GeForce Titan coming end of February

[2is]

But the average person who doesn't buy a viper doesn't buy one when a new more expensive viper is released either. So.... Yeah, I don't think it's going to have the effect you're suggesting nor do I think that's nVidias goal for releasing it. It just doesn't make sense.

 

[RussianSensation]

But the average person who doesn't buy a viper doesn't buy one when a new more expensive viper is released either. So.... Yeah, I don't think it's going to have the effect you're suggesting nor do I think that's nVidias goal for releasing it. It just doesn't make sense.

The Viper's main purpose is to create excitement about the brand and other products a company sells because it showcases what the company can do in terms of tech. That's the point of Halo products like Vipers or Titan GPU. NV almost always tries to have the world's fastest GPU, through any means possible, single or dual-GPU options. They want to say they have the fastest card to make their brand seem more premium on the PC. The Titan continues on that tradition. Considering GK110 sells for $3K+ as Tesla, it wouldn't make any sense to sell it in the consumer market unless NV had extras, or they didn't pass validation for Tesla, or they were launching it for other reasons than maximizing profits (such as brand value/marketing/hype/Halo product reasons).

 

[2is]

Of course they want to create excitement, that isn't what I'm disagreeing with. I'm disagreeing with the suggestion you made that its intent might be to lure console gamers to PC.

 

[VulgarDisplay]

LOl @ the New egg $1500 price tag. I only predicted $1300.
I'm looking forward to all the howls of protest on that one from loyal NV fans.
Even better will be its frame rate scores.Luckily it won't matter if the FPS are low because "it's all about smooooothness" right?:whiste:

Howls of protest about how AMD is responsible for the price somehow....

 

[OCGuy]

But the average person who doesn't buy a viper doesn't buy one when a new more expensive viper is released either. So.... Yeah, I don't think it's going to have the effect you're suggesting nor do I think that's nVidias goal for releasing it. It just doesn't make sense.

1. I wouldn't drive something as slow as a Viper

2. Can anyone point me to this link showing $1500 as the price-tag. Newegg posting cards before even initial reviews are out would be very interesting. By the time Newegg has a card in-hand, there would be several people at EVGA that would start leaking via 3rd parties.

 

[OCGuy]

Where do you see Newegg listing this @ 1500?

I agree, and would like to see this link as well.

Edit: Well I found the link to a ludicriously priced 1500 card.

http://www.newegg.com/Product/Product.aspx?Item=N82E16814121717

 

[Ibra]

If you were NV's management and you saw that your loyal fanbase is not too dissimilar to Apple consumers, paying more and more for your product, what's the takeaway? How would you price flagship Maxwell in 2014? The market is not efficient if the consumers are irrational. In the end that only gives AMD more room to raise prices too and then the average price of GPUs continues to go up.

And who is going to buy Xbox 720/PS4 just because they are featuring AMD hardware? +900$

 

[3DVagabond]

Some think AMD should go back to price/performance strategy of HD4000-6000 series instead. If GK114 launches, and we see HD7950 at $199 and HD7970 at $279 and HD7970 GE at $329, we'll put this theory to the test. According to the same sources that started hinting that HD8000 was not launching until Q4 2013, GTX700 has also been pushed back to a similar timeframe. NV might also postpone GK114 because if you say they'll launch a 2304 SP GK114, then it'll make the Titan even more overpriced than it already looks now. Instead, NV could just continue selling GTX680s for $400-450 and Titan at $900, milking both chips with highest profit margins, without even needing GK114. Since GTX670/680s are still selling in retail for $350-550, they must be selling well enough for NV not to have to lower their prices. Then why would NV rush GK114 if they can milk old tech just as the same as AMD will be doing with HD7000?

If both companies stand pat, except for Titan, then AMD could be OK. We'll assume that since nVidia hasn't started a price war already, that they have no intention of it. If they did, I think they could really hurt AMD. Although, it would hurt them as well.

If nVidia refreshes Kepler though, they could possibly increase performance at each price point by say 15%-20% and without cutting prices flip the field on the perf/$ advantage that AMD has now. If this year has proven anything it's that AMD has to be better perf/$, and by a fairly substantial amount, or not enough people are going to buy it.

