Kepler news and discussion

Page 18 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.

true, very true. They should of had a Dev team seperate for the new drivers
 
Yea like wise. but heres some bit of news for everyone....

Since tiwanese chip nm processes have decided to get slack on the whole chip manufacturing. Intel and NVIDIA are sharing facilities to .....well share!

So we should be able to get ahold of more 28nm gpu's from NVIDIA in the weeks to come but.... Stock for now is limitied with the rest of kepler come the way and also they have to same chips for the non reference but still stock is thin. I grow tired of people buying 20 as soon as they come in stock just to make a youtube video saying look at me i have 10 680's but i can only use 4....
 


I hope your wrong, I'm waiting for another week and in the worst case scenario grabbing an overpriced 7870. I really want to join team green. :cry: :cry: :cry: :cry: :cry: :cry: :cry:
 
I don't think Best Buy price matches online stores, only local places that are within like 50 miles.

It's worth a shot though. Meanwhile, my 680 devours everything I have been playing on it. Read it and weep AMD...

Jk, but I really do love my 680. Getting up early and sitting at my PC for an hour waiting was certainly worth it that day!
 
http://www.guru3d.com/news/zotac-geforce-gtx-680-extreme-clocked-at-12ghz/

Zotac's GeForce GTX 680 Extreme Edition was spotted over at Expreview. This new card features a base and boost clockspeed of 1200MHz (vs stock clockspeed of 1006MHz) and 2GB GDDR5 memory clocked at the standard clockspeed of 6000MHz. The standard core voltage is 1.212V and Zotac told Expreview that the card is capable of delivering up to 1.5V to assist in extreme overclocking. According to Zotac, a core clockspeed of 2GHz might be attainable in a best case scenario using extreme cooling, excellent overclocking skills and some luck. Apart from a strong VRM (detailed in the older article), the card is also backed by several nifty features, such voltage measurement points, extreme-cooling optimized BIOS, and adequate-looking VRM cooling. The USB mini type-B connector on the back of the card turns out to be not a connection to an OC module, but a USB connection to the motherboard. ZOTAC is routing the card's SMBUS to its Firestorm OC software over USB. Maybe they haven't figured out [how] this could be done through the NVIDIA driver interface, or they didn't trust NVIDIA. For example, NVIDIA blocks SMBUS access to components that they deem "sensitive".
 

agreed...
 

I think when they mentioned "extreme cooling", it was assumed they did not mean air cooling.

 

You're right, but it's 25 miles. Since I have Best Buy and Fry's right next to each other, I could use the listed price at Fry's to get them to match the price. By the way, it looks like Fry's has all the manufacturers in stock at $499.
 
I have a felling that Amd's strategy of overpricing their 7000 series is going to backfire. If I have to wait for Nvidia to fully release their cards for their prices to drop then I'm buying a Nvidia card and they lose my money. The only advantage that AMD has has now is compute performance.
 


bitcoin mining :lol:

Does anyone here have any experience overclocking the igpu of the i5-2500k?
 
More details of GK110 released...
Single Precision ~ 4TFLOPS
Double Precision~2TFLOPS
Die Size ~ < 520mm^2
GPC count ~ 6
CUDA count ~ 2304
VRAM ~ 4GB
Bus width ~ 512bit
Launch ~ Q3 2012

http://wccftech.com/nvidia-launching-gk110-features-4gb-memory-25-performance-lift-gk104-launches-q3-2012/
 
Those specs are truly nice.

From a lil' deeper in that article:

- ~ 550mm ² die size (28nm in production at TSMC)
- Round 6 billion transistors
- DirectX 11.1 Kepler architecture
- most likely (compared to the GK104) different structure of shader clusters (SMX)
- clearly more than in 2000 (1D) shader units (to 2500 shader units)
- 2 + Double Precision teraflops computing power, accordingly, probably 4 + Single Precision teraflops computing power
- 512-bit DDR memory interface
- Game Power consumption 250 to 300 watts
- Tape-out until after January 2012 launch not before August 2012

Cheers!
 
what makes me wonder is if this will have a GPU boost too.
And in this chip more percentage of the tdp is reserved for heavy compute tasks than gk104.So if this one does have GPU boost,then in games it'll make use of that excess power to overclocking and hence there will be a massive boost out of the box.
As a result,I'm being a bit skeptic about GPU boost being implemented on gk110.Or even if they do,they might set a limit for that boost or the card may become unstable.
 

Im not sure how the boost works thru n thru the card, but I do know, for DP it looks to be a no no
Tho, if its only for the core, it may work
 


Well, the die area is like 1.5x the size of GK104 and they are projecting a 25% increase over what lil' Kepler can do today... I'm guessing that's a best case scenario in some hand picked tests... Then, given that info, GK110 will muster in GPGPU, but won't be that far away from GK104 in gaming.

Now, current GTX680 owners probably won't crave this card THAT much if its indeed just a 25% increase in games, hahaha.

And in that same note, it seems the 8K series are also on their way according to that article. Any news around the red team?

Cheers!
 
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom