Clevo notebooks with 800M series coming out February 2014

Actually that's not my point at all. My point is that all this wait and wait for new tech and then it's purchased and you hear more bad than good things about it till it's fixed. So, just because it will be 20 vs 28 does not make it better till we actually get to see it in action. And you get allot of people talking about waiting all the time. If you wait 6 months something new will be out and so on and so on. It's been like that for years now. One person on these forums hasn't bought a new laptop in like 2 years. LOL. And is still waiting.

As to the review of the GPU only, not the machine it's in is something different.

 

Haha, Newton, nano, it's all the same! ;-)

 

Yeah, but tech leaps forward in huge steps at set times, so it does make sense to wait for that new tech to hit - architecture & die shrink (not 680M - 780M!). So, in my mind there's nothing wrong with waiting for 2 years, considering that's about the length of time between architectural changes & die shrinks. I think upgrading at the right time is sensible, the worst time to upgrade is to buy it just before a new architecture is released (which is now!).

 

Yes robbo99999, we heard this same argument 3 months ago. Wait for the 800 series to hit before buying.
I'm also not saying not to buy. I was going to buy a p570wm3, but why would I buy that with 780m's in it and I already have them in my Alienware. So I played the "waiting" game. Nothing wrong with this at all. What's wrong with this expectation is that it will not be 100 percent increase in performance. Or a better break down. 5 fps better than what we have now. Then you just waited to brag about having new tech that is 5 fps better. And if the game is a 100+ fps game, who cares about 5 frames more?

I would love to be wrong though!!!
If it can get any where near this, i'm all for it.
View attachment 109115

 

Hmmm, I don't really know what you're talking about, that's just a jumble of mixed messages to me, I don't really see that you're making any particular point there. So, don't really know what to say to that. I'll just say again that I think it makes sense to wait for high end Maxwell chips if it's possible to wait, because it looks like it's just a few months more wait. As we know technology leaps forward in a large step with a change in architecture & a die shrink, (not talking about the 680M - 780M - 880M small improvements on the same architecture & process size - especially if you release their overclocking potential with modified VBIOS which further nullifies performance differences between those chips).

 

I guess I never really *expect* a new tech 128-bit GPU to even compete with an "aged" two generation old 256-bit GPU, but it can come darn near close if what we're seeing about the 860m is true. The reduction of TDP is the most prevalent factor. Even coming close to 680m performance with half the TDP is quite stupendous.

I much appreciate your testing though johnksss, it definitely brings us down to reality, and absolutely worthwhile for users debating on whether to upgrade or not. I'm just looking at that performance and only wondering actually *what* a die shrink will offer. Even if it's "only" 30-40% you're looking at a single GPU performing close to 780m SLI at stock speeds.

 

I heard that the W230ss will have a glossy screen...

Clevo W230SS Notebook Display Replacement | LAPTOP & TABLET LCD SCREEN FROM $34.99.

 

Good find, although the stock screen will be 1920x1080 and matte.

 

It's simple. When it comes and it's not all cracked up to what you "thought" it was, don't be surprised. That's not rocket science. This is what we have been dealing with for years now. And it's the same old story. And now we come to maxwell. I take it you don't get to test allot of gpus? Sorry but im not easily impressed by paper stats. Until it's out in the open and we can see it in action, it's just that. Paper stats.

Im just hoping these paper stats hold true and the 870M is the best bang for the buck GPU.

I'm looking for that next "big time" jump in performance. So far from what we have found... The 780M is only 5 to 8 FPS behind the 880M. Even if it's a rebran, it's still an upgrade for those running anything lower than a 780M, but for those of us already running a 780M....Not so much. Best to really wait. And as of right now, the 880M is running on broken drivers and vbios. Curious to see how clevo addresses this out of the box. Been playing an older game. Dead Island and it's still down clocking with newer drivers. With older drivers it doesn't down clock, but it sure does crash while gaming, But not while benching. We have tested this with the AW15 and the AW M18xR2. Need Nvidia to come through on a decent driver.

Thanks, but we are waiting on you to test it on a clevo. LOL.

 

Awwwwww... Need to look for matte screen protectors.

 

Oh yeah, no doubt it's all theoretical discussion at the moment - that's a given. But there's no denying that Maxwell increased performance per Watt on the same process by 100%, and that's not even including a die shrink to 20nm which could theoretically nearly double the performance again (for a 400% increase over what is currently possible with Kepler 28nm). Given all this potential in the Maxwell architecture and the imminent likelihood of a shrink to 20nm, then it's definite that we will see a LARGE increase in performance from Maxwell high end GPUs. Which is why it's not particularly wise to upgrade to a Kepler GPU at the moment - this is my point.

(No I don't test a lot of GPUs, but I follow the industry and the specific cards (their specs & performance which I compare & contrast) on Notebookcheck.net, guru3d, amoungst other sites too. So, I don't think I need hands on experience (like you had with your 880M) to draw sensible conclusions.)

 

Coming soon to a forum near you.

 

Ok I read a bunch of pages here and there and this is my conclusion (take it with a grain of salt like all the other information out there)
w230st 765m gpu: 75watt TDP - ~4400 3dmark 11 score
w230ss 860m gpu: 60watt TDP - ~5300 3dmark 11 score

I just hope the laptop actually comes out this month, I've been needing a new laptop for a long time (xps 1210 - 7400go).

 

Hummm, Then I'll close with this.
Everything from the 580M up to the current GPU. Including your working GTX 670MX 3GB (@1.05V, 1124Mhz Core / 1150MHz VRAM Is working because of mine and svl7's work. Not Nvdia. Not Clevo. Not Alienware.
So Im thinking I know quite a bit on the workings of "what we are told" vs what we get and "fix"
I understand you want to sell people on maxwell and that is completely fine. I hope it's the next 8800M GTX, but if it isn't I will be right here to say. "I told you so" The same exact same thing was said about Fermi and Kepler. And the only killing each did was to AMD (In the mobile world). Not to each gen of Nvidia mobile GPU's. Hopefully AMD will come around again.

 

more 5800 than 5300 with the high clocks and 40W max with low voltage ram

 

Oh, come on please, you don't need to blow your own trumpet on the modified VBIOS front, and you don't need to add childish phrases like "I told you so". No doubt I'm very thankful to SVL7 for his modified VBIOS (and perhaps your contributions that I'm not wholly aware of), but that's not what's being discussed here, my points are referring to stock performance (also happens to apply to modified VBIOS performance too for the most part regardless). I'm not pushing any agenda with regards to Maxwell, and have no invested interest in it's success, but I am interested in it & I've drawn my conclusions from the info I've seen & the things I've learned. I'd rather AMD was more competetive too, and I hope they are too - competition is a good thing! I'm not going to bother reiterating any past points I've made to you as you seem a little loose on the comprehension front, and a little loose with formulating your arguments and keeping it focussed & relevant! It's become a waste of space.

 

I wouldn't complain if the real world scores are nestled anywhere in that range, if it gets above 5100 3DMark11 bench scores then that will be a winner in my book.

 

No worries, you don't understand. All good my friend. No need to get extra wordy.

 

You can think that if you want to.

 

That is what you can expect at stock.

Vs this at stock

 

Nvidia designed the chip and pcb capable of those clocks, whatever we do pales in comparison to the design and engineering that goes into these things.

 

To bad it never really works when it gets to us. Kind of the main point.

 

But it does work at stock speeds. It's only if you want to go over stock that you need to tweak things a little but that was done on purpose.

 

lol johnksss. Why are you so pessimistic about Maxwell?
GM107 was designed for mobile. They concentrated on lower the TDP (by half) and just increasing performance slightly (+25%) on desktops. If they can just use that TDP headroom which become available to build meaner chips for notebooks, we are in for some pretty hefty performance jump with Maxwell

Its like Maxwell is perfect for notebooks with our constrained TDP room

Also I can testify that my 770M`s can reach GTX 680M performance with new vbios and some hefty clocks (because 770M at stock is already overclocked sort of), but GTX 680M is clearly better when you start to overclock them again

 

Ummm, sorry to say. No it does not work at stock speeds. So if you were to open it up and start trying to play a game, it would down clock to non playable frames.

Not like that Cloudfire. I'm hoping it's all that and then some! We need this, but as with the last 2 versions the same talk was talked. And many people had to go back and retract prior statements made based off of "paper stats"
All i'm saying is...lets not go jumping the gun till it's actually here and proven that that's what it will do. So people can make an educated judgement call at that time.

The 680M was used as an example to say...It's 2 years old and a top tier card, but it's still plenty playable. So that generation of people are still getting their moneys worth. And should not feel the need to upgrade if they don't have to. This now applies to the 780M. And at the end of the day. For high-ed buyers.... TDP goes out the window the second you raise the clocks. Which about 99 percent of people in this tread will be doing. You only need to look at half of the signatures to see this. The other 1 percent will be trying to under clock....

 

Ah I see. Yeah, GTX 680M still holds its own even today. There have been several thread made by people who didnt know what to buy and I recommended GTX 680M since it only sell for around $500 now. It used to be $800-$900. Plus I know you are aware how much more you can gain with just a new vbios made by you guys

I know I`m buying the first high end Maxwell and sticking with it. The rest will just add heat and will have less overclocking headroom anyways. It seems like the most reasonable way to go. GTX 780M is nice too, which you have shown by your many benchmarks here on this forum. But I`m not as extreme as you and mod my own PSUs and stuff hehe. I`ll leave that to you pro`s

That said, GTX 680M start to show signs of aging. According to notebookcheck, there are 4 games it can`t play maxed out, and several in the danger zone as to what is playable FPS. We need better cards. I so hope GM204 will deliver this year. SLI two of them will be major overkill for 1080p. Would love to see higher res displays than 1080p on the next Alienware 18 but I`m not holding my breath

 

Im in agreement. It's aging this is for sure. And with many 680M vbios revisions it's come along way.

And will be right here when you do buy it!

As to the 780M it will run fine on a single psu with normal gaming. And of course, you can always control the TDP.

Will be loading the Witcher 2 & Crisis 3 & Metro 2033 soon. Since these 3 games are by far the most grueling of games for cards. Unless you have some other suggestions?

Also to add...would love to play at lenovo yoga pro 2 resolutions(Highend Version)! That will be the day that we have done something serious. LOL

 

Just a question, is it possible for a highly underclocked/undervolted dedicated GPU (such as 765m or 7870m) to consume less power than Haswell's IGP in web browsing and movies?

I don't really like using Intel's GPU drivers...

 

intel igp eat nothing when idle, dGPU will eat more power anyway..even if underclocked to the death (they already are when idle you know)

 

I'm with johnksss on this one. For the last couple generations Nvidia has been inexplicably locking down and gimping the capabilities of both their desktop and mobile GPU's, especially concerning overclocking and overvolting, e.g. Green Light Program and +135 MHz core overclocking limit. While we all truly appreciate the time and effort spent by people like svl7, johnksss, Prema, Saltius, and Mavke among many others to undo Nvidia's mistakes and poor decision-making with their BIOS/vBIOS mods, it's unfortunate that this is what it takes to unleash the real potential of our GPU's.

And as johnksss was alluding to, even at stock sometimes these Nvidia cards underperform relative to their stats on paper. Until Nvidia can prove to us that they will release products that won't limit or hurt the enthusiasts, or make purely profit-motivated moves like purposely gimping FP64 and compute performance on their "gaming cards" so more people will buy the much higher-margin Quadros and Teslas, then I remain skeptical of Maxwell.

 

I think the only thing that's going to hold Maxwell back from all the theoretical promise it can offer, is if NVidia choose to milk it slowly by only gradually releasing cards that are only a little faster, without exploiting the full potential of the architecture immediately. I do still expect a good step up in performance of the first high end Maxwell cards though (the replacement of the 880M), and then I would expect them to release gradually faster cards over the following 2 years, like they did with Kepler for 680M - 780M - 880M. I think it's sensible to be optimistic about the performance gains though given the 100% performance per Watt increase of the Maxwell architecture over Kepler (on the same process), and the quite soon to be released move to 20nm for even further potential improvements. I think this all adds up to more promise & potential of performance than we have seen in the past from a step up to the next architecture.

 

I'll say this...We already are going to set the bar mighty high for Maxwell from the get go. I guess we now find out if it gets delayed over time. lol

 

I do wonder if these improvements in power per watt will look as impressive once overclocked. Haswell shows great efficiency at stock, yet it's not the case once OCed.

 

Is that due to the innefficient thermal interface material between the core and the metal casing of the CPU (if talking about desktop), or is that due to the higher density of heat over a small CPU die that prevents efficient heat transfer to the heatsink? Or something else? Hopefully the same won't apply to Maxwell on the density of heat point.

 

Based on past history, this is exactly what they will do.. unfortunately.

 

But can we blame them? Remember the cost it involves getting 20nm up and running, the complicated designs, not just one, but many designs that involves everything from ES to QS to final retail design. On many chips. The marketing cost, the silicon cost, r&d, money to invest in next node jump etc

If they just released the meanest desktop and mobile GPU the day Maxwell was announced, everyone would buy them, and people would have to wait until 16nm. Not just because we wouldn`t actually need anything better, but because Nvidia couldn`t offer anything better until they can shrink the transistors further. Which takes research, new equipment, mature design which involves trial and error with bad yields at first, meaning several years to get it up and running to be able to produce it in mass scale for the public.

I would love a GTX 980M in July, but I know Nvidia and AMD need to do this the slow way to make a profit and keep themselves afloat. After all, its a pretty big company with many employees that need to get paid too. You can`t just design a new architecture with a manforce of 5 people

 

Let's be honest it's not about the employees that need to get paid, it's about padding their pockets. I agree they need some cash for future capital investment, but sometimes they just milk us a little too much.

 

Well yeah, its not like they are not making a ton of money

But I think its about finding the right balance between getting customers new exiting products and not letting them feel ripped off. As well as keeping the investors and stock holders happy. Lots of complicated stuff that anyone here is not able to fully understand I think

 

A lot of it is honest production/cost limitations.

 

Any new info on whether GTX Maxwell MXM cards will be compatible with current laptops with GTX Kepler MXM cards?

 

Hm. I wonder how many months it'll take for AMD to offer something that would compete against Maxwell.

Or are they betting on APUs at this point?

 

Maybe. But if they do, they really need to make some bigger chips, and solve their single thread CPU performance problem. Even if they do come up with some graphically fast APUs, with their current CPU design it's no-no.

 

I wouldn't exactly call $500M annual profits a pittance either. I'm not object to companies making money, that's how they survive. But sometimes when they gouge customers to make that profit especially in an oligopoly, is my concern. But I'll leave it at that.

And stop limiting their low and mid end CPU's to single channel RAM. Their IGP's are powerful, but as you stated, single and even dual thread performance is pathetic, and gimping GPU with single channel RAM is even worse yet.

 

There were a few APU laptops where they had 1333mhz RAM soldered in, so you don't even get the option to upgrade to faster RAM, all in the name of saving $5 to $10.

But my main gripe was that the laptop manufacturers frequently put the APUs in plastic 15" laptops that weigh 5lbs and with mediocre 768p TN displays and bad keyboards/trackpads. Not much AMD can do about those terrible laptop design choices.

 

performance wise i don't see a huge boost.

it;ll probably be a nominal difference.

 

Now compare that to intel.

 

Intel's in the same boat. It doesn't matter if you steal a TV or a car, they're both the same crime of a different scale.

The *ONLY* reason nVidia and Intel actually make progress and make chips semi-affordable is simply because AMD exists. I hope they never fold or fail because we would be in a world of hurt in the computer world. It's too bad other CPU and VGA chip makers died. Having three options to choose from would make the wheels turn more smoothly and for the better of the customer. Sometimes it makes me wonder if our advancements aren't stagnated because of this.

At least in the world of mobile devices you have multiple GPU vendors and many possible options.

 

Unconfirmed die shot of 870m. Perhaps it has some info for some of you.

Mobility Geforce GTX 870M Chip Exposed - 6GB and Different Codename, Are we looking at Maxwell?

 

tlprtr, I hate to be the pessimist but if the 870 is maxwell it will be so efficient and cheaper that it will make the already useless 880 even worse so it is very unlikely...

 

It's nearing the end of FEB. I've been waiting xD