Jump to content

[Rumor] GTX Titan II

jmaster299

Nvidia working on the next version of the GTX Titan, right now being called the Titan II. It's rumored to have more than 4000 Cuda Cores with a 512-bit memory bus. Yes I know people hate on the Titan because they don't understand the limited/niche market for it. But if Nvidia repeats what they did with the 780 Ti, we will get a more consumer friendly version of it. The 780 Ti and Titan Black pack 2880 Cuda Cores, so if the "over 4000 Cuda Cores" rumor stays true, that's at least a 39% increase in processing cores.

Quote:

"An Nvidia product going by the description of a “GM200 A1 Graphics Processor” has been spotted in shipment from Taiwan to Bangalore where it will arrive at an Nvidia facility for further testing. The A1 stepping signifies the pre-production status of the chip and it will be upgraded to an A2 stepping before being pushed into mass production for the consumer market. According to the source the GM200/GK210 codename GPU will be built on the 28nm fabrication process using Maxwell architecture. GM200/GK210 is rumoured to feature over 4000 CUDA cores and a widened memory bus of 512 bits. Given the size of the 28nm process this new GPU will be very larger, over 600 mm². Nvidia will have to rely on the efficiency gains of the Maxwell architecture in order to keep such a large chip running cool. The GM200/GK210 GPU will form the next generation of Titan flagship, for now called the GTX Titan II. Expect the launch to be in the first half of 2015."

http://www.eteknix.com/nvidia-preparing-gtx-titan-ii-over-4000-cuda-cores-and-512-bit-bus/

i7 2600K @ 4.7GHz/ASUS P8Z68-V Pro/Corsair Vengeance LP 2x4GB @ 1600MHz/EVGA GTX 670 FTW SIG 2/Cooler Master HAF-X

 

http://www.speedtest.net/my-result/3591491194

Link to comment
Share on other sites

Link to post
Share on other sites

GTX Titan: Return of the cuda cores

"How hard can it be?" - Jeremy Clarkson

"Speed has never killed anyone. Suddenly becoming stationary, that's what gets you." - Jeremy Clarkson

 

"There's an old saying that God exists in your search for him. I just want you to understand that I ain't looking." - Leslie Nielsen

 

Link to comment
Share on other sites

Link to post
Share on other sites

is the website trusted? well if its true then idk fuck it its not like il buy it lol.

My PC

[ I5 4690k (no oc) - Gigabyte Z97 D3H - 8GB Ram - Sapphire R9 280X Vapor-X ]

Link to comment
Share on other sites

Link to post
Share on other sites

because you know, we totally need more gpus to spend money on............ because its probably gotnna be like $1500, so why not get a Titan Z for double that and then u also get 2 gpus....

My Car: http://linustechtips.com/main/topic/274320-the-long-awaited-car-thread/?p=4442206


CPU: i5 4590 |Motherboard: ASRock H97M PRO4|Memory: Corsair Vengance 8gbs|Storage: WD Caviar Blue 1TB|GPU: ZOTAC GTX 760 2gb|PSU: Thermaltech TR2 500W|Monitors: LG24M35 24" & Dual 19"|Mouse:Razer DeathAdder 2013 with SteelSeries Qck mini|Keyboard: Ducky DK2087 Zero MX Red|Headset: HyperX Cloud|Cooling: Corsair 120mm blue LED, Lepa vortex 120mm, stock 120mm|Case:Enermax Ostrog Blue Windowed


 

Link to comment
Share on other sites

Link to post
Share on other sites

 

 

Nvidia will have to rely on the efficiency gains of the Maxwell architecture in order to keep such a large chip running cool. The GM200/GK210 GPU will form the next generation of Titan flagship

Well if it's GK210 it isn't Maxwell surely?

 

GF110 = Fermi

GK110 = Kepler

GM107 = Maxwell (750ti)

GK210 = Maxwell...? According to article.

 

So I'd assume it'd be GM210 to make sense. (I believe the 'big' silicone GPUs with large die size are usualy 110s).

Link to comment
Share on other sites

Link to post
Share on other sites

Well if it's GK210 it isn't Maxwell surely?

 

GF110 = Fermi

GK110 = Kepler

GK210 = Maxwell...?

 

So I'd assume it'd be GM210 to make sense.

Yes it is because Maxwell has been changed due to the delay at TSMC with the production of 20nm chips. We already the 750 Ti that uses a modified 28nm Kepler chip under the Maxwell "GM" name. Rumor is that because of the delay, which has since been fixed, Nvidia is going to skip 20nm for their GPUs. Supposedly we will get redesigned 28nm chips under the 800 series and Maxwell name and in a year or less time we Nvidia will jump to 16nm chips. Possibly keeping the 800 series name plate with some added identifier to separate them from the cards with 28nm chips.

