Titan Z announced - where are the reviews?

[blackened23]

Titan Z is primarily a compute card that happens to play games well. However, strictly for a gamer, it would not even be in the top 5 of their choice. $3K for a gaming card that cannot beat 780Ti SLI or 295X2 is absurd.

Correct. It's a CUDA developers card first and foremost, and while NV dropped the ball in marketing it partially towards gaming, the real differentiating fact of the Titan Z is the fully unlocked DP which makes it substantially faster than 780ti's for CUDA developers.

No, it isn't a good card for gaming. So while some AMD fans will be hellbent on keeping this thread bumped forever and telling us the obvious: that the card is not cost effective for gaming, this is something that we already know. If the card were "reviewed", it would tell us what we know. It isn't cost effective and would be slightly slower than 780ti sli. That gamers shouldn't really buy it unless they're a CUDA developer first that wants to fiddle with gaming on the side. Pure gamers? Skip it. We already know this basically.

I don't think anyone here ever suggested a gamer should buy this. So why hasn't it been reviewed? It would tell us what we already know. Reviewers would completely bypass the fact that it does more than gaming (ie CUDA development) and would simply offer up PC benchmarks. Is that the prime benefit of the card? No. So why send it to reviewers? Just giving us info that we would essentially know already.

Just, no joke. So stating that over and over, well, duh. It's obvious that it isn't a star card for PC gaming. Get the 295X2 instead if you want more bang for the buck in terms of gaming. This card will be bought mostly by CUDA developers as a Tesla/Quadro on a budget, since this card does 90% of what those cards do at a fraction of the cost - to that market, the Titan line has sold well.

At the end of the day, people here can complain about the price all day long but for CUDA developers they likely will not have an issue with the cost at all. Nvidia owns the professional market, and they know what that market will pay despite what anyone on a forum says. No, this card doesn't do everything Tesla or Quadro does but in terms of pure DP for CUDA development, it does all of that for a fraction of the cost of a Quadro/Tesla. But if you're a PC gamer, the obvious facts are is that it isn't cost effective and certainly is not the best buy for gaming. If you have tons of money to throw away and drive a 150k car? Well maybe get it, but everyone else should look to a different solution for gaming. The real market for the card would be CUDA development due to the unlocked DP which is not present on the 780ti. So because of that, I think it's fairly obvious why it hasn't been reviewed. The context of reviews would be PC gaming and not CUDA development.

 

[3DVagabond]

Yes, one must be an AMD fan to think this card is total fail. There are lots and lots of AMD fans it seems.

nVidia has marketed it almost exclusively as a gaming card. It doesn't offer professional support of any kind. It's not a pro card.

Tahiti has fully enabled DP just like the W9000 does. Is that supposed to be a professional card? What a steal at only $300. /sarc

 

[Grooveriding]

Card is a fail even for CUDA. You can get three Titan Blacks for the same price. Or just two that are clocked higher and cheaper. It's clear they got caught with their pants down by 295X2, it's why the marketing for the card seems so absurd now.

There is no value to be had in this card whatsoever. It's lolworthy.

 

[SlowSpyder]

Correct. It's a CUDA developers card first and foremost, and while NV dropped the ball in marketing it partially towards gaming, the real differentiating fact of the Titan Z is the fully unlocked DP which makes it substantially faster than 780ti's for CUDA developers.

No, it isn't a good card for gaming. So while some AMD fans will be hellbent on keeping this thread bumped forever and telling us the obvious

Interesting post. You call out AMD fans about this card, like they're the problem when talking about this card, like they're unrealistic with their critiques. At the same time you claim Nvidia incorrectly marketed their own card by calling this a gaming card. Seems like a lot of damage control on your part, yet you are calling out AMD fans.

I don't even like arguing about this stuff, but your post is kind of absurd.

Titan Z is definitely a gaming card plus, but it is quite obviously a gaming card first. That is what Nvidia has marketed it as. That is how Nvidia has named it. That is why Nvidia uses gaming card drivers for this card.

