New Pascal Titan X!

[n0x1ous]

I don't think anybody buying this really needs a review to make the purchase decision.

Probably not, but we still want to know in depth GP102 info and a professional review. I didn't say that as a potential purchaser just a GPU enthusiast. Weird comment bro.

 

[BFG10K]

First review: http://www.gamestar.de/hardware/gra...news-artikel/nvidia_titan_x,1010,3276621.html

 

[Arachnotronic]

First review: http://www.gamestar.de/hardware/gra...news-artikel/nvidia_titan_x,1010,3276621.html

25%-35% faster than 1080, right about in line with what you'd expect. With a reasonable overclock, this looks like a 4K/60FPS card.

Going to need much more grunt in the future to handle 4K/144Hz when those monitors become available, though!

 

[Timmah!]

1349 EUROs...oh well

 

[pj-]

Bought one to flip but it looks like they're not going out of stock.

Oh well, guess I'll have to return it and eat the shipping cost. Or keep it...

 

[Arachnotronic]

Bought one to flip but it looks like they're not going out of stock.

Oh well, guess I'll have to return it and eat the shipping cost. Or keep it...

Can't you cancel the order?

 

[tential]

I can't think of the last time a halo card was ever good in value. It will be interesting to see with the increase in bandwidth/px fillrate etc how this stacks up against a 1080 at 4k res.

980ti was the best high end value card in a long time. It's performance was justified at 650

 

[Arachnotronic]

980ti was the best high end value card in a long time. It's performance was justified at 650

Agreed, 980 Ti was an excellent value, much better than the 980 or the 780 Ti at launch.

 

[pj-]

Can't you cancel the order?

Nope!

Digital River sucks

 

[FatherMurphy]

The lack of coverage and reviews on "launch" day coupled with the unusual announcement at an academic gathering two weeks ago..... very strange for this uber high-end part. If you are Nvidia and you want to keep polishing that halo product, why the lack of marketing for Titan XP?

 

[Beavermatic]

The lack of coverage and reviews on "launch" day coupled with the unusual announcement at an academic gathering two weeks ago..... very strange for this uber high-end part. If you are Nvidia and you want to keep polishing that halo product, why the lack of marketing for Titan XP?

You guys keep talking like this is the "big pascal" titan.

It's just a titan x refresh. There's no doubt in my mind that in about 6 months, we're going to see the actual "big pascal" titan and 1080ti, likely with full 4000 Cuda cores and some hbm2 sweetness... likely just a gp100 gaming variant.

The fact the new Titan X is receiving little to no marketing attention or hype from Nvidia and was released so quickly against 1080 pretty much gurantees it. Just a minor product update to fill a gap.

 

[FatherMurphy]

You guys keep talking like this is the "big pascal" titan.

It's just a titan x refresh. There's no doubt in my mind that in about 6 months, we're going to see the actual "big pascal" titan and 1080ti, likely with full 4000 Cuda cores and some hbm2 sweetness... likely just a gp100 gaming variant.

The fact the new Titan X is receiving little to no marketing attention or hype from Nvidia and was released so quickly against 1080 pretty much gurantees it. Just a minor product update to fill a gap.

It's not a simple refresh (e.g. 680 to 770). It's an entirely new, relatively large chip. More importantly, this product further extends Nvidia's performance lead over AMD. I don't understand why Nvidia isn't sending it through the usual launch day marketing process (press release, reviews, etc.).

 

[tg2708]

Nvidia needs to give AMD time to breath. Geez. Solid performance lead over the 1080 though.

 

[sxr7171]

Anyone got their confirmation mails yet?

I did. Immediately after the order.

 

[Beavermatic]

It's not a simple refresh (e.g. 680 to 770). It's an entirely new, relatively large chip. More importantly, this product further extends Nvidia's performance lead over AMD. I don't understand why Nvidia isn't sending it through the usual launch day marketing process (press release, reviews, etc.).