It seems though that the R&D cost and rendering a product obsolete every year isn't the best business model for either company. They need to take a page from Intel's playbook and have separate lines, maybe. Have a mainstream series and an enthusiast series that can be refreshed independently. I think Titan and GK104 could offer this. GK104 is the mainstream product and Titan could be the enthusiast product.

Short of leaving a huge wall of text outlining the entire model, let's just say that it would require a whole new level of support, not dissimilar to what they do for Quadro, just aimed at the enthusiast rather than the pro user, and designing products to not be throwaway.

 

[CakeMonster]

It seems though that the R&D cost and rendering a product obsolete every year isn't the best business model for either company. They need to take a page from Intel's playbook and have separate lines, maybe. Have a mainstream series and an enthusiast series that can be refreshed independently. I think Titan and GK104 could offer this. GK104 is the mainstream product and Titan could be the enthusiast product.

That might work. But not with the rumored Titan prices. Intel's enthusiast series does not set you back that much. You need to have a large enough customer base for the enthusiast series for it to be meaningful in the first place. If you ask me, AMD and NV are wasting too much resources on multi GPU driver tweaking already. I don't want them to spend even more resources on driver support and development of $1000 cards that even fewer people own.

 

[3DVagabond]

That might work. But not with the rumored Titan prices. Intel's enthusiast series does not set you back that much. You need to have a large enough customer base for the enthusiast series for it to be meaningful in the first place. If you ask me, AMD and NV are wasting too much resources on multi GPU driver tweaking already. I don't want them to spend even more resources on driver support and development of $1000 cards that even fewer people own.

They don't actually have to be different than what they used to sell for ~$500. Just marketed and supported differently. What they do for them will have a trickle down effect.

 

[SirPauly]

I hope $800+ is not the future for single-GPU flagships and this is more of a marketing exercise like 8800GTX Ultra was. Otherwise, we might end up at $399-449 level becoming the new "mid-range". :biggrin:

This was exactly my fear when I did see the evolutionary and incremental price performance with a substantial and significant node and arch. Understand why companies desire to maximize revenue, margin, income and premiums but as a consumer how can anyone strongly defend significant jumps in MSRP or potentially 900+ for a single GPU? Appreciate choice; as choice is always welcomed and markets set pricing to me but it looks like GeForce 8800 Ultra days once again.

When companies have competitive advantages and opportunities they show their predator fangs as the feed and devour consumers and value. Pretty tough to strongly defend this.

 

[Crap Daddy]

Well, since Nvidia has been selling, quite successful I might add, midrange chips for high-end prices who's gonna stop them selling now high-end chips for ultra high-end prices. Certainly not AMD. Take it or leave it.

 

[Jaydip]

I think people are really looking at it differently.Lets see both 6970 and 580 launched on the same process node and how much improvement did they bring? ~15%. Now this card without a cpu bottleneck will almost match a 690 on the same process node, this is phenomenal.If this is not progress I'm unsure what is.When they meant 85% of 690 I believe they implies perfect scaling and no cpu bottleneck.Like everybody else I don't like to pay more than $550 for the top end gpu so maybe I will skip and add another 680.

 

[SirPauly]

Imho,

When AMD raised their MSRP was the best thing that could of happened for nVidia, " We can charge this for the GK-104?" As they were probably dancing in the halls at nVidia headquarters while their coffers were being filled with premiums.

The polar opposite of what happened with the GTX 280, " What? AMD is charging what? Damn, we have to reduce our MSRP and stop that silly 649 MSRP premium dance!"

 

[Jaydip]

Pauly I understand what you are saying but show me 60-70% speed bump at same fabrication process.It is a great achievement for them if it is true.

 

[SirPauly]

Absolutely, the technology has been substantial and significant and even more so potentially if GK-110 is offered.

 

[Grooveriding]

I think people are really looking at it differently.Lets see both 6970 and 580 launched on the same process node and how much improvement did they bring? ~15%. Now this card without a cpu bottleneck will almost match a 690 on the same process node, this is phenomenal.If this is not progress I'm unsure what is.When they meant 85% of 690 I believe they implies perfect scaling and no cpu bottleneck.Like everybody else I don't like to pay more than $550 for the top end gpu so maybe I will skip and add another 680.

This isn't relevant, you're talking about refreshes refining the same chip where you expect a small bump. The big jumps are in moving from one process to the next. 7800GTX to 8800GTX, 8800GTX to GTX 280, GTX 285 to GTX 480 all those are the major jumps.