AMD's Tahiti @ 1GHz is capable of about 1025 GFlop DP performance. Pitcairn only achieves about 160 GFlop DP performance @ 1GHz. We agree both the 7970 and 7870 are first and foremost gaming cards, right?

 

[omeds]

Card is a fail even for CUDA. You can get three Titan Blacks for the same price. Or just two that are clocked higher and cheaper.

Yes but Titan-Z has excellent power consumption in comparison which can be important to some, running compute tasks for extended periods, perhaps in multiple systems too. It also uses a single pcie slot, which again may be attractive to some, I'm not exactly sure who but the option is there. It fills a pretty specific set of needs that some may find attractive, but I dare say that is a very small market. I wonder how many of these things they're going to produce?

 

[blackened23]

I don't even like arguing about this stuff, but your post is kind of absurd.

It is what it is. NV's marketing are idiots for marketing it towards gaming, while JHH himself dubbed the card as a supercomputer on a budget. That would have been a better narrative for marketing, but then again, NV's marketing were essentially idiots about it. The card isn't cost effective for gaming. For CUDA developers, the card will be seen as a lower cost tesla/quadro variant and a lot of CUDA developers buy it for just that. I don't think Tahiti can do CUDA. CUDA has a pretty vast ecosystem for content creation - it has been used in both entertainment and scientific applications. Does Tahiti do CUDA? I'm not sure though.

Anyway, i'm just stating the obvious. No you shouldn't buy this for gaming. You can say this until you're blue in the face and nobody would disagree with you. But the card will still sell, no matter how much it makes you grit your teeth. You can say that its a terrible value and go on and on about that, but, given NV's dominant position in all dGPU markets (and especially the professional market) i'd say they more or less know what they're doing. Their marketing might not know what they're doing, but if it's 3000$, they have a certain niche of a market that will buy it for 3000$. If you have complaints about that, i'm sure you could write them or put your resume in and tell them that you know better than they do. Apparently the entire internet has a better strategy for NV increasing their 2/3rds market share than they do. Apparently internet experts know how to design GPUs for the professional and hybrid professional/gaming markets better than they do.

Nvidia owns the professional market and it'll do just fine there. Given their market share for dGPUs in the mass market and professional markets, if they price it where they do, people will buy it at that price. Probably not gamers, but CUDA developers? Like I said, no matter how much it makes you grit your teeth i'm sure that market will buy it. Granted, nobody on this forum should buy it and no PC gamer should buy it. Not a single person in this thread ever stated otherwise to my knowledge. So the obvious again: yes, it's a terrible value for gaming. But the card will still sell nonetheless because it has more capability than just gaming, regardless of how much it may bother some. End of story.

 

[Tohtori]

Yes but Titan-Z has excellent power consumption in comparison which can be important to some, running compute tasks for extended periods, perhaps in multiple systems too. It also uses a single pcie slot, which again may be attractive to some, I'm not exactly sure who but the option is there. It fills a pretty specific set of needs that some may find attractive, but I dare say that is a very small market. I wonder how many of these things they're going to produce?

I seriously doubt anyone is going to run extended enough compute loads on a card without ECC memory for power consumption to matter at all.

 

[TrulyUncouth]

NV's marketing are idiots for marketing it towards gaming, while JHH himself dubbed the card as a supercomputer on a budget.

I only saw a few snippets of his announcement of it, but he did talk quite a bit about gaming during the announce. I don't think their marketing messed up or that JHH did either- they simply didn't see the 295x coming. If it hadn't shown up, or hadn't used water cooling they would have had the fastest card and you would be saying "Well, some people are just willing to pay for the fastest gaming card ever made- this card is not made for most of the people on this forum."

But the card will still sell, no matter how much it makes you grit your teeth. You can say that its a terrible value and go on and on about that, but, given NV's dominant position in all dGPU markets (and especially the professional market) i'd say they more or less know what they're doing. Their marketing might not know what they're doing, but if it's 3000$, they have a certain niche of a market that will buy it for 3000$. If you have complaints about that, i'm sure you could write them or put your resume in and tell them that you know better than they do. Apparently the entire internet has a better strategy for NV increasing their 2/3rds market share than they do. Apparently internet experts know how to design GPUs for the professional and hybrid professional/gaming markets better than they do.

