

Can't wait to see a couple different reviews and what the X and Pro can get on the OC front. See if any patterns emerge...

Fury X will be on h20 though...should offer enormously higher OC room even if those aren't cherry-picked chips...and they should be cherry-picked to offset any perceived 9800ti advantage in OC headroom. I'd be shocked if we don't see at least 10% higher OC from Fury X, on top of the ~ 10% base higher speed, and I wouldn't be surprised to see it even more extreme than that. That water-cooled design is the one thing that might tempt me to finally get a truly high end card, even though I patently don't need it. I'm tired of hearing the fan ramp up a tiny bit when I play d3 with wizards or try to do too much at once on the screen in ciV..

 

[exar333]

Agreed. Reference Fury X should be compared to reference 980 TI, but the most interesting comparisons will be Fury X manually OC'd vs. reasonably priced aftermarket 980 TI's that are manually OC'd. If OC'd Fury X can keep up with OC'd 980 TI's like the G1, but ends up drawing 200 more watts, then it's just another factor to consider when balancing noise levels, power draw, price, etc.



I don't think there will be after market Fury X's with AIO water. How can the design be improved upon to make it worthwhile? Also, if we're talking about pluses and minuses to multi-GPU setups, then I think it's important to consider vram amounts. 6gb will be more useful at 4k in multi-GPU setups than 4gb. Other than that, I agree with the heat management issues.



$100 more for the Hybrid 980 TI isn't worth the price premium vs. going with the Fury X, IMO.

Based on some of the extreme versions we saw with the 980 and the upcoming 980Ti, there are always room for improvement on a top-tier card. I would gladly pay an extra $50 for better VRMs and components if that helps me get better or more stable overclocks.

 

[bryanW1995]

GTX 980 @ $400

They are or were down at 469$ a few days ago already.

I really hope they creep down closer to $350. That would be tough to resist...though even then I'd probably cheap out and get a 970 for $250 or whatever (and hate myself after all the 970 drama haha).

I think NV may indeed answer if the real world benchmarks indicate a need for NV to step up.

It wouldn't surprise me to find that NV has something up their sleeve.

They have nothing "up their sleeve", it's right out there for everyone to see with tons of cards available. A 9800ti with 6gb and 20% higher clocks, badged as 9800 Ultra or 9800ti + (my favorite haha). They might not have them sitting in warehouses yet, but I'd bet that within a month or 2 they could bring that card to market with enough volume to trash AMD's party even if Fury X pulls out something spectacular like "15% faster overall on 1440 and 4k with same power draw in games". Also, AMD isn't the only manufacturer who knows how to put a water block on a card.

If the FuryX OC performance is within 5%(+/-) of the 980Ti OC performance, I'll probably get the FuryX just to so I can say I have HBM in my PC. Depending upon how close the Fury OC is to the Fury X OC, I may pocket $200 instead (since I'm buying two of them).

Don't cheap out on your video cards if you're going Xfire, factory water cooling is ideal for that setup imho. Well, unless you already have your entire rig on water ofc.

AMD's releases were largely inline with the leaks. The Nano was a nice little surprise.

AMD > Nvidia in business strategy.

What? JHH is a freaking genius, and a fantastic Entrepreneur. He built NV from the ground up into the most dominant company in the industry within just a few years, and has done a great job of transitioning them over to newer industry as the market has changed. AMD has had 4 CEO's in 4 years. AMD's business strategy can be roughly summarized as "I want to pick 2 David vs Goliath battles just b/c I'm cool like that". Don't think that NV has poor business strategy just b/c it's not popular here in the forums with many posters.

 

[Headfoot]

I really hope they creep down closer to $350. That would be tough to resist...though even then I'd probably cheap out and get a 970 for $250 or whatever (and hate myself after all the 970 drama haha).

The 970 and 980 are good chips priced poorly -- the problem is that they're in the middle of the stack from a die size perspective like GTX x60 chips of old, yet they were priced as fully fledged high end cards.

At proper x60 pricing they are great chips. The GTX 980 at ~$300 (like x60 chips of yore, gtx 260, 460, 560 ti, etc.) is a great card. Its a terrible card at $500. A $250 970 would be a fantastic perf/$ card. Very easy to recommend, just like those nicely priced factory oc 460s

 

[bryanW1995]

Pretty bizarre how many are saying yes without seeing a review. Especially from a company has a rich history of over promising and then not delivering at launch everything they've promised. At least this time around AMD isn't touting any software features which inevitably won't make it into the release day drivers. How often do you see NVidia announce a date, just to announce a product without any reviews? What on earth? Obviously even after all this time, something still isn't ready for these cards, or they would have been released to the public today with reviews.

