New details about Nvidia`s Maxwell

I completely agree, especially considering that was with Turbo Fan on.... on my GT60, turbo fan drops my GPU temps substantially. Granted, MSI has a very slow non turbo fan speed. For comparison, my 780m runs mid 70'sC with turbo fan on all day long while under heavy use.

This is my concern and reservation for ordering a Razer 14".... if only they stuck with the 860m Maxwell this generation...

 

Sorry, I thought you were considering the 15" laptop not the 17". The 15" Clevo's only get a 180W PSU. The 17" ones get the 230W, and you should be fine.

But it makes no sense to me why Clevo do this because the only difference between the two is obviously the larger screen, and the 17" offers two 2.5" bays, 15" only one. They should both get the 230W PSU IMHO.

 

No worries. I know you were trying to be helpful (as always). :thumbsup:

 

I'm considering this same laptop now. But I think I'll wait a few more months and see what comes of Maxwell for high end cards.

 

I wonder if alienware is gonna change up their lineup , its highly unlikely they are gonna have a completely different design but maybe slightly thinner

Sent from my LG-D800 using Tapatalk

 

Or maybe at least make the 14" a true 14" and shave off at least half an inch of height and lose at least 2lbs. Seriously that thing is a monster compared to the competition.

 

It really needs to atleast have a MXM GPU considering it's size

 

Or they should give it double battery life for its size.

 

Or both make it thinner and increase battery life by, say, 50%. (I don't get the current AW 14 - there are 17" notebooks with around the same gaming prowess and a borderline lighter package, not to mention the Aorus x7...)

 

I may finding myself making use of the MXM feature if there is a huge difference.

Sent from my SPH-L720 using Tapatalk

 

Nice temps but turbo fan is a joke..

 

Although they could do something thinner, Dell is going to sit on their recent design until Broadwell. I'm guessing early to mid-2015 until we see anyrhing new.

 

I hear nothing...

 

It's because you don't have a MSI with outstanding dynaudio !! (jk)

 

Same can't hear a thing .

 

i edited the video, sorry my bad

and thats 50% only..


or for the GE series :

 

"Golf tango six zero, you are cleared for immediate takeoff, over and out."

 

octiceps : exactely what I said haha !

 

Reminds me of this video I saw a few years back:

 

Stick your finger in it

 

GE series is badly designed.
Small fan, few thin heatpipes to the CPU and GPU.

GT series is the only ones that cut it. The fan is pretty beefy and can move a lot of air. The WHOOOSH sound this beefy fan makes isnt so bad actually.
I owned the GT70 with 680M and I never had to use the turbo fan mode. Neither did it go in full speed automatically either.

Another story with the GT60/70 with 1GHz 780M (880M) I guess. I don`t think the cooling system in the GT series isnt capable for that sort of heat

 

Lets hope so, because they are the first to release their 870M/880M enabled laptops

 

Yeah, that's been an issue since I can remember.


<iframe width='480' height="360" src="//www.youtube.com/embed/PFZ39nQ_k90" frameborder='0' allowfullscreen></iframe>

 

GT70 with GTX 870M
86C for 870M and 82C for the CPU

GT70 with GTX 880M
89C for GTX 880M and 96C for the CPU

All running a simple 3DMark Vantage test.

LOL
MSI GT70 Dominator & Dominator Pro with nVidia GTX880M, GTX870M Review Benchmarks Unboxing - YouTube

 

That's in US, come to see what it will be like in Tunisia (ambiant temp in summer ? 34°C and it's not the max)

 

Extra airblower required. Or running that turbo fan mode all the time

 

is it recorded with 7.1 surround sound? i should listen to it with a proper headset to understand just about how loud the thing gets

 

Maxwell chips are gonna make the heat go away for a while but I think nvidia will push to 100W TDP to create the refresh high end GPU for laptops !

 

I'm drooling at the idea of how powerful a 100W 20nm gm204 or gm200 will be ....

Sent from my LG-D800 using Tapatalk

 

Now imagine how powerful a 20nm 250W desktop card would be. THAT would be something truly worth drooling over.

 

Where I live (Chennai, India), ambient temps are 37-38C, the GE series would replace an induction cooktop.

 

Any answers?

 

Hi guys!
I'm just one of the thousands who follow this thread for the latest information about Maxwell. You've really got some knowledgeable people here.
Anyway seeing that we were veering off topic slightly from gpus to the origin of human life, just thought I'd contribute my two cents and clear the matter.
I work often in ambient tempratures of 50-55°C in 100% relative humidity with NO VENTILATION.
Despite sweating like a pig and despite "wikipedia" I'm fine.
Hope that settles it.
Cheers !