Nvidia owns the professional market and it'll do just fine there. Given their market share for dGPUs in the mass market and professional markets, if they price it where they do, people will buy it at that price. Probably not gamers, but CUDA developers? Like I said, no matter how much it makes you grit your teeth i'm sure that market will buy it.

I think the only thing that makes anyone on here grit their teeth, and the reason so many are very happy this card failed is that prices are getting out of hand and we all want to see NV/AMD taken down a peg on pricing.

Its ridiculous as hell that we are getting cards released on a 3+ year old mature process priced considerably higher than flagships have been historically priced. It would be great if both NV/AMD had failed launches on their halo cards and realized they are gouging just a bit too much.

 

[Gikaseixas]

So while some AMD fans will be hellbent on keeping this thread bumped forever and telling us the obvious: that the card is not cost effective for gaming, this is something that we already know. If the card were "reviewed", it would tell us what we know

Us , we ... who are you refering to? Nvidia?
This tells a lot about you. It hurts you so much, seeing a thread like this going on that you had to post this attack and labeling others as fans LOL. Grow up

 

[Techhog]

It is what it is. NV's marketing are idiots for marketing it towards gaming, while JHH himself dubbed the card as a supercomputer on a budget. That would have been a better narrative for marketing, but then again, NV's marketing were essentially idiots about it. The card isn't cost effective for gaming. For CUDA developers, the card will be seen as a lower cost tesla/quadro variant and a lot of CUDA developers buy it for just that. I don't think Tahiti can do CUDA. CUDA has a pretty vast ecosystem for content creation - it has been used in both entertainment and scientific applications. Does Tahiti do CUDA? I'm not sure though.

Anyway, i'm just stating the obvious. No you shouldn't buy this for gaming. You can say this until you're blue in the face and nobody would disagree with you. But the card will still sell, no matter how much it makes you grit your teeth. You can say that its a terrible value and go on and on about that, but, given NV's dominant position in all dGPU markets (and especially the professional market) i'd say they more or less know what they're doing. Their marketing might not know what they're doing, but if it's 3000$, they have a certain niche of a market that will buy it for 3000$. If you have complaints about that, i'm sure you could write them or put your resume in and tell them that you know better than they do. Apparently the entire internet has a better strategy for NV increasing their 2/3rds market share than they do. Apparently internet experts know how to design GPUs for the professional and hybrid professional/gaming markets better than they do.

Nvidia owns the professional market and it'll do just fine there. Given their market share for dGPUs in the mass market and professional markets, if they price it where they do, people will buy it at that price. Probably not gamers, but CUDA developers? Like I said, no matter how much it makes you grit your teeth i'm sure that market will buy it. Granted, nobody on this forum should buy it and no PC gamer should buy it. Not a single person in this thread ever stated otherwise to my knowledge. So the obvious again: yes, it's a terrible value for gaming. But the card will still sell nonetheless because it has more capability than just gaming, regardless of how much it may bother some. End of story.

Still waiting on your answer for why CUDA devs would buy this instead of 2-3 Titan Blacks. If you don't have a response for that, you're arguing a moot point. Literally the only reason to buy this is if you're a CUDA developer who needs two GPUs but you don't want to buy a motherboard which can support two cards, so you choose to spend an extra $1000 to get both GPUs on a single card. Now, read what I said, and try to figure out what part of it actually makes sense.

And before you say "miniITX", show me a miniITX case with three expansion slots. Unless you can do that, then only the waterblock version makes sense, which narrows the market even more.

 

[PPB]

Still waiting on your answer for why CUDA devs would buy this instead of 2-3 Titan Blacks. If you don't have a response for that, you're arguing a moot point. Literally the only reason to buy this is if you're a CUDA developer who needs two GPUs but you don't want to buy a motherboard which can support two cards, so you choose to spend an extra $1000 to get both GPUs on a single card. Now, read what I said, and try to figure out what part of it actually makes sense.

And before you say "miniITX", show me a miniITX case with three expansion slots. Unless you can do that, then only the waterblock version makes sense, which narrows the market even more.

You are up for a long wait then. No sane person can argue about the value proposition of the Titan Z. It is laughable when you compare it as a pure gaming solution vs SLI/Xfire or the 295x, it is even more of a joke when you compare it for compute vs 3 Titan blacks (which, by the way, are way higher clocked than this card). So even with 3 Titan Blacks you will actually be getting more than just 1.5x the perf of a Titan Z on CUDA.


People in the arch viz business lately have been buying the top end Geforce SKU for their compute capabilities and putting it up for production rendering via CUDA. So there is always a market for people that cant afford a tesla or a quadro yet they still want to use GPGPU. With GK104 these people where ultimately shafted as the GK104 in CUDA was having higher rendering times than GF110. They needed the extra 0.5GB Vram for rendering more complex scenes, but werent willing to lose 15% perf over it, so they stayed with 580 3GB versions. Then Titan launched, and while it's value proposition was still bad vs multi 580's builds for rendering, there was a niche of a market for people that didnt care about perf/$ and cared for perf/watt/PCIE Slots (yeah, new metric to make a product look good when everything else fails). So there is a market for these cards.


But even if there is such market target, you have other solutions from NV that will be a lot better value over this "supercomputer on a budget", used Titans or new Titan blacks will do the job better for you than this card. That is why that not even using the GPGPU argument will save this card from disgrace. What a joke of a launch this turned out to be LOL.

 

[SlowSpyder]

It is what it is. NV's marketing are idiots for marketing it towards gaming, while JHH himself dubbed the card as a supercomputer on a budget. That would have been a better narrative for marketing, but then again, NV's marketing were essentially idiots about it. The card isn't cost effective for gaming. For CUDA developers, the card will be seen as a lower cost tesla/quadro variant and a lot of CUDA developers buy it for just that. I don't think Tahiti can do CUDA. CUDA has a pretty vast ecosystem for content creation - it has been used in both entertainment and scientific applications. Does Tahiti do CUDA? I'm not sure though.

Anyway, i'm just stating the obvious. No you shouldn't buy this for gaming. You can say this until you're blue in the face and nobody would disagree with you. But the card will still sell, no matter how much it makes you grit your teeth. You can say that its a terrible value and go on and on about that, but, given NV's dominant position in all dGPU markets (and especially the professional market) i'd say they more or less know what they're doing. Their marketing might not know what they're doing, but if it's 3000$, they have a certain niche of a market that will buy it for 3000$. If you have complaints about that, i'm sure you could write them or put your resume in and tell them that you know better than they do. Apparently the entire internet has a better strategy for NV increasing their 2/3rds market share than they do. Apparently internet experts know how to design GPUs for the professional and hybrid professional/gaming markets better than they do.

Nvidia owns the professional market and it'll do just fine there. Given their market share for dGPUs in the mass market and professional markets, if they price it where they do, people will buy it at that price. Probably not gamers, but CUDA developers? Like I said, no matter how much it makes you grit your teeth i'm sure that market will buy it. Granted, nobody on this forum should buy it and no PC gamer should buy it. Not a single person in this thread ever stated otherwise to my knowledge. So the obvious again: yes, it's a terrible value for gaming. But the card will still sell nonetheless because it has more capability than just gaming, regardless of how much it may bother some. End of story.

Dude, I have no idea what you're going on about now. Look at my post, number 55.

I think this card should be priced closer to the Radeon 295, but this card really isn't aimed for value shoppers that compare price. Nvidia will make a small number of these (relatively speaking) and sell what they do produce. I don't think this card makes any sense at $3k, but I was never Nvidia's target audience with this card.

I said that Nvidia will sell all of these they produce and that they know that. I don't know why you think I would 'grit my teeth' over that. From a business perspective this card makes sense for Nvidia, they know they have some loyal customers that will pay their premium for this part despite it's shortcomings. What makes you think I don't understand that or it bothers me?

