Next Gen Clevo

Oh I missed the 'flagship' part. My bad, you're right. But the 640m came before the 680 so we may be seeing the Kepler refresh sooner than later, for the midrange notebooks, at least (and definitely earlier than 2014).

 

I have not heard about nvidia using a new chip, just slight power tweaks on the current ones.

 

That is correct.

So expect a 20% performance increase TDP to TDP.
I think that is a decent increase IMO.

 

I don't expect it to be that much since unlike the 40nm process this one was not broken.

 

True, but isn't the architecture a bit tweaked as well ?

 

I suspect that a new line will be slated shortly after the release of Haswell.

 

What he means is that the Kepler architecture was not as power-unoptimized to begin with as Fermi's conception was. Therefore revising kepler might improve power efficiency a tad, but there isn't much to increase, unlike with Fermi.

 

Also when you talk about slight tweaks, with fermi they didn't because it's so complicated to do and expensive. So we will see but a 20% increase for no power gain is too optimistic IMO.

 

Ok. I will rephrase my initial statement, what I meant was that it will be at most 20%. Usually at these types of refresh is around 15%. I would be somewhat disappointed if it's less and happy it it's more.

 

I'd say 10-15% too, hopefully we get something between the current 660M and 670MX too as it's quite a gap.

 

What's the story with the 680MX? Is that just an Apple exclusive thing? Are Clevos ever going to see one?

 

Too high TDP for current machines.

 

Isn't it kind of stupid to call it a mobile graphics card if it can't be used in any mobile devices?

 

Shhhh...shhhhh

Sent from my SCH-I535 using Tapatalk 2

 

Well it's not really a desktop card either since it's in the same form factor as mobile cards, Nvidia did not want to create a whole new website section for it i'm sure.

 

I believe somebody mentioned that MXM 3.1 (which we currently use) can draw 120W power, however it might not be possible to support two 120W cards + a 55W CPU without making the power brick beyond massive. Limiting the 680MX to machines like the P170EM might work, as they can only take single cards and 1 680MX --> 1 680M, however 2 680Ms ---> 1 680MX for the P370EM model... but then how much would a 680MX cost? Would it be worth it? The 680M is already VERY pricey, and lots of people choose the 7970M over it in spite of AMD's less than stellar driver performance and previous Enduro headaches. In fact, you can get two 7970Ms for only $100 or so more than a single 680M in a P370EM, and the possibility of buying a single 680MX (without the chance to SLI) in a P370EM should be possible too... *head in overdrive*

Sorry, just my thoughts rambling.

 

According to Clevo's site, the P370EM, the P570WM, and the P170EM all use MXM 3.0b.

 

I think that's a little outdated though... I've seen their site say that their machines are limited to X amount of RAM and X range of video cards and on release resellers and rebranders have made more RAM or different video cards available. Isn't MXM 3.1 the PCI/e 3.0 version of MXM? The same person that said they had allowed 120W per card was the one who had said that they change from MXM 3.0b to MXM 3.1 was PCI/e 2 to PCI/e 3. I just can't remember who said it or where, but it was on this forum... somewhere.... Not necessarily on this thread either.

 

first one to give us detailed infos on this was mythlogic

and yep its true, pci-e jumped from 2.0 to 3.0 on the latest mxm revision. i have seen the claims about 120w tdp raise, but not sure about that tho...

edit: there u go:

http://forum.notebookreview.com/sager-clevo/650077-mxm-standard-3-1-new-one-up-3-0-a.html

no mention of any tdp raise as i remember. so who made that one up?

Sent from my Galaxy Nexus using Tapatalk 2

 

But do the new notebooks use the 3.1 standard?

 

yup, the EM series sports pci-e 3.0, so its mxm 3.1

Sent from my Galaxy Nexus using Tapatalk 2

 

There was no change to tdp as far as I was aware and it makes no difference if the motherboards are not changed.

 

Huh. Guess I'm wrong then. Good to know! Less chance to make myself look like an idiot elsewhere XD

 

The 100W limit is flat out a myth anyway.

 

what makes u say that?

Sent from my Galaxy Nexus using Tapatalk 2

 

Real world usage has proven it. Look at every overvolted and overclocked GTX 580M. MXM 3.0 has no problem with giving them the power they need, even though they draw waaay over 100W. And they also prove that the high-end laptops have cooling solutions which can cool such 120W-130W power draws.

 

Ok so here is the official word (we are part of the MXM group)

Is there a HARD rule on 100W? Nope, you can design whatever crazy GPU you want to as long as you can power it and cool it, however, there are HARD rules when it comes to thermals, including on the backside (which is NOT allowed to be cooled per MXM spec) and so its the thermals that really limit how big of a GPU you can put in there.

