NVIDIA Kepler speculation thread

Yes. You realize its roughly the same performance as the 7970m, which is quite a bit smaller (pitcairn).

Mobile GPUs are traditionally more humble than their desktop brothers. The point remains that NV intended all along to use GK104 for the mobile space as their mobile top dog, despite that many insisted it won't be the case.
 
Mobile GPUs are traditionally more humble than their desktop brothers. The point remains that NV intended all along to use GK104 for the mobile space as their mobile top dog, despite that many insisted it won't be the case.

They had not really the choice if they wanted something too bring on for laptop with high end mobile gpu before the half of the year.

The 680M is still a 670 cores running at 720mhz. The performance are closer of a GTX660 instead of a 660TI-670...

Even the 675MX and 670MX run really low under the performance of the a 660 GK106 desktop with low clock speed. The line up is really declined from the performance of the first released 680M and go down then.

I hope Nvidia is planning a real refresh for the future laptop gpu's and not just an increase on clock..
 
Last edited by a moderator:
No idea for their future mobile plans, but I would not suggest that a GK110 would make it into notebooks at all.
 
Yes. You realize its roughly the same performance as the 7970m, which is quite a bit smaller (pitcairn).

The 680M seems to use less power than that 7970M. With equal wattage it could have been made faster. GK106 based higher clocked solution probably wouldn't be able to compete against the 680M.
 
The 680M seems to use less power than that 7970M. With equal wattage it could have been made faster. GK106 based higher clocked solution probably wouldn't be able to compete against the 680M.

The difference is not extreme in watts and performance wise... I dont really see the 680M in front of the 7970M .. ofc with higher clock, this will make pass the 680M in front of the 7970M. and again, not by a big margin

http://www.notebookcheck.net/Review-GeForce-GTX-680M-vs-Radeon-HD-7970M.77110.0.html

Summarizing all gaming benchmarks, the old generation (HD 6990M & GTX 580M/675M) is roughly 30 percent slower on average. Without Enduro, the HD 7970M is marginally faster than the GTX 680M. With Enduro, the AMD model is slightly outperformed.
However, several results suggest that the GTX 680M could do better as well. We were somewhat surprised that several Nvidia-favored games actually ran more smoothly on the HD 7970M. This might be the fault of the driver version. Nonetheless, we will update as necessary and possibly do a follow-up test with an MSI barebones notebook soon.
 
Last edited by a moderator:
The underlying topic under discussion was, whether the GK104 can work and perform as a valid mobile chip and it easily can, in fact from your link:

"It’s a fact that Nvidia have produced the overall better GPU with the GeForce GTX 680M."
 
Lol what MSI is doing there looks very fishy.
I'm not quite sure though (and even the full article doesn't explain that) why this actually results in higher boost clock - the actual voltage increase is fairly small anyway. Is the RT8802A delivering some feedback to the gpu which gets messed up in this case so the gpu always thinks the power target hasn't been reached yet?
A clever hack I must say though I wouldn't be so confident as msi that powering the chip with a voltage 30% over even the max absolute rating (not to mention of course the guaranteed working rating) is such a good idea long term.
 
The underlying topic under discussion was, whether the GK104 can work and perform as a valid mobile chip and it easily can, in fact from your link:

"It’s a fact that Nvidia have produced the overall better GPU with the GeForce GTX 680M."

No, the underlying topic was whether it was initially intended as a mobile part. Not that it couldn't work. I'm sure you could cripple any part enough to get under 100W or whatever the limit would be, that doesn't mean it was the desired goal.
 
Lol what MSI is doing there looks very fishy.
I'm not quite sure though (and even the full article doesn't explain that) why this actually results in higher boost clock - the actual voltage increase is fairly small anyway. Is the RT8802A delivering some feedback to the gpu which gets messed up in this case so the gpu always thinks the power target hasn't been reached yet?
A clever hack I must say though I wouldn't be so confident as msi that powering the chip with a voltage 30% over even the max absolute rating (not to mention of course the guaranteed working rating) is such a good idea long term.

Here is some more information.
http://www.brightsideofnews.com/new...proving-quality-or-strangling-innovation.aspx
 
No, the underlying topic was whether it was initially intended as a mobile part. Not that it couldn't work. I'm sure you could cripple any part enough to get under 100W or whatever the limit would be, that doesn't mean it was the desired goal.

Of course it was going to be made into a mobile part. Gf114, 104 made it into the mobile space, heck even the gf100 initially made it into the mobile space. Nvidia has never been scared to use larger GPU in their mobile parts.

If you believe that gf104 wasn't intended for the mobile sector, then you must think AMD engineers are incompetent. That being, Nvidia somehow designed a better mobile GPU than AMD by designing a purely focused desktop and somehow it overcame AMD ability to design a mobile part.

The low power characteristics compared to earlier Nvidia designs made it an ideal flagship part.
 
If you believe that gf104 wasn't intended for the mobile sector, then you must think AMD engineers are incompetent. That being, Nvidia somehow designed a better mobile GPU than AMD by designing a purely focused desktop and somehow it overcame AMD ability to design a mobile part.

Same speed with only 35% more die space...
 
With lower power consumption. In mobile space that makes a lot of sense.

Bingo. It is the most important metric along with performance. Unless they are hardly charging any money for the part, size is hardly an issue, the close to 400 dollar premium ensure that they are making money. Power consumption and heat are a much bigger deal in the mobile space.

Gk104 is hardly a big part for nvidia in the mobile space considering their history.
 
Bingo. It is the most important metric along with performance. Unless they are hardly charging any money for the part, size is hardly an issue, the close to 400 dollar premium ensure that they are making money. Power consumption and heat are a much bigger deal in the mobile space.

Gk104 is hardly a big part for nvidia in the mobile space considering their history.

You can say that again; especially for the cases where they used salvage high end cores in the past.
 
I'm sorry, what's going on exactly?

Nvidia and their partners involved in a big scandal. I think NV as usual play the dirty role with negative consequences to all except probably to them.

Comments from NVIDIA regarding the voltage scandal with graphics cards

http://translate.google.sk/translat...andalu_s_napryazheniem_videokart.html&act=url

Yesterday's news plunged into mourning not only the supporters of NVIDIA, and overclocking in general. The company has banned at least two of its partners to provide users with the means to increase the voltage on the GPU GeForce GTX 680. Discerning minds gathering that generates any technical conference, immediately determined that these steps are aimed at reducing NVIDIA percent warranty return cards and increase profitability.

http://www.tomshardware.co.uk/MSI-GTX-660-670-overvolting-PowerEdition,news-40278.html

Our colleagues over at Tom’s Hardware Germany seem to have caught MSI red handed, overvolting the GPUs on the GTX 660 Ti and GTX 670 Ti Power Edition boards (Google Translate) to achieve a higher and longer lasting GPU boost state by basically circumventing the PWM controller. In other words, MSI was cheating. Perhaps no one would ever have known if it hadn’t been for one side effect. The increased voltage can cause the system to refuse to POST.

That's rubbish from NV and that team in Germany. WTH people, what's wrong with you?
 
Back
Top