Thursday, May 17th 2012

NVIDIA Readies GK104-based GeForce GTX 680M for Computex

NVIDIA is readying a high-performance mobile GPU for a Computex 2012 unveiling. Called the GeForce GTX 680M, the chip is based on its trusty 28 nm GK104 silicon, but with about half its streaming multiprocessors disabled, resulting in a CUDA core count of around 768. Reference MXM boards of the chip could ship with memory options as high as 4 GB, across a 256-bit wide memory interface. With the right craftsmanship on the part of NVIDIA, the GTX 680M could end up with a power draw of 100W. A Chinese source had the opportunity to picture the reference board qualification sample, and put it through 3DMark 11, in which it was found to be roughly 37% faster than the GF114-based GeForce GTX 670M, scoring 4905 points in Performance preset. The test bed was driven by Intel Core i7-3720QM quad-core mobile processor.

Source: VideoCardz
Add your own comment

54 Comments on NVIDIA Readies GK104-based GeForce GTX 680M for Computex

#1
KooKKiK
by: Benetanegia
You can laugh all you want. Stock HD7970M does ~5700 as shown by the link I provided. So I was not far off.

HD7850 does around 5500 too and it's clocked close, 860 Mhz.

So laugh, laugh all you want, you only look like an idiot.

PS: People who know me for a long time know my track record of nailing the performance of upcoming cards based on specs and I don't use methods much more complicated than that. Only difference is I adjust based on where I think or know there will be a bottleneck and some other things. And if you think it's not as simple as that, maybe you should think about taking some computer science classes.

EDIT: And/or read below.



Yeah I didn't notice that when I replied, then after that I moved on to other things. 1035 Mhz... hmm

P5746 @ 850 Mhz (stock)
P6951 @ 1035 Mhz

Let's do an stupid thing, let's normalize the 6951 result to stock clocks. You know that soooo stupid and laughable thing...

6951 * 850/1035 = 5708 what? wait... but no because that would be so stupid. To think that...
ok ok, THE 680M gonna beat 7970 even the desktop version.

dream on, boy :laugh:
Posted on Reply
#2
Benetanegia
by: KooKKiK
ok ok, THE 680M gonna beat 7970 even the desktop version.

dream on, boy :laugh:
And you keep looking like an idiot. I never said it would be faster. I said it would be 10% slower, maybe some more, because the HD7870 results I used for my calculations were apparently old, old drivers I mean, or simply in the low-ish side for whatever reason. In any case 10% or 15% (going by the 5746 OR 6951 results) difference does not fit any definition of "being eaten alive".
Posted on Reply
#4
Benetanegia
previous argument lost -> move to next argument eh? ;)

GTX 580 does beat HD6970 for more than 15%, and I've never ever heard anyone say that it eats HD6970 alive or HD6950 for that matter.

PS: Someone, a long time ago made a post requesting that users stop using expressions like crush, devour, eat alive, etc. when talking about performance differences. My opinion back then was "man leave people express themselves as they see fit". Now... :laugh: Wow. Imagine how many posts and wasted time would have been avoided if you had not used that lame expression or if you had explained what it really means for you in your first repply instead of posting inconclusive scores and stupid laughin smilies...
Posted on Reply
#5
KooKKiK
nope, thats all same argument ( as i told you "already said" ) ;)


and you havent answered me about the photograph either ;)




ps. got beaten by ( 5746 ( from your score ) - 4905 ) / 4905 = 17.14% in favoring benchmark.

ok, thats really competitive for me :laugh:
Posted on Reply
#6
Benetanegia
by: KooKKiK
nope, thats all same argument ( as i told you "already said" ) ;)


and you havent answered me about the photograph either ;)




ps. got beaten by ( 5746 ( from your score ) - 4905 ) / 4905 = 17.14% in favoring benchmark.

ok, thats really competitive for me :laugh:
I think you don't understand the use of smilies...

or you are flirting with me... unless you are Charlize Theron, I'm not interested.
Posted on Reply
#7
Andy77
by: Benetanegia
And what about that? That I find it very very unlikely and posibly fake, considering that not even the desktop HD7870 is so fast.
No! You forgot to read... the rest, Einstein. CPU is Sandy at 4,5GHz, GPU at 1035 MHz. These make up that score not some reference figures you thought of.