I used Ati/AMD cards exclusively for over a decade. Not any more. You've burned me enough. Until they can demonstrate an ability to deliver what they promise and release stable drivers on a consistent/frequent basis, I am steering clear of AMD.

Where have been for the past 18 months? 290x has steadily narrowed the gap and surpassed most of the NV cards due to more-consistent and steady driver improvements (and possibly just a better arch).

Reviews out tomorrow btw...hardly a huge time gap from the reveal at e3 to the reviews.

A week in advance? And, none of the companies you listed sell components. What does Google sell that anyone needs benchmarks for? Microsoft and Apple both sell products whose performance is derived from a number of different components, all produced by someone else. Both AMD and NVidia used to paper launch quite frequently, but I can't ever remember NVidia publically announcing a reveal date for a new architecture just to announce they didn't actually have anything ready for release. And even more embarrassing, AMD didn't even have press samples available for reviews. How much of an improvement is AMD going to be able to muster from the drivers in a week? Why don't major review sites have reviews up yet of a card supposedly ready for public release in a week? In recent years NVidia has been quite good at having cards ready the day they are officially announced, with reviews.

Plenty of sites have already completed sample testing...it's just that clearly-biased sites like legitreviews (who even admitted their bias) didn't get samples and threw a fit over it. I think that after seeing the fury specs, and seeing how NV has responded aggressively in the past to solid AMD launches, it's not unreasonable for AMD to hold their cards close to the vest this time.

Don't let your bitterness to AMD b/c of past experiences blind you to at least the possibility that this time they've done something right.

nvidia this, nvidia that... cut it...

Fermi... enough? Not?

a Clown with Pascal GPU with HBM without interposer made entirely out of wood.

Go back to your cave, troll.

Yes, we all remember the infamous Wooden Video Card. However, Pariah mentioned why he is skeptical of AMD's claims, and I at least have never seen him post in the VC&G forums. Plus, he's an elite member, so he probably knows a thing or 2 about tech in general and Video Cards in particular. Perhaps YOU are the one who needs to apologize here.

 

[Despoiler]

What? JHH is a freaking genius, and a fantastic Entrepreneur. He built NV from the ground up into the most dominant company in the industry within just a few years, and has done a great job of transitioning them over to newer industry as the market has changed. AMD has had 4 CEO's in 4 years. AMD's business strategy can be roughly summarized as "I want to pick 2 David vs Goliath battles just b/c I'm cool like that". Don't think that NV has poor business strategy just b/c it's not popular here in the forums with many posters.

I think you missed the context of my post. That is to say, it's a statement of this round of Fury vs Maxwell.

Also, Nvidia has been around since before DX was a thing. They are not new. JHH bet the wrong way when DX came out and almost torpedoed his company in it's infancy. He has bet the wrong way numerous times. I don't think he would still be around if he didn't own the company. Nvidia's current dominance is more due to AMD's missteps than his business acumen.

 

[blastingcap]

AMD has had 4 CEO's in 4 years. AMD's business strategy can be roughly summarized as "I want to pick 2 David vs Goliath battles just b/c I'm cool like that". Don't think that NV has poor business strategy just b/c it's not popular here in the forums with many posters.

AMD's strategy was more like

1. Let's tie together two no. 2 players in their respective industries (AMD and ATI)
2. ??? (APU???)
3. Profit!

It took AMD FOREVER to get effective APUs going, and by the time they did, Intel crushed that momentum via selling better perf/watt CPUs with good-enough iGPUs that are only getting better with time. (APUs don't make much sense for most of the desktop market; and for mobile, Intel's price/perf gives it the nod over AMD on all but the lowest-end wannabe-gaming notebooks. The higher-end gaming notebooks often use Intel CPU + NV dGPU.)

AMD made lemonade out of lemons by taking the console market with APUs but there's limited profit to be made there compared to servers and professional graphics and scientific computing/supercomputing, which are Intel/NV dominated.

I think you missed the context of my post. That is to say, it's a statement of this round of Fury vs Maxwell.

Also, Nvidia has been around since before DX was a thing. They are not new. JHH bet the wrong way when DX came out and almost torpedoed his company in it's infancy. He has bet the wrong way numerous times. I don't think he would still be around if he didn't own the company. Nvidia's current dominance is more due to AMD's missteps than his business acumen.

Um, no. JHH has a history of experimenting and seeing what succeeds; you can't just cherrypick his failures and ignore the successes. E.g., his pushing for G80 paved the way for NV's continued dominance in non-gaming graphics... he already had Quadro, now there is Tesla. 93% of NVidia employees approve of him being CEO. http://www.glassdoor.com/Reviews/NVIDIA-Reviews-E7633.htm

 

[IEC]

I would think on a perf/W basis Fury X has a real chance to beat out 980 Ti/T-X. But we'll have to wait for apples-to-apples comparisons and solid reviews to find that out.