No, the 7970 doesn't do CUDA but that is completely irrelevant to the point. The 7970 could be used as a low cost replacement for AMD professional cards, just like the Titan Z could be for Nvidia professional cards. Is the 7970 not a gaming card first and foremost, even with it's much better DP performance compared to the next gaming card range down, the 78xx? Titan Z is a gaming card with some compute abilities. The 7970 is a gaming card with some compute abilities, like mining. Even though the 7970 can be used as a low cost replacement for compute abilities or for professional AMD cards, it is still a gaming card just like Titan Z.

I don't think you realize how defensive and off base you sound (while calling out AMD fans...) about this. You just got done praising Nvidia about their dominant position in the market and their ability to produce a card slower than their competitor while being able to charge a huge premium and selling every single one, but then you call their marketing idiots. You call them idiots because they made a gaming card that you are trying so hard to convince the world isn't a gaming card because you feel you have to try and save face for Nvidia's halo product for some reason (hint: fanboyism) :whiste:

 

[blackened23]

Look, I don't disagree with a single person who states the 295X2 is a superior gaming card for the money. I'd say, the Titan Z could probably trade blows with it but it surely is not worth double the price for a gamer. But you say it's "obviously" a gaming card. The card has full double precision for CUDA developers while the 780ti SLI doesn't, so maybe it's not so "obvious" as you suggest. Yes, they partially marketed it toward gaming. But they also, for prior Titan cards in particular and at all GTC presentations have gone over the CUDA development capabilities which are a differentiator for the Titan Z. So maybe it isn't so obvious a gaming card. Because it's obvious the capabilities between the 780ti and Titan Z are different - and DP does not help gaming, but it does help for CUDA development which is used in a wide variety of medical/scientific/professional dev environments.

Basically, I just have to do a double take at people suggesting that they know more or less better than NV how to design and price their GPUs. Yes, it's a gaming ripoff. But if it's 3000$, there's a certain non gaming segment that will buy it. I sure as heck wouldn't buy it, and no gamer should buy it unless they're just filthy rich, but NV isn't exactly stupid about positioning and pricing their cards to sell everything that rolls off the assembly line. They've proven it time and time again and have the stats to back it up. So the internet experts talk about how "stupid" nvidia is about these things....so....when the internet experts chime in about this sort of thing, I have to do a double take. It isn't "obviously" just for gamers if the capabilities are more than the gamers card, the 780ti. It is obviously a ripoff for gaming, though. On that I would never disagree with anyone.

And this is essentially why it hasn't been reviewed. It would tell us what we already know: it would trade blows with the 295X2 but overall the conclusions would be that it's a rip-off at twice the price for gaming. But it does things the 295X2 can't do. So it's not "obviously" an apples to apples comparison, which is presumably why both this and the Titan Black weren't sampled to reviewers. No reviewer would benchmark the CUDA development capabilities, which are the differentiator for the Z. They would simply run games and nothing more.

 

[chimaxi83]

Basically, I just have to do a double take at people suggesting that they know more or less better than NV how to design and price their GPUs.

What a joke. You do a double take at people suggesting they know more or less better than NV on GPU design, but you yourself are suggesting you know more or less better than NV how to market their card. Uh huh.

 

[SlowSpyder]

We know you won't disagree that the Radeon 295 is better for gaming, you can't. So you're trying to convince the world that Titan Z isn't a gaming card despite what Nvidia itself says about the card, the GeForce name, and the gaming drivers it requires. But you're ignoring the fact that a pair of Titan Blacks would cost $800 less (Newegg prices) while being faster at everything Titan Z can do, like CUDA. From a consumer point of view this card really has almost no redeeming qualities, yet you keep trying to defend it (and I'd like to point out again, do so while mentiong AMD fans as somehow an issue here).

 

[frowertr]

I seriously doubt anyone is going to run extended enough compute loads on a card without ECC memory for power consumption to matter at all.

+1

If, as Blackened mentioned, the Titan-Z was really developed for the workstation market, then surely they wouldn't have overlooked slapping ECC memory on this thing just like the Tesla and Quadro models. Especially considering the price of this thing.