The rumor that Nvidia is skipping the 20nm chips is also supported by the fact that Apple is now being given first priority for 20nm chips for their next set of iPhones and iPads coming in September. TSMC has already stated delivering those chips to Apple.

i7 2600K @ 4.7GHz/ASUS P8Z68-V Pro/Corsair Vengeance LP 2x4GB @ 1600MHz/EVGA GTX 670 FTW SIG 2/Cooler Master HAF-X

 

http://www.speedtest.net/my-result/3591491194

Link to comment
Share on other sites

Link to post
Share on other sites

This isn't good news, it just means that the best maxwell chip will be locked behind a 1000$ paywall for months before we get a gaming version and that will be expensive as heck too. It's not like I'll be upgrading my r9 290 anytime soon, but still.

Don't ask to ask, just ask... please 🤨

sudo chmod -R 000 /*

Link to comment
Share on other sites

Link to post
Share on other sites

Is this even necessary? I mean are games evolving that fast to use all of this cuda cores? 

Link to comment
Share on other sites

Link to post
Share on other sites

Is this even necessary? I mean are games evolving that fast to use all of this cuda cores? 

Yes, especially with 4K becoming more common and affordable. A 780 Ti, which again has the same number of Cuda Cores as a Titan Black, can barely maintain 60fps at 1440p, and can't come close to 60fps at 4K. Not unless you significantly reduce the quality settings.

http://www.guru3d.com/articles_pages/geforce_gtx_780_ti_review,23.html

http://www.guru3d.com/articles_pages/geforce_gtx_780_ti_review,25.html

i7 2600K @ 4.7GHz/ASUS P8Z68-V Pro/Corsair Vengeance LP 2x4GB @ 1600MHz/EVGA GTX 670 FTW SIG 2/Cooler Master HAF-X

 

http://www.speedtest.net/my-result/3591491194

Link to comment
Share on other sites

Link to post
Share on other sites

Is this even necessary? I mean are games evolving that fast to use all of this cuda cores? 

these high end cards aren't really meant for gaming

 

edit: or you can use them for 4K

build log: diagonalmod (RIP?)


i know i use many of these: ( ) and these: ... (i really do... (sry...) ) edit: and edits

Link to comment
Share on other sites

Link to post
Share on other sites

Yes it is because Maxwell has been changed due to the delay at TSMC with the production of 20nm chips. We already the 750 Ti that uses a modified 28nm Kepler chip under the Maxwell "GM" name. Rumor is that because of the delay, which has since been fixed, Nvidia is going to skip 20nm for their GPUs. Supposedly we will get redesigned 28nm chips under the 800 series and Maxwell name and in a year or less time we Nvidia will jump to 16nm chips. Possibly keeping the 800 series name plate with some added identifier to separate them from the cards with 28nm chips.

The rumor that Nvidia is skipping the 20nm chips is also supported by the fact that Apple is now being given first priority for 20nm chips for their next set of iPhones and iPads coming in September. TSMC has already stated delivering those chips to Apple.

I'm slightly confused  :(

 

GM107 = 28nm Maxwell

GK210 = 28nm Maxwell

 

I must be missing something important here  :o

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

Yes, especially with 4K becoming more common and affordable. A 780 Ti, which again has the same number of Cuda Cores as a Titan Black, can barely maintain 60fps at 1440p, and can't come close to 60fps at 4K. Not unless you significantly reduce the quality settings.

http://www.guru3d.com/articles_pages/geforce_gtx_780_ti_review,23.html

http://www.guru3d.com/articles_pages/geforce_gtx_780_ti_review,25.html

 

 

these high end cards aren't really meant for gaming

 

edit: or you can use them for 4K

 

 

Ok I see now. Well it looks like we will have to sell some organs to afford it. 

Link to comment
Share on other sites

Link to post
Share on other sites

 

I'm slightly confused  :(

 

GM107 = 28nm Maxwell

GK210 = 28nm Maxwell

 

I must be missing something important here  :o

 

 

 

GK210 is Kepler, NOT Maxwell. You misread what they are saying. They never said it was Maxwell. The quote you provided only said that "GM200/GK210 will form the next generation of Titan flagship". These first Maxwell chips are nothing more than modified Kepler chips, so what they are saying is the GM200 will be based off of a modified GK210. We also mays see a modified GK210 chip, under the GM210 name, with the same lower 2880 Cuda Core count that will benefit from the higher efficiency and lower temps that all the GM chips are supposed to have.

 

 

Ok I see now. Well it looks like we will have to sell some organs to afford it. 