And yes, 10% means eating it alive!... Kepler can barely overtake Tahiti at default clocks, put Tahiti to its popper clocks and OC3D shows you it can beat any GK104.
Posted on Reply
#8
theoneandonlymrk
jebus guys , whats it matter, 680M or 7970M , ill take two of both, either or just one of any

in the whole thread i found this that was worth a read/mention

by: DarkOCean
Cutting the sp count in half seems a little drastic where's gk 106?
+1 , i thought NV announced its release?? or was that in my head , must be as surely it would be more efficient etc.
Posted on Reply
#9
dieterd
7970M will eat this alive and in breakfast, because 7970m in reviews show performance like 675m SLI (again 1x7970M = 2x675M)! and if Nvidia do not even dear compare 680m vs 675m (old 580m) then there wont be any +10% performance increase.
check out: http://www.notebookcheck.net/AMD-Radeon-HD-7970M.72675.0.html

7970M is based on 7870 desktop and cloked down to 7850 desktop performance level and it is HUGE power in mobile segment it is like 6950 desktop - if year before someone would say that there will be single mobile card at HD 6950 desktop level I would give him :slap: immediately
Posted on Reply
#10
Over_Lord
News Editor
by: KooKKiK
not to believe from user review. thats ok for me.

but...



where did you get that 5500 score from ??? an official review ??? HA-HA-HA :laugh:


and yeah, WE WILL SEE ;)
Some people refuse to believe haha.

Yes buddy, rightly said. We WILL SEE :nutkick:
Posted on Reply
#12
Xzibit
by: Benetanegia
* This slide came from Nvidia too, and it's almost entirely true, so I don't see your point, show me a single Nvidia slide in the last 3-4 years that didn't turn out true (bare a slight exageration), it's not AMD's marketing department we ar talking about here ;)
Thats easy.

Take any Nvidia slide that compares Kepler to competition and claims faster performance without specifing Game GPU. Even Fermi is a faster number cruncher then Kepler so there you have it.

:nutkick:

Believing in marketing slides? You'd have to believe them from both sides.:rolleyes:
Posted on Reply
#13
Benetanegia
by: Xzibit
Thats easy.

Take any Nvidia slide that compares Kepler to competition and claims faster performance without specifing Game GPU. Even Fermi is a faster number cruncher then Kepler so there you have it.

:nutkick:

Believing in marketing slides? You'd have to believe them from both sides.:rolleyes:
Show one or STFU. I have not seen such claim made by Nvidia, no released Kepler (aka GK104 or GK107) has been marketed as a GPGPU powerhouse, so you fail, big time. Every single slide I've seen has the games listed in the X axis so... nice try, learn a little reading comprehension.

EDIT: I should not reply to you, because I give it 3 posts, max, before you talk about NVDA stocks and how that somehow means something related to this thread...
Posted on Reply
#14
Xzibit
by: Benetanegia

Show one or STFU. I have not seen such claim made by Nvidia, no released Kepler (aka GK104 or GK107) has been marketed as a GPGPU powerhouse, so you fail, big time. Every single slide I've seen has the games listed in the X axis so... nice try, learn a little reading comprehension.
Then i guess you have selected memory.

05/24/12 GeForce / DT - Jeff Fisher

Just browse thru it and those marketing slides dont have games just statements. 3 / 7 / 12 compared to Fermi

Dont get mad :) You asked i replied.
Posted on Reply
#15
Benetanegia
by: Xzibit
Then i guess you have selected memory.

05/24/12 GeForce / DT - Jeff Fisher

Just browse thru it and those marketing slides dont have games just statements. 3 / 7 / 12 compared to Fermi

Dont get mad :) You asked i replied.
:laugh: FAIL. There's only 2 performance slides. Both have games listed... :nutkick:

How could I be mad when you just proved you wrong?? :roll:

And of course, no single mention of GPGPU performance.

What's more the entire marketing material is gaming oriented, it says it in the first page: Jeff Fisher, SVP GeForce: if your reading comprehension fails that means Jeff Fisher Senior Vice President of the GeForce division. GeForce division is the gaming division fyi. I know I'm patronising, I somehow felt like I should, don't know why... :laugh:

And wait there's more, here's the GK110 white paper, you know the one and only Kepler chip Nvidia has ever claimed to be aimed at GPGPU: http://www.nvidia.com/content/PDF/kepler/NVIDIA-Kepler-GK110-Architecture-Whitepaper.pdf

First page, no need to delve more into it:
The Fastest, Most Efficient HPC Architecture Ever Built
Versus
The fastest, most efficient GPU ever built
See the difference??
Posted on Reply
#16
Xzibit
I love his reasoning

Nvidia has never been wrong in its marketing slides ever in 3-4 years. Find one.

GeForce = Games only so i'm still right. Don't think thats what you asked unless you meant something else in your head.

Even slide 7 has 20% faster performance then competitor. I'm pretty sure it was 50/50 split and the games it came out infront the % wasnt 20%. Unless W1zzard is wrong.

Got to love these blinded fanboys. Even marketing is true for them to the letter.
Posted on Reply
#17
Benetanegia
by: Xzibit
I love his reasoning

Nvidia has never been wrong in its marketing slides ever in 3-4 years. Find one.

