No HBM stranglly on ti or titan. But only for better bus and more vram i will wait for ti.

TheAdAgency
Is it normal for the Titan and Ti to have lower clock speeds than their stock counterpart (in this case the 1080)?

If the bus is wider then it can still have more thruput.
 
The table in OP has been discussed in a couple other forums, and there is a lot of discussion about whether or not its a real data leaked, or made up data.
 
 

bojerka
so "ti" version be released or no ? I want "ti" but if not be released I buy normal version. I don't know if wait or no :/

All we KNOW now is that the FE 1080 will be available on the 27th and FE 1070 will be available on the 10th of June.
 
On the 27th, EVGA will be free to talk about other models of their 1080 cards.
 
There are hints that other versions will be available soon, but exactly when is not known today. However, the general belief is that a Ti version will not be available right away. Usually its 6 months or more.

I want this to be true.
 
A Ti with two power sockets, 3456 CUDA cores, that fat 384 bit mem bus, 96 ROPs...yes please!
 
With water you could blow the 1080 out of the sky overclocking that bad boy.

Yup they will need to have HBM2 and the full chip on the workstation end on the Pascal Titan for it to be a full success. 

unmeaty
I just can't bring myself to care about HBM2. I don't think it's the holy grail some people think it will be.
 
Those listed specs would make for an amazing piece of hardware with serious OC possibilities. HBM2 and a full GP100 would send the price into the stratosphere and I just don't think it would be THAT big of an improvement.
 
It would also seem to indicate that the Ti would be coming sooner rather than later.

I'm not saying the card wouldn't be awesome. I just want to buy nvidia's top end chip. GP102 isn't it. GP100 is.

make a titan x2 with gp102 base then a titan x2 extreme with hbm2 and gp100 base :)

Actually if nvidia is releasing a gp102 titan then they may be doing the original titan, 780 ti, & titan black crap all over again. 
 
1080 would be the like the original titan.
1080 ti/Titan P would be like the 780 ti.
Titan P Ultra would be the Titan Black.
 

jlp209
Definitely holding out for "bigger" Pascal and will resist all urge to get a 1070 / 1080 to tide me over, assuming I stay with Nvidia. The sole reason I'd even think about getting rid of my 980 Ti now is to limit loss of value of the card. All these card variants are ridiculous, but Nvidia is in the business of making money so I get it. Learned my lesson with the 780 and 780 Ti. It's a shame G-sync as a whole (the tech and the monitors that support it) is far more polished than FreeSync. I still have a bitter taste in my mouth from the way AMD treated the 7970m Clevo laptop debacle and have bought only Nvidia ever since.

It's based on what you want. You didn't "learn your lesson" with your 780 because you didn't know (I don't think anyone knew until they announced it). That stunt was only because Nvidia was trying to beat AMD at the time, complete accident. That's the only reason. The 1080 will be awesome for you either way, but if you decide to wait on the 1080TI that will be good too, but know in a year (or less) from the time that you buy the 1080TI, that they should then be making announcements of the release of Volta. So, it's all about what you want to get.  

yes with enuff fps and fast enuff monitors u don't even need g-sync. i'm running 3 2560 by 1440 at 144hz in sourround with my 4/way titan black hydro coppers with g-sync turned off. and all most every game I play is over 100 fps. so not needed lol.

I am a little confused about Nvidia's release cadence. Has it changed from Fermi/Kepler where they essentially had 2 generations pr architecture? Given that Maxwell only had 1 generation but was called GM20x I assumed that they skipped gen 1 for some to me unknown reason. Also, we only saw Ti versions of gen 2 cards, so GK20x and GM20x chips, which would lead me to believe that a Pascal Ti won't be around until GP20x ? Am I way of on this ?

Sajin, what if they keep the GP100 on the Tesla/Quadro cards and don't give it to the consumer level? I mean, maxwell cut the double precision on the GM100 consumer card, so what was the point? The original Titan/Titan Black had double precision, which made it the clear winner over the 780/780ti without the dp for those that wanted that feature. The Titan X had no real advantage over the 980Ti once it was released, aside from the extra vram. Double precision didn't even exist on the Titan X like it did on the original Titan and Titan black.

If they don't release a Gp100, will you just skip this generation?

I would like to upgrade soon.. I'm running two 780Ti's in SLi.  They are just not performing very well @ >1440p.. So the big question for me is a familiar one - Upgrade now, or wait for the 1080Ti?  I would be purchasing two cards.

it seems to me nVidia has the focus completely on the GTX 1080, looking at the specs.

Nephross
Has it changed from Fermi/Kepler where they essentially had 2 generations pr architecture? Given that Maxwell only had 1 generation but was called GM20x I assumed that they skipped gen 1 for some to me unknown reason.
Also, we only saw Ti versions of gen 2 cards, so GK20x and GM20x chips, which would lead me to believe that a Pascal Ti won't be around until GP20x ? Am I way of on this ?

Since Fermi, Kepler and Maxwell were all on 28nm, I tend to see the latter two as refreshes and refinements of the original.  A GM10x would be the first design, while replacing the 1 with a 2 implies a second try at the same design.  I'm thinking Maxwell was meant for a smaller process than 28nm and making the rest of Maxwell an emergency second revision, hence GM20x series codenames.  There was a GM10x, but it only showed up in mobile and low-end chips, i.e. GM107 in GTX750/750Ti; why it happened this way was due to AMD-- at the time, the GTX650/650 Ti had not be succeeded in over a year, nVidia was milking the low-end and AMD was catching up.  
 
 
It is easy to figure out the cadence by looking at nVidia's threats: AMD, on the mainstream-on-up graphics side; technically, both AMD and Intel are attacking nVidia's entry-level.  Anyway, we've known for a while (those of us that watch rumors) that only nVidia's Pascal and AMD's Polaris were planned for this year.  But then we found out that Polaris is going after nVidia's Maxwell-based mid-range on down and high-end is what the first Pascal products were going after.  This is where the fun begins.  In any business, midrange on down pays the bills, while high-end is pure profit because everything else pays for everything. But if you lose bread & butter sales, high-end starts to pay the bills, and that's bad.  You'd think the next move is to bring Pascal midrange on down, but no, partly because those Maxwell counterparts are new debuts.  The other part is that because AMD decided to react to the gap that GTX1080/1070 have with Polaris, they brought Vega closer into this year.  Vega is Polaris super-sized, it is a bigger die, it was originally planned for 2017-- guess what else was planed for 2017?  Big Pascal, which would make the next Titan and the 1080 Ti.  Rumormill is claiming it's being brought closer, as Vega is a threat to GM204.  Now it is a race for which company pushes out the true flagship first.
 
As for those of you confused why Big Pascal has no HBM2, high costs are pushing usage out to next year, this keeps both camps from using the second version too early; hence why Big Pascal for consumers is using GDDR5X.  Of course, P100 has HBM but isn't for consumers.