Positives for power efficiency:
-GloFo 28nm vs TSMC 28nm
-HBM1 vs GDDR5
-AIO CLC vs air cooling (cooler GPU = less power consumption at same performance level)

Negatives:
-8.9B transistors vs 8.0B

 

[bryanW1995]

So you tell people to wait for reviews when the subject is pro AMD...


...But you're more than happy to scream that 4GB isn't enough

This forum is really something

Ha, that was very politely written. I don't remember thinking that he was so strongly in one camp a few years ago...perhaps all of the recent NV dominance has left its mark I suppose.

True, for brand agnostic/objective gamers, 980Ti OC vs. Fiji OC is a key comparison. For all those other gamers who for years ignored HD7950/7970/7970Ghz overclocking though and constantly linked us after-market 680 reviews vs. a reference HD7000 card, it's an interesting position how now a reference Fiji is a fail unless it outperforms a max overclocked after-market 980Ti card.

I still remember the feeling of "WTF has happened to AT!??" when Ryan Smith posted that gtx 460 oc review here years ago. And I loved my 460 - 768, it just felt dirty to see AT go along with such a thinly-veiled dirty trick.

 

[bryanW1995]

...dp

 

[bryanW1995]

I think you missed the context of my post. That is to say, it's a statement of this round of Fury vs Maxwell.

Also, Nvidia has been around since before DX was a thing. They are not new. JHH bet the wrong way when DX came out and almost torpedoed his company in it's infancy. He has bet the wrong way numerous times. I don't think he would still be around if he didn't own the company. Nvidia's current dominance is more due to AMD's missteps than his business acumen.

Nice attempt to spin. I'm going to take the fantastically successful business entrepreneur over a steady succession of wanna-be/can't win CEO's or a random poster on a tech forum.

And fwiw, jhh isn't the only person who bet wrong on DX...he's just the only one who was able to successfully navigate those dangerous waters and lead his company to greatness. Your claim that he wouldn't still be around if he didn't own the company is true, however...most great entrepreneurs can't coexist with a boss, they must BE the boss.

Nobody bets right all the time in business unless he's running an as-yet-undiscovered pyramid scheme...but put jhh in charge of AMD 20 years ago and the steady stream of terrible AMD ceos at NV and AMD would be the better company today instead of nearly bankrupt.

That 93% approval rating from his employees is no accident...and I hear that he's a VERY demanding boss. His employees don't necessarily LIKE him, but they certainly respect him and they doubtless like the fact that their company is still in existence all these years later.

 

[IEC]

So to bring this thread back on topic instead of pointless argument over executive leadership at graphics card companies, who else is going to be buying Fury/X?

 

[Pariah]

I've said rock solid drivers with GeForce 6, 8, HD4000, 6000 and 7000 series. You must be talking about CF situations then because AMD's drivers have mopped the floor with NV's when it comes to GCN vs. Fermi/Kepler generation. If anything, AMD's driver team has been way more rock solid than NV's in the last 3 years when it comes to single GPUs.

Last AMD card was a 7970. Wasn't a CF issue. Was a multi-monitor issue where I finally threw in the towel.

 

[xthetenth]

So to bring this thread back on topic instead of pointless argument over executive leadership at graphics card companies, who else is going to be buying Fury/X?

If it overclocks to within 10% of a 980 Ti oc on reasonable power, I am.

 

[jackstar7]

A long-term feature I'm hoping to see is mixed-rez Eyefinity. If Fury X does have that, then I'm even more sold as the idea of flanking my 3440x1440 with 2 2560x1440s makes me laugh like a merry fool.

 

[happy medium]

Reviews out tomorrow btw...hardly a huge time gap from the reveal at e3 to the reviews

really? I thought it was the 24th?

 

[sze5003]

really? I thought it was the 24th?

24th is release date.

 

[n0x1ous]

I think tomorrow is nda date for 300 series not fury

 

[guskline]

To answer the OP, ChuckFX's question, I'm fascinated by the GTX980TI, Titan X and now the AMD Fury/Fury Pro and especially Fury X. That being said I have 2 R9 290 Sapphire Tri-X gpus (1110 core/1300 mem) under water with EK blocks and use a 2560 x 1440 monitor (specs below). I really have no need to upgrade unless AMD would "abandon the 200 series in software support. Since the 300 series 390/390x are really refinined and tweeked Hawaii chips I should be in good shape for quite awhile.

I suspect I'll wait till next year when the HBM2 cards start to appear.