GeForce = Games only so i'm still right. Don't think thats what you asked unless you meant something else in your head.

Even slide 7 has 20% faster performance then competitor. I'm pretty sure it was 50/50 split and the games it came out infront the % wasnt 20%. Unless W1zzard is wrong.

Got to love these blinded fanboys. Even marketing is true for them to the letter.
No, I love your reasoning, you gotta love it. You take marketing material from the gaming department which talks exclusively about gaming and pretend they are talking about anything but gaming. Bravo. Like I said they are very especific, in their GPU marketing material they talk about graphics performance, in their HPC marketing material they talk about HPC performance. You can try to distort that as much as you want, you are not right and cannot provide a single slide to support your idea.

My initial statement was clear, I never said Nvidia slides are accurate, I said that they have not been false, outright false, and hence their claim of 20% faster or whatever is not going to suddenly become into performance in which the 7970M "eats it alive". Again try to distort it as much as you want.
Posted on Reply
#18
Xzibit
by: Benetanegia
No, I love your reasoning, you gotta love it. You take marketing material from the gaming department which talks exclusively about gaming and pretend they are talking about anything but gaming. Bravo. Like I said they are very especific, in their GPU marketing material they talk about graphics performance, in their HPC marketing material they talk about HPC performance. You can try to distort that as much as you want, you are not right and cannot provide a single slide to support your idea.
I think i just did.

The problem is your brain cant accept it.;)
Posted on Reply
#19
Benetanegia
by: Xzibit
I think i just did.

The problem is your brain cant accept it.;)
You might think you did. You didn't. They are talking about GPU not HPC or GPGPU. First learn the terms then come and discuss. If it's not the fastest Graphics Processing Unit on the planet, show me the one that it is.

Here, some help:



If it's not the GK104 in the GTX680, then which one?
Posted on Reply
#20
Xzibit
by: Benetanegia
You might think you did. You didn't. They are talking about GPU not HPC or GPGPU. First learn the terms then come and discuss. If it's not the fastest Graphics Processing Unit on the planet, show me the one that it is.
I dont think i did. I know.

You might want to start asking every reviewer out there to make the performance gap 20% or greater now.

Please go fix every review out there on the net please ;)
Posted on Reply
#21
theoneandonlymrk
ok where my posts go.. tut

anyways then

imho it would be worth awaiting a direct unbiased review prior to purchaseing this or a competitors solution, and that either way , this gpu looks very good for mobile gameing




I actually dont mind them(post) goin if cut:)
Posted on Reply
#22
Benetanegia
by: Xzibit
I dont think i did. I know.

You might want to start asking every reviewer out there to make the performance gap 20% or greater now.

Please go fix every review out there on the net please ;)
There's several games where it is more than 20% faster, who's to say that's not the real performance of the card? Who decides what is what? You have the same problem as theoneandonlymrk, you think you can take other people's words and decide what they mean. They don't say it's 20% faster overall, they say it's 20% faster period, which s something that cannot be completely tested or refuted. Just as an example if a card is bottlenecked by a CPU it's no longer faster than another card? That is only an example, what if games are not optimized for it, drvers have yet to implement optimizations, etc? It make the card slower, or is the other conditions which make the SYSTEM slower?

Like I said, there's plenty of games where it's 20% faster. In the ones where its fps are not 20%, it could be because it's bottlenecked, because it's not properly optimized for, because of many other unknown issues or because it's just plainly slower. The fact is that you don't know and that makes the claim as true as it could be false. And that's why PERFORMANCE slides exist, where you can see exacty where they say that it is faster, and I never said you have to believe them, but like I did say Nvidia's ones have been far more honest lately: they included Deus Ex in which it's slower and they had absolutely no need for that.

They are not saying it's 20% faster in any specific situation, only that it's 20% faster and in dozens of cases that is just plainly true. So stop taking ONE SINGLE statement, make your own interpretation of it and pretend they are saying what they are not saying. Hint: they are not saying it's 20% faste in GPGPU.

And like I said, in the end it IS faster and as such 680M IS going to be faster than 7970M, it's not going to be "eaten alive", and that has been my only point. In fact until mrk did it, I did not mention the slides AT ALL. I linked to that page for the specs, not the slides, it's not my problem if all of you can't see past some shiny marketing slides. based on specs 680M is far more likely to "eat the 7970M alive" than being the dinner...
Posted on Reply
#23
Xzibit
by: Benetanegia
There's several games where it is more than 20% faster, who's to say that's not the real performance of the card? Who decides what is what? You have the same problem as theoneandonlymrk, you think you can take other people's words and decide what they mean. They don't say it's 20% faster overall, they say it's 20% faster period, which s something that cannot be completely tested or refuted. Just as an example if a card is bottlenecked by a CPU it's no longer faster than another card? That is only an example, what if games are not optimized for it, drvers have yet to implement optimizations, etc? It make the card slower, or is the other conditions which make the SYSTEM slower?

