gtx680m specs revealed

[tviceman]

7 SMX's enabled (as many shaders as the gtx670), 720mhz base clock, 256-bit 1800mhz memory for 115.2 gb/s bandwidth.

http://www.engadget.com/2012/06/04/nvidia-geforce-gtx-680M/
http://forum.notebookreview.com/alienware-m17x/668384-news-about-upcoming-gtx-680m.html

Of course it will cost an arm and a leg, but the amount of power at 100 watts is pretty amazing.

EDIT: A gtx675m is only a rebranded gtx580m. Wow there is a massive difference between Nvidia's top mobile GPU and second-best mobile offering. Uh.... maybe they should have named this one the gtx685 and slotted something else between it and the gtx675m.

 

[exar333]

Besides that fact I am getting very tired of just about everyone using the graph-style linked at the top to exaggerate comparisons, it looks very good. If they provide a cut-down mobile with approx. the power of the previous 580M for 40-50w, that would be a AWESOME mainstream part as well.

 

[ShintaiDK]

It was obvious that they would do even better in the mobile space than desktop when we saw Keplers power/performance efficiency.

 

[Skurge]

Interesting choice of games. Will see in independent reviews. I must say though Packing previous gen high end performance into notebooks (be it large ones) is impressive from both AMD and Nvidia.

 

[SolMiester]

The graph is easy to understand if you get over looking at it as a comparison of FPS!...Its a graph of % improvement over the other and is does not show exaggerated performance at all!..LOL

 

[exar333]

The graph is easy to understand if you get over looking at it as a comparison of FPS!...Its a graph of % improvement over the other and is does not show exaggerated performance at all!..LOL

Did you see the first graph? So 2x graph height for a 20% improvement is accurate? Thats tricky marketing by any stretch of the truth. Those graphs are intentionally misleading, and unfortunately everyone in the industry seems to like them recently.

 

[Cookie Monster]

Such a huge gap from the GTX670M which uses a GF114 core iirc. Its quite impressive for them to fit an almost full fledge GK104 in there. Wait so 720MHz isn't the boost clock??

 

[SolMiester]

Did you see the first graph? So 2x graph height for a 20% improvement is accurate? Thats tricky marketing by any stretch of the truth. Those graphs are intentionally misleading, and unfortunately everyone in the industry seems to like them recently.

Read it again...the increments are 20%, there is no 2 x at all.....the radeon is the 100%, the NV is showing performance above that!

 

[iCyborg]

Read it again...the increments are 20%, there is no 2 x at all.....the radeon is the 100%, the NV is showing performance above that!

Yes, there is 2x. The increments are 20%, but the 20% more is visually presented as a 2x longer bar.
Do you think that this graph conveys the proper relation of being 15% better:

It's hard to fault them for this as he says, because everybody including AMD is doing the same, but it is tricky marketing.

 

[ShadowOfMyself]

Yup, its insulting to the customer, thinking they are dumb enough to look at the bars and go "WOW" without actually checking the numbers...

... And the worst part is the average person will probably fall for it /facepalm

 

[SolMiester]

LMAO....please show me where it show 2 x !!.....80>100>120>140 are all even 20% incremental......The Radeon stands at 100%, this graphs show quite clearly the % improvement over the Radeon which is what it is supposed to do!

 

[SolMiester]

Yes, there is 2x. The increments are 20%, but the 20% more is visually presented as a 2x longer bar.
Do you think that this graph conveys the proper relation of being 15% better:

It's hard to fault them for this as he says, because everybody including AMD is doing the same, but it is tricky marketing.

Your graph doesnt, as it is not showing %......what is the point of showing the graph below 80%?

 

[Jaydip]

If I not much mistaken this will probably be a desktop equivalent of Gtx 660.They should know how to create accurate graphs though.

 

[SolMiester]

If I not much mistaken this will probably be a desktop equivalent of Gtx 660.They should know how to create accurate graphs though.

Are you serious about the graph?.....Ive got a bee in my bonnet about this because really all its doing is showing peoples ignorance....

 

[(sic)Klown12]

LMAO....please show me where it show 2 x !!.....80>100>120>140 are all even 20% incremental......The Radeon stands at 100%, this graphs show quite clearly the % improvement over the Radeon which is what it is supposed to do!

Look at the graph without the side. It shows the bar being twice as long which would make anyone who quickly glanced at it think it was twice as fast. Both companies do this to exaggerate the performance increase because they know most won't bother to look at the side.

 

[lopri]

Did you see the first graph? So 2x graph height for a 20% improvement is accurate? Thats tricky marketing by any stretch of the truth. Those graphs are intentionally misleading, and unfortunately everyone in the industry seems to like them recently.

Yeah. But what do you expect? It's marketing. You won't see such graphs in reputable reviews, though.

 

[SolMiester]

Look at the graph without the side. It shows the bar being twice as long which would make anyone who quickly glanced at it think it was twice as fast. Both companies do this to exaggerate the performance increase because they know most won't bother to look at the side.

The Red BARS show quite easily where the horizontal axis is.....and why would you knock off the left hand side which shows the numbering clearly at 100%....
The vertical axis is proportional so we can quite easy see the % above the Radeon.....I dont see any 2 x and the space you are referring too show the red BARS 1/2 up anyway!....THERE IS NO TWICE AS LONG..FFS

 

[iCyborg]

Your graph doesnt, as it is not showing %......what is the point of showing the graph below 80%?

Are you serious?? It doesn't, but only because I didn't put % signs?
Imagine there are % signs, or do you want me to repost the graph with %...

 

[SolMiester]

Are you serious?? It doesn't, but only because I didn't put % signs?
Imagine there are % signs, or do you want me to repost the graph with %...

Okay dude, whatever...I guess Excel isn't a program you use often...and I am dismayed by the number of people who are confused by this graph!

Edit - the graph starts at 80%, as if it was 100%, you wouldnt see any Radeon bars...does that make it easier for everyone?

 

[Jaydip]

Are you serious about the graph?.....Ive got a bee in my bonnet about this because really all its doing is showing peoples ignorance....

Sol if u chart a graph on excel where two data values are 30% apart the graph won't look twice as big

 

[SolMiester]

The graph type is to show % differential cleanly as in #% on the Y axis..... But whatever, not all graphs start at zero.....

Do you see a glass hall empty or half full?
/rant over....sorry.

 

[Jaydip]

If it's whiskey I always see it half empty
But honestly speaking professional chart shouldn't cause so much confusion

 

[3DVagabond]

Nobody is disagreeing with what you are saying, SolMiester. You are just completely ignoring the point they are presenting.

It would be a more accurate visual representation if the graph started at zero. Then 20% better performance would be represented by a graph that was 20% longer, rather than twice as long.

 

[SirPauly]

I don't know, what I did get from the graph was nVidia was up to 30 percent faster than their competition. Understand the point about the size of the bars but the context of the graph was clearly noted.

 

[ShadowOfMyself]

Its just a mind trick really... Even though they clearly state "30%", the huge bars make those 30% look a lot more impressive than they are, and that kind of thing gets into peoples subconscious, even if they dont notice

Its even funnier when you consider they used 80% base to compare vs AMD, and 40% to compare against their own

Eagerly awaiting 99% base graphs, you know its just a matter of time