So thats why the recommended design target is 100W as that will keep you within the thermal spec easily. The electrical maximum on the MXM cards is IN THEORY 20V @ 10A (DC) which is WAY above 100W

 

Well MSI ignored the back cooling for their 680M models so even those rules can be bent....

100W is what the current cooling technologies are designed to handle, 125W systems would look quite a bit different, extra heatpipes and more aggressive fans/more fin area would be needed.

 

Maybe better cooling materials, like Nickel-Cooper instead of pure cooper... ?

 

Well every material has downsides, cost, weight or heat dissipation. If there were better materials we would see them in high end desktop coolers.

Nickel plating is mostly for aesthetics.

 

So, guys, to sum it up:
- new mobile CPU's will come end of june
- they won't be much of an improvement compared to Ivy
- 2W higher TDP (no idea why)

= gpu's wont arrive until october?
= new barebones will use the old series gpu's?
= new gpu's won't be much of an improvement too

Well, seems like its pretty much a failed year for hardware. No big leaps forward. Nothing interesting coming out in the high-performance notebook sector.



Also, our forum is getting terribly spammed by what seems to be random advertizing for women's jewelry and footwear.
Which makes me even more sad.
I feel like I should cry.
Im going into a little corner to cry right now.

;(

 

It's 2w higher because the vrms and some maybe some of the other stuff that on the pch/chipset get moved to the CPU. Overall TDP should be the similar if not lower.

 

But I've seen that the high-performance water-cooling solutions are usually Acetal-Nickel:
Aquatuning -Your watercooling and modding distributor - EK Water Blocks EK-Supreme LTX - Acetal+Nickel CSQ (Sockel 775/1366/1155/56/2011) EK Water Blocks EK-Supreme LTX - Acetal+Nickel CSQ (Sockel 775/1366/1155/56/2011) 10443

I have some slight amendments:
- new mobile CPU's coming in laptops at the end of june, CPU performance increase is around 10-15% higher than what we have now.
- new GPUs will probably arrive sooner, but yes, the 780M will be more of a rebrand with 10-15% performance increase over 680M

Overall, the new laptop series coming this summer will be between 10-20% faster then the current EM series.

 

Chances of Clevo adding a manual switch:

0% or Zero percent?

 

You want the truth or do you want us to lie to you

 

What's a manual switch?

 

Being able to directly tell your machine to use only GFX card or Processor Integrated GFX without relying on the automated magic (Optimus/Enduro) which can be hit/miss.

 

I want you to manually add it because somehow Changeman were Legendary Beasts and therefore you being Mythlogic can do it.
I don't know how that makes any sense, but it's my reason for now.

 

Manual switching requires direct PCB design, so yeah :s

 

Trust me if we could do it with a soldering iron we would have done it by now

 

Can you guys get back to Clevo with a "Listen, your customers want this feature. They're screaming for it. Please add it." ?

 

It does add cost and increases power use...

 

Alienware bundles their notebooks with a switch that controls wether you want to use the IGP or the dedicated GPU. Much less prone to errors.

It looks like this. You get to pick if you want a red, white or black one. Mine is white since my wall is white too. The switch in upward position is IGP. Flip it down, its the GPU.



You have to put it on the wall just like any other electrical switches. Although you get less errors, you still have to deal with ton of wires going back and forth on the floor from the switch to the computer. I don`t see whats with the all fuzz about it. I took a picture of how it looks like from the switch to my computer in my room. The white ones are the power supply, the blue ones is carrying the picture to the monitor. As you can see, they take a lot of room there on the floor. It works though, I don`t have to deal with Optimus or Enduro.



Some smart people have found a way to put the wires under the floor to the computer. Not everyone have room for that. Lucky people

 

Cloud living up to his title there

 

A year and a half ago or so a reseller stated on the forums that the switches cost about $10 and thus were not pretty costly, however I didn't know about any kind of power drain increase. How much does it require? And can you verify the price for the switches?

 

You would have to ask clevo since they are the only ones who can implement them. Also remember every display output has to be muxed.

 

AFAIK that's GPU manufacturers who fight against muxed HybridGPU implementations - it proven to be a headache for them to support, since every manufacturer devised own contraption, but then people came screaming to them that "your driver's not working/crashing/etc!"

 

Well you could maybe get a hybrid setup like our new machine which has the internal and VGA wired to the integrated and the HDMI and DP hooked up to the dedicated card letting you be optimus free on the external at least.

 

This kind of workable, but still will be somewhat difficult. Unless dGPU will carry own extended desktop, there will be no way to easily & seamlessly "transfer" primary desktop to external monitor when you plug it in - without reboot/relogin, mainly because it will require to change the GPU which will handle desktop rendering.

 

The transfer is seemless and is done on the fly automatically The primary render target must just be the non optimus display so make the external primary and you are good to go. No special drivers. No restarts. No optimus.

So if sager/clevo go that way it's all in place to be used now.