The GTX 680 broke the mold in the characteristics of the flagship chip and delivered half the usual performance gain we had seen in those node changes from the past. Now they are releasing GK110, a chip that fits the mold in characteristics and performance for that node movement of the past. Except they're charging $900 instead of $500.

Your expression of being impressed with it is a byproduct of nvidia conditioning the market by releasing a lacklustre flagship in the GTX 680 and looking at it as a refresh. The GK110 is not a refresh like 280 to 285 or 480 to 580, it's a completely different chip.

If the GK110 had released as GTX 680 it would of been really something as it's looking to be their most impressive flagship since the 8800GTX in terms of the performance gain over the past node's flagship (GTX 580). It's still going to be very impressive on the performance front, the price if accurate is what is nuts.

 

[Jaydip]

This isn't relevant, you're talking about refreshes refining the same chip where you expect a small bump. The big jumps are in moving from one process to the next. 7800GTX to 8800GTX, 8800GTX to GTX 280, GTX 285 to GTX 480 all those are the major jumps.

The GTX 680 broke the mold in the characteristics of the flagship chip and delivered half the usual performance gain we had seen in those node changes from the past. Now they are releasing GK110, a chip that fits the mold in characteristics and performance for that node movement of the past. Except they're charging $900 instead of $500.

Your expression of being impressed with it is a byproduct of nvidia conditioning the market by releasing a lacklustre flagship in the GTX 680 and looking at it as a refresh. The GK110 is not a refresh like 280 to 285 or 480 to 580, it's a completely different chip.

If the GK110 had released as GTX 680 it would of been really something as it's looking to be their most impressive flagship since the 8800GTX in terms of the performance gain over the past node's flagship (GTX 580). It's still going to be very impressive on the performance front, the price if accurate is what is nuts.

You raised some good points.Lets see, when 480 launched it consumed almost as much power as 5970 while being slower, 580 rectified that a bit but not much.So now comes 680 which consumes as much power as 560 Ti and how much faster it is(not checking exact figures just giving a rough estimate)? If you include power draw(which you should) 680 doesn't look that bad.Sure in some cases 680 is a disappointment compared to Fermi but NV made power draw Kepler's one of the most important metric.If you look at HPC, power consumption is a huge thing and sadly neither NV nor AMD can ignore this going forward.If Titan consumes as much as 680 but offers >=60% more performance color me impressed.

 

[Crap Daddy]

If the GK110 had released as GTX 680 it would of been really something as it's looking to be their most impressive flagship since the 8800GTX in terms of the performance gain over the past node's flagship (GTX 580). It's still going to be very impressive on the performance front, the price if accurate is what is nuts.

The best thing that happened to NV was in fact AMD's inability to produce a faster flagship card and the pricing of that card at 550$. From that point on it was clear that their midrange chip (remember the rumors about pricing GK104 cards at 300$, the true value of those cards?) can beat AMD's flagship and sell for almost double the amount it was initially thought. Now the real Kepler flagship is ready. Guess what? They are alone on the market with this so why not charge more? Will they sell out? Remains to be seen.

 

[BallaTheFeared]

If Titan consumes as much as 680 but offers >=60% more performance color me impressed.

Color me skeptical

 

[Crap Daddy]

If Titan consumes as much as 680 but offers >=60% more performance color me impressed.

It's impossible. Compare it with GF110, last node and arch and you'll be amazed.

 

[RussianSensation]

And who is going to buy Xbox 720/PS4 just because they are featuring AMD hardware? +900$

You missed all the other points I made, like NV's earning call soon. You think investors are going to be happy to hear that NV lost PS4's contract? Investors sometimes just react without understanding it might have been better for NV to not even do the contract. Here NV presents the Titan and reassures them that they are still a class leader in graphics. Also, it's not about having AMD GPU in PS4 for gamers. It's about NV trumpeting up excitement about PC gaming. You must have not watched the live video unveiling of GTX690 last year. NV made it a big event to hype up PC gaming. You think NV doesn't care at all about attracting console gamers to the PC? Titan is not going to get people to not buy a PS4 and instead get a $900 Titan. It's about showing off the capabilities of what PC gaming offers today. If someone has $500 and have a decent desktop they might decide that it's better to just buy a GTX670 and forget the PS4, especially if PS4 locks used games, has always on DRM/internet connection, etc. Whatever NV needs to do to promote the PC gaming platform, they will do it because they want those console gamers to become more aware of PC gaming.