 

Back on topic.

The statement below was based on reading old articles... sorry.

<strike>I have read a few articles about TSMC and Global Foundries having difficulty getting 20nm FinFET process to work. nVidia needs this process to get the performance out of Maxwell. If the fabs can't do it, then Maxwell isn't going to double the performance of Kepler.

Thoughts anyone?


</strike>

 

Great. So now Nvidia wIll probably have to pay Intel a license fee to use their tech.

Beamed from my G2 Tricorder

 

I just realized I was reading articles from mid 2012. Here is something a bit more recent and from the horse's mouth.

TSMC Tweaks 16nm FinFET to Match Intel - Electronics360

 

20nm FinFET won`t exisit. Just a normal 20nm process.
They will go straight to 16nm FinFET which is already being produced at TSMC.

Worst case scenario is that they skip 20nm and do Maxwell on 28nm and do 16nm FinFET GPUs next year. But 20nm isnt a midstep they skip like 22nm was. And 16nm will be even more expensive than 20nm and harder to produce, so the theories involving cost and yields about 20nm means we never get any new nodes.

Nope. Nvidia and Intel have made an agreement a couple of months back to do cross licensing. Intel get to use Nvidia`s licenses, and vice versa.

 

You were the one saying he could not survive at that temperature without stating that the humidy there might be taking effect.

Now that we have reached the proper conclusion however: (yes I am a stickler for accuracy, though usually I am corrected tech mistakes, not biological)

Intel was the first to launch finfet but it did not originally come up with the idea AFAIK, so it's not their technology as such just something they figured out how to use before anyone else.

The 20nm node should be exciting enough by itself for now though.

 

So why not use 22nm finfet instead then? It clearly works for Intel.

Beamed from my G2 Tricorder

 

exactly, we wont be getting any new nodes cause it is too expensive for our all mighty lord, Nvidia..

28nm and above are here to stay for forever. And u shall\ be grateful that our lord even blesses us with such a wonderful joy as purchasing new GPU!!


who even said it?

 

Because the only company that is physically capable of producing 22nm chips is Intel and they're not interested in selling that production capacity to other companies right now.

 

Be careful sir, once you down that path of "good enough" you may never find your way back

 

I don`t think Nvidia pay anything to Intel regarding semiconducting. That is up to TSMC and whoever owns right to any process manufacture method and designs. Nvidia just pay TSMC whatever the wafers cost, silicon, production cost and IPs and everything included.

As Alternai pointed out. I don`t think TSMC had the equipment nescessary to produce 22nm FinFET, Intel is ahead here. And maybe it just wasn`t worth it, performance and cost wise. I don`t know.
The good news is that TSMC is ready with their 3D FinFETs and will be using them on 16nm. That should give some benefits. FinFET is pretty interesting since you get bigger current flow and less leakage. All in all a more efficient chip.

But we are talking maybe late 2015, early 2016 here. Intel will probably be at 10nm when its available. Intel is in a pretty sweet spot. Ahead in semiconducting industry, and ahead in CPU designs. And now managed to beat AMD in their graphic APU game. No wonder they make a crapload of money each year.

They could probably sell just as cheap CPUs as AMD and still make a much better profit than AMD which is fab less and have to pay for wafers, but why bother when you are so far ahead? Sweet position indeed

 

Volta will be dominant at this time in 2016-date of release.

 

Hey guys, nice thread and a lot of good info(though a lot more speculations ). Please keep it on topic.

 

It will for sure be very interesting when that time comes
I wonder what "stacked VRAM" can do...

 

Tomorrow could be the day where more is revealed regarding Maxwell

Watch LIVE STREAM here:
Twitch


Here is some of what Jen-Hsun Huang shared with us last year :thumbsup:

Spoiler

<iframe width='560' height="315" src="//www.youtube.com/embed/BYJ1-XQzHx4?list=PLZHnYvH1qtOY0ZrWQgnQlj4dwGZ1pVgoj" frameborder='0' allowfullscreen></iframe>

 

Actually, I have heard that Intel has excess high end (22nm? 14nm?) capacity at the moment and is allowing certain companies to use their fabs.

 

I think there should be something about Maxwell, not 100 % related to 800 announcement but main list of new Maxwell features, like with dx12.

and of course the gpu (yes, I know , no need to tell me) everyone of us has been itching to buy, 790!!

 

Not their direct competitors: Nvidia and AMD obviously