I understand. I had maxwell titan x's, prior to my pascal 1080s. I know the titan x's I just ordered are also based on that same pascal architecture as the 1080s.

But there's a much, much more ferocious version of pascal that released ahead of either the gp104 1080 or new gp102 titan x.... the gp100 tesla.

Obviously, the tesla is a hpc card... thus never intended directly for gaming nor priced for such. But it has all the potential to do so... and I figure they'll release a true Titan X successor based closer to that gp100 in the coming months, and the current gp102 is just to hold over the "where's our big pascal?!" whiners.

 

[Det0x]

The lack of coverage and reviews on "launch" day coupled with the unusual announcement at an academic gathering two weeks ago..... very strange for this uber high-end part. If you are Nvidia and you want to keep polishing that halo product, why the lack of marketing for Titan XP?

Could have something to do with this ?

SK Hynix Adds HBM2 to Catalog: 4 GB Stacks Set to Be Available in Q3

http://www.anandtech.com/show/10527/sk-hynix-adds-hbm2-4-gb-memory-q3

Maybe Vega is incoming sooner then expected, and NV is trying to saturate the high high-end market before it launches..

I mean that october rumor could turn out to be true, although i doubt it.

Bucket load of salt needed

AMD Pulls Radeon "Vega" Launch to October

https://www.techpowerup.com/222403/amd-pulls-radeon-vega-launch-to-october

AMD allegedly pushes VEGA GPU forward to October

http://videocardz.com/59808/amd-vega-gpu-allegedly-pushed-forward-to-october

Report suggests AMD has pulled 'Vega' GPU launch forward to October

http://www.techspot.com/news/64797-report-suggests-amd-has-pulled-vega-gpu-launch.html

After Vega launches Nvidia could repeat history, with their own launch of the full die gp102 to combat the HBM2 flagship, making some of the very same users in this thread upgrade again :thumbsup:

Profit for NV

 

[FatherMurphy]

How does a fully functional GP100 have the potential to exceed the gaming performance of a fully functional GP102? From what I can tell, they have the same amount of FP32 CUDA Cores. The extra HPC hardware in GP100 (the NVlink bits, the dedicated DP cores) bloats the die and potentially eats up the power budget. Surely, in gaming, the GP102 will run at higher clocks and be more than competitive than the GP100?

I understand that the GP100 has the HBM advantage, but I strongly doubt that the GP102 is so bandwidth starved that the bandwidth boost from HBM is going to make an incredible difference.

EDIT: @ Detox

If AMD has an incoming Vega chip, all the more reason to market the Titan XP.

 

[Det0x]

How does a fully functional GP100 have the potential to exceed the gaming performance of a fully functional GP102? From what I can tell, they have the same amount of FP32 CUDA Cores. The extra HPC hardware in GP100 (the NVlink bits, the dedicated DP cores) bloats the die and potentially eats up the power budget. Surely, in gaming, the GP102 will run at higher clocks and be more than competitive than the GP100?

I understand that the GP100 has the HBM advantage, but I strongly doubt that the GP102 is so bandwidth starved that the bandwidth boost from HBM is going to make an incredible difference.

EDIT: @ Detox

If AMD has an incoming Vega chip, all the more reason to market the Titan XP.

The current Pascal Titan X is a cutdown version.. Maybe the full die (3840 CC + faster GDDR5X) is needed to combat Vega.

Its one of the few things "both sides" agrees on.. NV is selling a cut Titan for 1200$, and people are buying it

3584 CUDA cores, eh? No way this is full GP102. Bet that fully enabled dies with 3840 CUDA cores (1.5x GP104, just like GM200 was 1.5x GM204) will arrive later with faster GDDR5X.

$1200, just like many of us predicted that NV will once again raise prices.
3584 Cores is a repeat of the OG Titan where NV released a cut-down flagship and only later followed up with the full Titan Black. Let the milking begin!
...$1500-2000 Titan X Black 3840 CC slated for 2017.