1. I wouldn't drive something as slow as a Viper

You must drive a Corvette ZR1 or faster then? The Viper would destroy a GT500 in nearly every performance metric that actually counts.
http://fastestlaps.com/tracks/laguna_seca.html

There isn't even a point in comparing a live rear-axle GT500 with its massive curb weight against the Viper.

Well, since Nvidia has been selling, quite successful I might add, midrange chips for high-end prices who's gonna stop them selling now high-end chips for ultra high-end prices. Certainly not AMD. Take it or leave it.

Yup. It's business as usual, except it could have significant long-term consequences. Same thing happened in the audiophile market if you head over to Head-fi.org. Companies like Sennheiser said you know what we'll replace HD650s with HD800 and double the price. No problem said audiophiles. Next thing you know the average price of an awesome headphone went up 2-3x on the high-end as other companies saw that audiophiles were willing to pay 2-3x the historical pirce. What Nvidia is doing is testing how far they can raise prices. They are taking full advantage this opportunity because AMD won't really have a card to respond to the Titan. They conditioned the consumers that $1K is OK with GTX690 and they are testing this again 2nd time in a row. NV is going to milk for as much as possible to test higher and higher price levels to see just how inelastic NV's own high-end GPU consumers are in hopes of establishing new normal due to the dying sales of sub-$100 GPUs. Where do you get growth if most people are moving to tablets/laptops/smartphones? You try as hard as possible to raise prices in your other remaining product segments and see if the market will bear it. And just like this happened in the audiophile headphone market, we could just as well wake up to a mid-range 300mm2 chip for $500 and high-end 500-550mm2 Maxwell chip at $800-900 next gen.

Imho,

When AMD raised their MSRP was the best thing that could of happened for nVidia, " We can charge this for the GK-104?" As they were probably dancing in the halls at nVidia headquarters while their coffers were being filled with premiums. The polar opposite of what happened with the GTX 280, " What? AMD is charging what? Damn, we have to reduce our MSRP and stop that silly 649 MSRP premium dance!"

Pretty much. Same thing with Intel. People keep claiming Intel isn't raising prices on us. They are in a very sneaky way. Over the last 3-4 years the die sizes keep getting smaller but prices stay the same which means Intel isn't passing on that free performance to us. They are benefiting fully from the die shrinks by propping up their gross margins and all we get are reductions in power consumption with minimal performance gains; and still same # of cores from i7 920 days. By now Haswell should be a 6-core $325 part but it's not. If NV had issues with GK110 last year or not doesn't really matter in the grand scheme anymore. They got away selling 294mm2 die for $500 and it outsold the 580, while the real flagship that was supposed to replace the 580 is now rumoured to drop at $800-900.

 

[tviceman]

God I hope Valve comes out with a steam-pc. I won't care if it's Linux and won't have all the AAA titles right from the get go, I'm seriously kinda over upgrading my PC every 12-18 months. If valve makes their own hardware that A) runs PC games B) uses steam workshop for mods C) allows user to choose between kb&m or controller and D) has higher frame rates and/or better graphics than consoles, I will officially quit visiting this forum. LOL.

 

[paul878]

God I hope Valve comes out with a steam-pc. I won't care if it's Linux and won't have all the AAA titles right from the get go, I'm seriously kinda over upgrading my PC every 12-18 months. If valve makes their own hardware that A) runs PC games B) uses steam workshop for mods C) allows user to choose between kb&m or controller and D) has higher frame rates and/or better graphics than consoles, I will officially quit visiting this forum. LOL.

I wish they just add kb&m to the console, then there won't be anymore pc vs console debate.

 

[2is]

God I hope Valve comes out with a steam-pc. I won't care if it's Linux and won't have all the AAA titles right from the get go, I'm seriously kinda over upgrading my PC every 12-18 months. If valve makes their own hardware that A) runs PC games B) uses steam workshop for mods C) allows user to choose between kb&m or controller and D) has higher frame rates and/or better graphics than consoles, I will officially quit visiting this forum. LOL.

You don't need to upgrade every 12-18 months unless you're upgrading to the bare minimum all the time or you want to max out all the latest greatest games which a steam PC won't help you with.