Rumor mill amends Titan release date to 4Q16/1Q17

E

evilsofa

[H]F Junkie
Joined
Jan 1, 2007
Messages
10,078
April 2016 - GTC: Tesla (GP100)
June 2016 - Computex: GTX 1080 (GP104) & GTX 1070 (GP104)
4Q 2016 - GP106 & GP107
4Q 2016 or 1Q 2017 - Titan (GP100)
1Q / 2Q 2017 - GP108 (GM108 rebrand?)

Bits And Chips rumor article, Google Translated to English

The June release date for the 1080 and 1070 is not new; that's been fairly consistent for awhile now. That the Titan will not be coming out in April is in conflict with previous rumors but perhaps we knew better than to actually expect that, when we know what Titan GPUs are - what's left of GP100 after demand for much higher priced Tesla and Quadro is sated.

I'll remind you all that Tesla cards are for server farms, not gaming cards, and even if they were nice enough to include a DVI port, a likely 5-digit price tag will keep it out of your hands anyway.

Due to what I posted in another thread, I speculate that if the next Titan is delayed that far, it will quite possibly be a dual GPU card like the Titan Z, because Pascal's greatly reduced thermals would be great for dual-GPU cards.
 
Not till September launch more likely. GDDR5X is not ramping production till June.
We plan to be in full volume production this summer.
Click to expand...
https://www.micron.com/about/blogs/2016/february/gddr5x-has-arrived

HBM2 will be too expensive for the 1070/80. I would bet HBM2 would be on the Next Pascal Titan/Ti Version. I would fully expect 1070/80 with Gddr5 at launch in September, like they did with the 900 series followed by March 2017 Titan Pascal and May launch of 1080Ti HBM2. It just fits with how/when Nvidia launches. Just IMO of course.
 
The "Ti" version of Pascal will definitely be my next upgrade (along with jumping to a 4K monitor), but I am in no rush. If it's still a full year out, that's just fine. Very happy with the 980Ti SLI/30" LCD setup as it easily handles everything I throw at it. Gaming @ 2560x1600 just doesn't push the cards all that hard... but I can see 4K being a bit more of a challenge that the "Ti" version of Pascal should easily handle.
 
This stuff is getting to be very boring.

Technology is moving at a snail's pace.
 
sounds good.
2560x1400 96+hz doesn't work on a 780 anymore. I'd love a june release for a "1080". Would be extremely disappointed if its closer to September (start of new semester)
 
Shit I just upgraded to a 660ti and 1440p monitor.... I'm way behind. Although I love the rumor mill
 
SnowBeast said:
Not till September launch more likely. GDDR5X is not ramping production till June. https://www.micron.com/about/blogs/2016/february/gddr5x-has-arrived

HBM2 will be too expensive for the 1070/80. I would bet HBM2 would be on the Next Pascal Titan/Ti Version. I would fully expect 1070/80 with Gddr5 at launch in September, like they did with the 900 series followed by March 2017 Titan Pascal and May launch of 1080Ti HBM2. It just fits with how/when Nvidia launches. Just IMO of course.
Click to expand...

I think this is pretty much spot on.

Unless AMD force otherwise, it's in nvidia's best interest to delay and milk Pascal for as long as possible.
 
jwcalla said:
This stuff is getting to be very boring.

Technology is moving at a snail's pace.
Click to expand...

Next year my motherboard will be 6 years old and still running 24/7. I thought AMD and Intel wanted to sell me a new pony. Intel did say that they will sell me a slower pony next year that is more energy efficient. That way I can knock at least $1 off my $150 electricity bill for my all electric home.

Darn exciting times that were living in as far as new PCs goes.
 
SnowBeast said:
Not till September launch more likely. GDDR5X is not ramping production till June. https://www.micron.com/about/blogs/2016/february/gddr5x-has-arrived

HBM2 will be too expensive for the 1070/80. I would bet HBM2 would be on the Next Pascal Titan/Ti Version. I would fully expect 1070/80 with Gddr5 at launch in September, like they did with the 900 series followed by March 2017 Titan Pascal and May launch of 1080Ti HBM2. It just fits with how/when Nvidia launches. Just IMO of course.
Click to expand...