Anyone taking odds on when this will be priced sub-$1500 to match up with 295X2??

 

[Techhog]

Look, I don't disagree with a single person who states the 295X2 is a superior gaming card for the money. I'd say, the Titan Z could probably trade blows with it but it surely is not worth double the price for a gamer. But you say it's "obviously" a gaming card. The card has full double precision for CUDA developers while the 780ti SLI doesn't, so maybe it's not so "obvious" as you suggest. Yes, they partially marketed it toward gaming. But they also, for prior Titan cards in particular and at all GTC presentations have gone over the CUDA development capabilities which are a differentiator for the Titan Z. So maybe it isn't so obvious a gaming card. Because it's obvious the capabilities between the 780ti and Titan Z are different - and DP does not help gaming, but it does help for CUDA development which is used in a wide variety of medical/scientific/professional dev environments.

Basically, I just have to do a double take at people suggesting that they know more or less better than NV how to design and price their GPUs. Yes, it's a gaming ripoff. But if it's 3000$, there's a certain non gaming segment that will buy it. I sure as heck wouldn't buy it, and no gamer should buy it unless they're just filthy rich, but NV isn't exactly stupid about positioning and pricing their cards to sell everything that rolls off the assembly line. They've proven it time and time again and have the stats to back it up. So the internet experts talk about how "stupid" nvidia is about these things....so....when the internet experts chime in about this sort of thing, I have to do a double take. It isn't "obviously" just for gamers if the capabilities are more than the gamers card, the 780ti. It is obviously a ripoff for gaming, though. On that I would never disagree with anyone.

And this is essentially why it hasn't been reviewed. It would tell us what we already know: it would trade blows with the 295X2 but overall the conclusions would be that it's a rip-off at twice the price for gaming. But it does things the 295X2 can't do. So it's not "obviously" an apples to apples comparison, which is presumably why both this and the Titan Black weren't sampled to reviewers. No reviewer would benchmark the CUDA development capabilities, which are the differentiator for the Z. They would simply run games and nothing more.

And you're still avoiding the Titan Black SLI issue. 🙄 Look, if you don't have an arguement, just don't argue. I don't know why you feel that you have to justify this, but the fact or the matter is that you can't. If you could, you would already have an answer for why this makes more sense than 2-3 Titan Blacks for CUDA. It has no market and you know it, so unless you work for Nvidia PR I don't understand what or why you're arguing. Instead, you're just accusing everyone of looking at it as a gaming card when we are humoring you with the CUDA thing. Either come up with a counter point, or admit that you've lost.

 

[NTMBK]

Speaking as a developer who uses CUDA from time to time, the Titan Z just makes no sense. A pair of Titan Blacks is 2/3rds of the price, and offers better performance. This card has no viable market.

 

[SlickR12345]

Look, I don't disagree with a single person who states the 295X2 is a superior gaming card for the money. I'd say, the Titan Z could probably trade blows with it but it surely is not worth double the price for a gamer. But you say it's "obviously" a gaming card. The card has full double precision for CUDA developers while the 780ti SLI doesn't, so maybe it's not so "obvious" as you suggest. Yes, they partially marketed it toward gaming. But they also, for prior Titan cards in particular and at all GTC presentations have gone over the CUDA development capabilities which are a differentiator for the Titan Z. So maybe it isn't so obvious a gaming card. Because it's obvious the capabilities between the 780ti and Titan Z are different - and DP does not help gaming, but it does help for CUDA development which is used in a wide variety of medical/scientific/professional dev environments.

Basically, I just have to do a double take at people suggesting that they know more or less better than NV how to design and price their GPUs. Yes, it's a gaming ripoff. But if it's 3000$, there's a certain non gaming segment that will buy it. I sure as heck wouldn't buy it, and no gamer should buy it unless they're just filthy rich, but NV isn't exactly stupid about positioning and pricing their cards to sell everything that rolls off the assembly line. They've proven it time and time again and have the stats to back it up. So the internet experts talk about how "stupid" nvidia is about these things....so....when the internet experts chime in about this sort of thing, I have to do a double take. It isn't "obviously" just for gamers if the capabilities are more than the gamers card, the 780ti. It is obviously a ripoff for gaming, though. On that I would never disagree with anyone.