 

Not necessarily. The 28nm being old technology are actually cheaper to manufacture now,. Which happens with all chips, they get cheaper to make over time. the rumors floating around almost all say these first 800 series cards will be less expensive than their 700 series counterparts.

i7 2600K @ 4.7GHz/ASUS P8Z68-V Pro/Corsair Vengeance LP 2x4GB @ 1600MHz/EVGA GTX 670 FTW SIG 2/Cooler Master HAF-X

 

http://www.speedtest.net/my-result/3591491194

Link to comment
Share on other sites

Link to post
Share on other sites

GK210 is Kepler, NOT Maxwell. You misread what they are saying. They never said it was Maxwell. The quote you provided only said that "GM200/GK210 will form the next generation of Titan flagship". These first Maxwell chips are nothing more than modified Kepler chips, so what they are saying is the GM200 will be based off of a modified GK210. We also mays see a modified GK210 chip, under the GM210 name, with the same lower 2880 Cuda Core count that will benefit from the higher efficiency and lower temps that all the GM chips are supposed to have.

In the article:

According to the source the GM200/GK210 codename GPU will be built on the 28nm fabrication process using Maxwell architecture

But the 2nd part of your post makes more sense to me, thank you. As they changed how the streaming multiprocessors are made in blocks of 32 so each Maxwell unit is 128 CUDA cores down from 192 in Kepler.

 

Took me a while to figure out why it was different but I got there :P

Link to comment
Share on other sites

Link to post
Share on other sites

In the article:

But the 2nd part of your post makes more sense to me, thank you. As they changed how the streaming multiprocessors are made in blocks of 32 so each Maxwell unit is 128 CUDA cores down from 192 in Kepler.

 

Took me a while to figure out why it was different but I got there :P

You are still misunderstanding what they are saying. By saying "GM200/GK210" they are saying "The GM200, which is based off the GK210, will use the Maxwell architecture.". Which is exactly what I said as well. All of these new chips are modified Kepler chips but will use the GM and 800 series name.

i7 2600K @ 4.7GHz/ASUS P8Z68-V Pro/Corsair Vengeance LP 2x4GB @ 1600MHz/EVGA GTX 670 FTW SIG 2/Cooler Master HAF-X

 

http://www.speedtest.net/my-result/3591491194

Link to comment
Share on other sites

Link to post
Share on other sites

Maxwell has 1,5x more performance per Cuda Core compared to Kepler.
If it actually has 4000 Cuda Cores we are talking here around 10Tflops or about 2,5x GTX780 TI performance.
It would make sense tho because such a card would be actually able to play games in 4K.

 

RTX2070OC 

Link to comment
Share on other sites

Link to post
Share on other sites

Will we have 2 dozen threads a week about people complaining how its too expensive when they can't even afford it and all they'd use it for is surfing Facebook?

Link to comment
Share on other sites

Link to post
Share on other sites

I'll take all of them.

Main rig on profile

VAULT - File Server

Spoiler

Intel Core i5 11400 w/ Shadow Rock LP, 2x16GB SP GAMING 3200MHz CL16, ASUS PRIME Z590-A, 2x LSI 9211-8i, Fractal Define 7, 256GB Team MP33, 3x 6TB WD Red Pro (general storage), 3x 1TB Seagate Barracuda (dumping ground), 3x 8TB WD White-Label (Plex) (all 3 arrays in their respective Windows Parity storage spaces), Corsair RM750x, Windows 11 Education

Sleeper HP Pavilion A6137C

Spoiler

Intel Core i7 6700K @ 4.4GHz, 4x8GB G.SKILL Ares 1800MHz CL10, ASUS Z170M-E D3, 128GB Team MP33, 1TB Seagate Barracuda, 320GB Samsung Spinpoint (for video capture), MSI GTX 970 100ME, EVGA 650G1, Windows 10 Pro

Mac Mini (Late 2020)

Spoiler

Apple M1, 8GB RAM, 256GB, macOS Sonoma

Consoles: Softmodded 1.4 Xbox w/ 500GB HDD, Xbox 360 Elite 120GB Falcon, XB1X w/2TB MX500, Xbox Series X, PS1 1001, PS2 Slim 70000 w/ FreeMcBoot, PS4 Pro 7015B 1TB (retired), PS5 Digital, Nintendo Switch OLED, Nintendo Wii RVL-001 (black)

Link to comment
Share on other sites

Link to post
Share on other sites

Maxwell has 1,5x more performance per Cuda Core compared to Kepler.

If it actually has 4000 Cuda Cores we are talking here around 10Tflops or about 2,5x GTX780 TI performance.

It would make sense tho because such a card would be actually able to play games in 4K.

 

A little off on your numbers, Nvidia says it's 135%, or 1.35x, not 1.5x. Also have to keep in mind those numbers are typically inflated. But either way it's not 1.5x.

nvidia-maxwell-performance-kepler.jpg

i7 2600K @ 4.7GHz/ASUS P8Z68-V Pro/Corsair Vengeance LP 2x4GB @ 1600MHz/EVGA GTX 670 FTW SIG 2/Cooler Master HAF-X

 

http://www.speedtest.net/my-result/3591491194

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×