I don't see GTX Pascal parts launching without GDDR5X, aside from maybe GP107 @ 7ghz GDDR5 and obviously GP100 with HBM2. Pascal likely won't have the same benefit of large performance gains at the same bandwidth that Maxwell had over Kepler. There is only so much to wring out in the way of texture compression and bandwidth efficiency, and 3 generations in a row with the same memory bandwidth (going back to Kepler) will create a huge bottleneck in GPU's performance. It will be better for Nvidia to launch at higher prices with GDDR5X and bring prices down as memory gets cheaper than to launch poorly performing next generation cards and get a bad rap from the start.
 
Last edited:
My prediction: first GP100 card will almost assuredly be a Tesla part, and Titan X successor with HBM2 won't land until Q4 2016 at the very very earliest.
 
n=1 said:
My prediction: first GP100 card will almost assuredly be a Tesla part, and Titan X successor with HBM2 won't land until Q4 2016 at the very very earliest.
Click to expand...

This is how I'm leaning as well, with Q4 still being VERY unlikely.
 
I think we're looking at 2017 for sure until the GTX 1080 Ti MSI Lightning Edition :(
 
Half expected, I thought there was something wrong with Titan V being released so soon, and so far away from Ti.
 
n=1 said:
My prediction: first GP100 card will almost assuredly be a Tesla part, and Titan X successor with HBM2 won't land until Q4 2016 at the very very earliest.
Click to expand...
There is definitely again an old problem for NV how to keep Tesla and top Titan ranges separate and discourage professionals from using Titan instead of much more expensive Tesla. Separation in time between launches is temporary solution. On the hardware side Tesla will support NV link which Titan may not. Could it also be that only Tesla has chip with HBM memory while Titan, at least in the beginning, would come only with standard VRAM?
 
wirk said:
There is definitely again an old problem for NV how to keep Tesla and top Titan ranges separate and discourage professionals from using Titan instead of much more expensive Tesla. Separation in time between launches is temporary solution. On the hardware side Tesla will support NV link which Titan may not. Could it also be that only Tesla has chip with HBM memory while Titan, at least in the beginning, would come only with standard VRAM?
Click to expand...

I can't see nvidia releasing a Titan part without HBM, given the competition already beat them to it last year.
 
wirk said:
There is definitely again an old problem for NV how to keep Tesla and top Titan ranges separate and discourage professionals from using Titan instead of much more expensive Tesla. Separation in time between launches is temporary solution. On the hardware side Tesla will support NV link which Titan may not. Could it also be that only Tesla has chip with HBM memory while Titan, at least in the beginning, would come only with standard VRAM?
Click to expand...

HPC Cluster Blog - GTX vs Tesla » Advanced Clustering Technologies

But all in all, the Titan sits comfortably in the middle of the specs of the various Tesla cards. So, our customers often ask, why purchase a Tesla at all? Why not just use one of these high-end GTX Titan cards for all cluster development and deployment needs?

The answer to this question is threefold. First, and most importantly, the high-end GTX GPUs like the Tesla do not use ECC (error checking and correction) memory. ECC memory includes extra memory bits designed to detect and fix memory errors, which is of paramount importance to the successful completion of high performance, double-precision code. ECC memory ensures that the results of computations run on a Tesla are the same every time; the same tasks run on a high-end GTX card like the Titan can vary from job to job. Clearly, for scientific computing, the Tesla offers the best consistency.

Secondly, since Titans are not designed for constant high performance computing in a dense environment, we’ve seen a much shorter longevity with these cards when used for this purpose as compared to the Teslas. One of the reasons for this is that Titans use active cooling (with fans) as opposed to the passively cooled (no fans) Teslas; when installed in a standard rack-mount server case, the small space causes the Titan’s fans to overwork, thus increasing the likelihood of their failure and the subsequent overheating of the Titan card. If several Titans fail, then the overall cost paid is equivalent to simply purchasing a Tesla at the outset. Additionally, NVIDIA provides full support, bug fixes and feature requests for high performance computing with all varieties of Tesla, whereas the Titan cards are supported only by their third-party manufacturers.