And this is essentially why it hasn't been reviewed. It would tell us what we already know: it would trade blows with the 295X2 but overall the conclusions would be that it's a rip-off at twice the price for gaming. But it does things the 295X2 can't do. So it's not "obviously" an apples to apples comparison, which is presumably why both this and the Titan Black weren't sampled to reviewers. No reviewer would benchmark the CUDA development capabilities, which are the differentiator for the Z. They would simply run games and nothing more.

So you know more than Nvidia chief, all their developers, all their marketing team? I mean are you god, or maybe demigod to know so much?

All of Nvidia Loudly and aggressively and strongly advertised this card as a GAMING CARD at every opportunity for over 2 months because its meant for gaming, not anything else.

Once they got caught with their pants down when AMD released the 295x they stopped ALL talk about the Titan Z and didn't deny even ONCE that this card was somehow magically after 3 months of gaming card advertisement that it all of a sudden was a cuda developers card, not ONCE!!!!!!!!!!!

So it was meant for GAMING, all of Nvidia were saying that for over 3 months until AMD ruined their parade and released 5x better card at half the price and destroyed Nvidia.

You and others like you are suffering from Stockholm syndrome and learned helplessness, so you are literally victims to Nvidia.

 

[rtsurfer]

Yes but Titan-Z has excellent power consumption in comparison which can be important to some, running compute tasks for extended periods, perhaps in multiple systems too. It also uses a single pcie slot, which again may be attractive to some, I'm not exactly sure who but the option is there. It fills a pretty specific set of needs that some may find attractive, but I dare say that is a very small market. I wonder how many of these things they're going to produce?

You know that with the $1000 difference in price, I could run the TitanZ till the time that it will be slower than my phone GPU & still wouldn't make up the $1000 in electricity costs as savings from SLI.

Try again.

 

[3DVagabond]

+1

If, as Blackened mentioned, the Titan-Z was really developed for the workstation market, then surely they wouldn't have overlooked slapping ECC memory on this thing just like the Tesla and Quadro models. Especially considering the price of this thing.

Anyone taking odds on when this will be priced sub-$1500 to match up with 295X2??

It's obviously a gaming card, as you point out. I personally don't think it will ever be $1500. I think they will sell the ones they have and as quietly as they introduced it, they'll EOL it.

 

[NTMBK]

It's obviously a gaming card, as you point out. I personally don't think it will ever be $1500. I think they will sell the ones they have and as quietly as they introduced it, they'll EOL it.

There may be some value in this as a collector's piece, if it's going to be quite that scarce!

 

[Spanners]

Snip...

This is your take on it, but as you've said this certainly isn't the way Nvidia have marketed the card. If this was being marketed as a CUDA card then I'd agree that the gaming performance/price argument would be a waste of time and review space. It is however being sold as a gaming card so I think it's more than fair to assess/review it as such.

There is also a pretty strong argument to be made that there are better choices for CUDA at the same or lower price points.

"GeForce GTX TITAN Z is a gaming monster, built to power the most extreme gaming rigs on the planet. With a massive 5760 cores, 12GB of 7GBps GDDR5 memory, and the most advanced power delivery system, GTX TITAN Z offers truly mind-blowing performance. It's easily the fastest graphics card we've ever made."

 

[3DVagabond]

I wouldn't be surprised if and when we do see reviews it will be of this one.

With a full cover water block it will be faster than the 295x2. (I would hope so anyway)

 

[TreVader]

I wouldn't be surprised if and when we do see reviews it will be of this one.

With a full cover water block it will be faster than the 295x2. (I would hope so anyway)

Not unless they clock it well above 1150 core stock

It will probably be competitive in synthetics and Gameworks stuff but it won't "beat" the 295 by any significant margin. Not without absurd clocks.

The whole vapor launch/bad press/295x2 release/more bad press has really killed this as ever being a well reviewed product IMO.