Gtx 660m

[Cloudfire]

Actually I think the GT 635M should be faster or very close to the GT 650M.
635M is no Kepler but a rebrand and is basicly a rebranded 555M and with two versions, one with 128bit and another one with 192 bit memory bus

The GT 555M is like 30% (sometimes more) faster than 550M. I don`t know how much faster the 650M will be compared to 550M but it have to perform quite a bit to surpass the 635M (555M).

[Kevin]

If the 635M ends up being faster than the 640M and 650M, Nvidia will have crossed into a new plane of ridiculousness.

You want to talk about confusing customers? Bigger means better, not slower.

[Cloudfire]

Fully agreed

[gamba66]

If you compare the picture of the N13P-GT (650m kepler codename) with the specs of the N12P-GT (550M), you might see some similiarities... Also this underlines that the codename n13p-gt might be very real!

Check out the text from the source of the picture!

While at this year's CES, we received a tip from an unnamed source that there were Kepler GPUs located in some Samsung laptops on the floor. So, we decided to go and check it out for ourselves and surprisingly enough, there were indeed some Mobile 600 series GPUs.

If you take a look at our images below, you will see that we saw two different laptops with one major variation, their GPUs. These were Samsung's Series 5, 550P5 and 550P7 laptops. The real major difference here was that this one was a Kepler GPU and the other was simply a rebadged Fermi mobile GPU. Both laptops also support NVIDIA's optimus switching technology.

If you look at the shader cores you can see that the N13P-GT is going to be a mid-range performance part as it features 2GB of dedicated GPU memory as well as 96 CUDA cores and a 405MHz core clock. What's interesting is that if you compare this to the GT 630 GPU in the 550P7 laptop, you can see that both of them have the same amount of CUDA cores (shaders). The GK-107 part, the N13P-GT is actually clocked lower than the GT 630 GPU in the 550P7 laptop. Either this indicates to us that NVIDIA's Kepler mobile GPUs really do have an improved performance per clock, or the N13P-GT isn't as high-end of a part as had been rumored. Both GPUs have the exact same 2GB of DDR3 memory and both have the same 1.8GHz effective memory clock.

The similarity between clocks is really what confuses us, though, since the N13P-GT is a 405MHz part and the 630 GT is a 475MHz part. Those clock differences are followed by similarities in shader clock, which also mimic the GPU clock differential. We're just not sure why NVIDIA would make two so similar parts unless one was a newer generation part and the other was simply a rebadge.

We're still not entirely sure whether or not N13P-GT is a Kepler part or not considering its relative similarity to the GT 630 which has been known to be a Fermi rebranded mobile GPU. We are hopeful that we've stumbled upon the first Kepler part in a mobile devices, but we remain somewhat cautious about it being one considering all of the similarities.

Source: Samsung Laptops Show New NVIDIA 600 Series Mobile GPUs - Bright Side Of News*

In the original article they compare the specs to a 630m, which is in another samsung notebook next to the n13p-gt notebook.. But now compare the specs to the N12P-GT

The 550m is basically an overclocked 540m (which the 630m is a rebrand of) thats why the similiarities in the article

GeForce GT 550M 96@740MHz
Codename	N12P-GT
Pipelines	96 - unified
Core Speed *	740 MHz
Shader Speed *	1480 MHz
Memory Speed *	900 MHz
Memory Bus Width	128 Bit
Memory Type	DDR3/GDDR5
Max. Amount of Memory	1536 MB
Shared Memory	no
DirectX	DirectX 11, Shader 5.0
technology	40 nm
Notebook Size	large
Date of Announcement	06.01.2011

So if we look at the charts again, we see that the refresh or followup of the 555m (codename N12E-GE-B) isnt actually the 650m, but rather the 660m (Codename N13E-GE)!


[/QUOTE]

Only true if the codenames are real, but the article from the samsung notebooks may indicate that they could certainly be true

[TheBluePill]

yes but check out the picture, the 555m has 144 cuda cores and the 550m has 96, so it could really be that the 650m is the followup of the 550m and not 555m

ofcourse we dont know how the new architecture works maybe its still faster

I see what you are saying.

The 555M is now the 635M.

I would figure the 650M will be the replacement for both the 550M AND 555M from the last generation.. And i think a lot of us figure it will be on-par with the 560M from the last generation.

The 660M/670M/675M and Later 680M will improve upon the segments that the 500 series GPUs all occupied.

[maxheap]

If the 635M ends up being faster than the 640M and 650M, Nvidia will have crossed into a new plane of ridiculousness.

You want to talk about confusing customers? Bigger means better, not slower.

oh yes, craziness, I hope they will keep up the bigger the better notion for the sake of all

[gamba66]

guys, im not hoping that the 650m will be weaker then the 635, and obviously the 640m and 650m should be stronger since they have bigger numbers

but for me it looks like the 650m is just a die shrink of the 550m, I do think it brings more perfomance, but how much more is the question? It looks like kepler is faster clock-per-clock but atm we can only guess how much faster kepler really is.. We will just have to wait and see

And like I said, based on the facts I have posted, I think the real follow up for the 555m isnt the 650m, but rather the 660m, since the Codename and TDP fit the description.

BTW, another problem might be availability

Check out these current article on NVIDIA

http://semiaccurate.com/2012/02/15/n...nference-call/

[TheBluePill]

guys, im not hoping that the 650m will be weaker then the 635, and obviously the 640m and 650m should be stronger since they have bigger numbers

but for me it looks like the 650m is just a die shrink of the 550m, I do think it brings more perfomance, but how much more is the question? It looks like kepler is faster clock-per-clock but atm we can only guess how much faster kepler really is.. We will just have to wait and see

And like I said, based on the facts I have posted, I think the real follow up for the 555m isnt the 650m, but rather the 660m, since the Codename and TDP fit the description.

BTW, another problem might be availability

Check out these current article on NVIDIA

Nvidia comes somewhat clean on their quarterly conference call | SemiAccurate

I don't think anyone thought you were advocating for a backward naming-convention. I think we were just confused a bit.

nVidia has a very long, proud, tradition of screwing the pooch on product names. I mean.. Surely an 8800-GTX is Faster than a GTX-280? Right? oh.. Wait.. How about a GeForce 405, has to be a LOT faster than a GeForce 295.. I mean.. its a whopping 110 more!!

HA!

[funky monk]

As far as bigger=better goes, you have to understand one thing, which is that the first digit can more or less be ignored when comparing unless you're looking at cards from a good few years back. For example, a 580 is only a little faster than a 480, but the 480 would completely destroy a 520.

[Kevin]

via Clevo:



Believable? Discuss.