Lastly, Tesla cards are optimized for cluster usage, including full support for InfiniBand and RDMA (remote direct memory access) to allow for high-throughput and low-latency inter-node communication. They also include built-in tools for GPU and cluster management. If a developer’s intent is to use CUDA to design a program for use in a cluster, a Tesla will best support this effort.
Click to expand...

Those who are serious about compute (ie livelihood depended on it) would never consider a Titan anyway.
 
After they lied about the Drive PX2 (held up a 980M based board) and Semi-Accurate debunked the Zauba shipping manifest leak, I'd be surprised if tney release anything by April

(So no link to that article, on tablet atm)
 
I think they will stick with their schedule as they have the last few years: Release the highend workstation card first, then release teh "Titan" and go from there. May/June release.
 
n=1 said:
HPC Cluster Blog - GTX vs Tesla » Advanced Clustering Technologies
Those who are serious about compute (ie livelihood depended on it) would never consider a Titan anyway.
Click to expand...

Time and again it has been proved that Teslas/Quadro and Titans are of no real difference for workstation type of applications. NVidia tried to discourage professionals from Titans by saying that Tesla has drivers optimized for high-end professional software and offered better support which has some truth but in the end those who buy Teslas are organizations for which 1000 or 4000$ added hardware cost per seat makes no real difference. Now for the server/cluster HPC with tons of cards rthe Tesla is better, who will put hundreds of individually air cooled cards into 24/7 operation (though servers with 8 air cooled Titans are readily available) and ECC RAM may have some significance too. But here we are talking more about workstation type of things.
 
jwcalla said:
This stuff is getting to be very boring.

Technology is moving at a snail's pace.
Click to expand...

Agreed. No more competition in these sectors. Both Intel and Nvidia can hang out and do nothing because nobody else is driving them. This 6700K machine I built is hardly any faster at all than my 2600K both at 4.6 in anything other than benchmarks and FSX. I built it for FSX and it is wayyy faster than the 2600K at that but that's really the only thing.
 
wirk said:
Time and again it has been proved that Teslas/Quadro and Titans are of no real difference for workstation type of applications. NVidia tried to discourage professionals from Titans by saying that Tesla has drivers optimized for high-end professional software and offered better support which has some truth but in the end those who buy Teslas are organizations for which 1000 or 4000$ added hardware cost per seat makes no real difference. Now for the server/cluster HPC with tons of cards rthe Tesla is better, who will put hundreds of individually air cooled cards into 24/7 operation (though servers with 8 air cooled Titans are readily available) and ECC RAM may have some significance too. But here we are talking more about workstation type of things.
Click to expand...

Well you said "professionals", which is kinda ambiguous.

In any case, organizations who buy Tesla cards are not suddenly going to cheap out and buy a Titan, which at the end of the day is still a GeForce part no matter how nVidia markets it. Quadro/Tesla cards are very expensive mainly because of the excellent support nVidia provides for them, in addition to much better build quality. So as I said earlier, those whose actually need a Tesla will continue to buy Tesla cards and won't bat an eyelash at the Titan. Those who would consider a Titan likely never would've bought a Tesla in the first place anyway.
 
cageymaru said:
Darn exciting times that were living in as far as new PCs goes.
Click to expand...

I haven't gone this long without feeling like there were any genuinely compelling cpu/gpu upgrades since I started building my own PCs in the 90's.
 
Coming from a 7970 Lightning last year to a 980ti has given me that anxiety again also. Whilst i agree with sustainability in the market, this is fucking ludicrous. Wheres my 10^2 million MIPS CPU XD
 
I'm starting to feel how squints felt before he kissed Wendy Peppercorn in The Sandlot.
 
You must log in or register to reply here.
Back
Top