Like I said, there's plenty of games where it's 20% faster. In the ones where its fps are not 20%, it could be because it's bottlenecked, because it's not properly optimized for, because of many other unknown issues or because it's just plainly slower. The fact is that you don't know and that makes the claim as true as it could be false. And that's why PERFORMANCE slides exist, where you can see exacty where they say that it is faster, and I never said you have to believe them, but like I did say Nvidia's ones have been far more honest lately: they included Deus Ex in which it's slower and they had absolutely no need for that.

They are not saying it's 20% faster in any specific situation, only that it's 20% faster and in dozens of cases that is just plainly true. So stop taking ONE SINGLE statement, make your own interpretation of it and pretend they are saying what they are not saying. Hint: they are not saying it's 20% faste in GPGPU.
Looks like you agree with me.

You could have just said I was right instead of try'n to explain it away and contradict yourself.

I guess it hurts too much to admit your wrong. Let me know if you need a hug.
Posted on Reply
#24
Benetanegia
by: Xzibit
Looks like you agree with me.

You could have just said I was right instead of try'n to explain it away and contradict yourself.

I guess it hurts too much to admit your wrong. Let me know if you need a hug.
You have some serious problem with comprehension, but ey I guess it's compfortable there inside your mind. I didn't agree with you at all and I'm still waiting for that proof where Nvidia claimed GK104 to be faster on anything but gaming.

You can't admit defeat and will fight any slight remain you think you have left, but you have none. You can't provide a single slide that isn't essentially true. And that's the key to all of this as I've explained several times, we are not discussing slides in general, their truthness. If you are just doing that, pff you couldn't go more off-topic. We are discussing the slides that pertain to this thread and how they relate to the real performance of the 680M. And we are not discussing about a difference of 10%, as in they said GTX680 (desktop) is 20% but it's only 10% faster. No, we are discussing the fundamental change between what was said before, the HD7970M beating it by a substantial margin, 20% or more and the reality being the absolute oposite. If you are discussing the truthness of slides based on the premise that they are false because they are off by 10%, you are doing something really stupid, because there's no single/absolute truth like that, not even W1zz reviews can represent that, they only represent a much wider array of results and that's all. Change the game selection or the settings and you will find another "truth". I never told you to find slides which were slightly off the reality, I told you show me slides which are false. There's none, you lose, time to move along.
Posted on Reply
#25
Xzibit
Awe come on now. Stop being a sore looser.

You even went back and tried to edit what you said out..:laugh:

by: Benetanegia
* This slide came from Nvidia too, and it's almost entirely true, so I don't see your point, show me a single Nvidia slide in the last 3-4 years that didn't turn out true (bare a slight exageration), it's not AMD's marketing department we ar talking about here ;)
^I like the "Almost entirely true" part and then continues to ask for slide thats isnt true in 3-4yrs when he doesnt even believe the one hes refering to as 100%.

Somehow you have a problem with competition slides. I wonder why that is, Just curious.

So tell me again how you can prove or disprove Nvidias 20% performance claim

by: Benetanegia
The fact is that you don't know and that makes the claim as true as it could be false. And that's why PERFORMANCE slides exist, where you can see exacty where they say that it is faster, and I never said you have to believe them
Well if your going by Performance slides. Do all reviewers score said GPU at 20% or more all the time since after all Nvidia didnt specify the situation. Do Nvidias own results give a 20% increase all the time ?
05/24/12 GeForce / DT - Jeff Fisher
On page 8 where the games are I only see 5 games out of 13 meet the 20% or better performance claim. Less then 50% of the games.
It's not like i'm taking them out of context its the very next page.

by: Benetanegia
They are not saying it's 20% faster in any specific situation, only that it's 20% faster and in dozens of cases that is just plainly true. So stop taking ONE SINGLE statement, make your own interpretation of it and pretend they are saying what they are not saying. Hint: they are not saying it's 20% faste in GPGPU.
So like you said yourself that claim only applies to a dozen cases as you see it. I didnt see that footnote in the slide but I guess you know how to translate Nvidia graphs:)
That would make the 20% claim false the rest of the time. By there own graph 8 games out of 13 it doesnt meet 20% faster performance claim so I dont know how you can say its plainly true. :laugh:

So even you know its not right yet you continue to argue the fact that you said show me 1 slide from Nvidia thats not true.

How much more clearer can you get when you said it yourself. Do you not beleive yourself now ? :laugh:
Posted on Reply
Add your own comment