NV is a genius at marketing. I am expecting 1080 / 1080 SLI owners to be putting up their cards on eBay and upgrading to this thing. Well played.

You can bet "Pascal Titan X Black" will get much more press and coverage, its needed to fight Vega after all.. (in cherry picked dx11 games from 2014 if the reviewers wish to receive one )
And NV have made ~twice the profit :thumbsup: (even some diehard fans will have upgraded their graphic card 3 times in a span of ~6 months)

Good for the shareholders, not so much for green users.

 

[Beavermatic]

The current Pascal Titan X is a cutdown version.. Maybe the full die (3840 CC) is needed to combat Vega.

You can bet it will get much more press and coverage.. And NV have made ~twice the profit :thumbsup:

Good for the shareholders, not so much for the users.

For some reason, I keep thinking full die is 4096 CC's, or is that incorrect?

 

[n0x1ous]

For some reason, I keep thinking full die is 4096 CC's, or is that incorrect?

3840 is the full die based on the released info on the Quadro P6000

 

[xpea]

Titan X Pascal in SLI review:
http://www.pcworld.com/article/3102...ew-titan-x-is-absolutely-decadant-in-sli.html

edit: beautiful pcb on this bad ass (more VRMs than 1080 FE)

 

[2is]

How does a fully functional GP100 have the potential to exceed the gaming performance of a fully functional GP102? From what I can tell, they have the same amount of FP32 CUDA Cores. The extra HPC hardware in GP100 (the NVlink bits, the dedicated DP cores) bloats the die and potentially eats up the power budget. Surely, in gaming, the GP102 will run at higher clocks and be more than competitive than the GP100?

I understand that the GP100 has the HBM advantage, but I strongly doubt that the GP102 is so bandwidth starved that the bandwidth boost from HBM is going to make an incredible difference.

EDIT: @ Detox

If AMD has an incoming Vega chip, all the more reason to market the Titan XP.

The power savings from HBM2 will more then offset the additional power consumption of the larger die IMO.

 

[alcoholbob]

Is it just me or does the buy it now button not work on nvidias website? Doesnt seem to work on my cell phone or 2 different pcs.

 

[FatherMurphy]

The power savings from HBM2 will more then offset the additional power consumption of the larger die IMO.

Assuming that is true, you are still looking at a much larger die in GP100 plus expensive HBM, but not really any more gaming performance, at least from what we know and can infer. We are confident that the two dies have the same number of FP32 ALUs. Now, we do not know (as far as I am aware) what kind of fixed function graphics hardware (TMUs, ROPs) the GP100 has. However, given its primary purpose as an HPC chip, I strongly doubt that the GP100 has any advantage over GP102 in this area.

So, without the promise of more clock speed or ALUs, the only advantage from a gaming perspective that GP100 has over GP102 is HBM2. For that to significantly matter in gaming, you would have to assume that GP102 is horrendously bandwidth starved even with GDDR5X on a 384-bit bus. I really doubt that.

So, again, from strictly a gaming perspective, I fail to see how Nvidia with the GP100 is going to set the bar for performance much higher than a fully functional GP102. Which brings me back to my original perplexity over Nvidia's (lack of) marketing of this chip. I do not believe we'll see a GP100 consumer chip, it just doesn't get Nvidia anything in the gaming sector. So, I think the best we'll see of the Pascal generation is a fully enabled GP102, which is not going to be much better than the cripped GP102 in Titan XP. It seems far-fetched to think Nvidia is going to offer in just a few short months an even more titan Titan (the Titanest of them all) which may offer, what, 15 or 20% more performance tops, at, what, $1500?

 

[Det0x]

@ FatherMurphy

Think its only you thats talking about the GP100 here.. My bet is that we will never see it(gp100) on a consumer based gfx card.
Read my previous updated post.