Dropping the M doesn't mean anything. We have had full core chips for years now, just with massively reduced clock and memory speeds. From the sound of it, they're just going to drop the M and confuse everything. Imagine the nightmare of trying to get the specs on a 1080 notebook if there is no designation for the difference between the notebook and desktop versions. It's just going to confuse customers that don't know any better. It also sounds like boost bins are going to be much larger than they currently are in an attempt to "close the gap" without actually closing anything. It's a fancy way of marketing an inferior product. It's a page ripped right out of Intel's play book. Lock a product to a set TDP and say it can go up to X but it never really gets there. My 4940MX is a great example. 3.1GHz base with a 3.8GHz 4 core boost that out of the box at the stock 57W TDP and voltage actually throttles down to 3.4-3.5GHz when under a heavy load. For all intents and purposes my chip is a 3.5GHz chip that cost me 1k.
-
I would imagine the Pascal power use at clock rates we want to use shows that it requires more power and more heat dissipation than the Maxwell cards have.
The current laptops that can take the 980 desktop mobile laptop GPU, will need to improve power and cooling for the 1080 desktop mobile GPU(s).
It's likely all of the current 980m/970m laptops will need to improve power and cooling, or be forced to throttle the 1080m/1070m potential.
There are spots for both a full desktop mobile 1080 and 1080m/1070m, all cut from the same core, but tuned to fit the build - the differentiation of naming them 1080 / 1080m / 1070m / 1060m / etc helps buyers know what they are getting.Last edited: Jun 2, 2016transphasic, bsch3r and Georgel like this. -
GTX 1080m as a GTX 1070ish card is perfectly fine for most people. Why waste perfectly good fully enable dies when dumping your broken ones works. Especially you can sell them at the same price.
Mobile Maxwell is probably treated as the garbage dump for seriously messed up GM204 cores.hmscott likes this. -
Yeah. 780m was a full GTX680, with lower clocks and mem clocks. However, I do think they can easily do away with the whole M rebrand of mobile GPUs and simply name their GPUs accordingly. Lower end, thin notebooks could probably use up to a 1060, which should be something like GTX 970 performance at low enough TDP.
I mean, polaris 11 will bring GTX 950 performance at 50w or so right? I am sure nvidia has a slight edge on per/watt, so they can make a 50w GTX960 (close to 970m) etc. -
Right. There will have to be at least mid-range GPU's like the current 950m and 960m. Unless they plan on continuing to use Maxwell architecture for those, but the power savings alone for a mid-range chip would be perfect for the smaller, lighter, thinner notebooks. They likely just mean for the high end chips like the 1070 and 1080 they will not be called "M". Although that's dumb because they are mobile chips and it just designates them as such regardless of whether they're cut down or not.
-
OMG, dropping the whole M lineup ? WTF. No mobile pascal at computex and this now.. I'm just speechless, RIP upgradeability ?
MSI's statement on upgrade kits and new machines with SLIs with upgradeablity proves it otherwise, this is getting more dark now..I just want MXM 3.0b cards FFS !!Last edited: Jun 1, 2016jaybee83, Papusan, TBoneSan and 1 other person like this. -
They could just drop the M off, call it the 1060 but use a different chip from the line compared to the desktop for example. Marketing trick that's not really illegal, just extremely deceptive and something I wouldn't put past nVidia.
I think we are slowly getting to the point where user upgrades are going to disappear completely. Consumers are keeping their laptops longer, especially with the lack of processor innovation, so the easiest way for everyone to make money is to solder everything.jaybee83 likes this. -
Sounds interesting? Only 450w psu for Sli
Maybe the cards, but not the whole machine
Ashtrix likes this. -
Man, I guess I can't upgrade my NP9377 GPU then...
-
It's not yet confirmed, but I'm worried a bit..
Tweaktown said the mobile 1080m and 1070m are going to be shipped soon, PCgamer said 1070 /1080 are coming soon, I'd say cramming a 200W & 150W GPUs without proper cooling would rekt the machine apart and the high power CPU too, So the advantage of BGA is observable, soldering the machine with those high TDP parts is going to pose risk.
Clevo, MSI use full MXM (Maybe ACER too ?, Predator, ASUS uses MXM but custom, still it's not full BGA rt) and I highly doubt they'd go the BGA route given the MSI's guarantee of upgradabilty, IIRC MSI is a keyplayer so I believe we will get the MXM3.0b maybe Nvidia was waiting for the AMD Polaris's notebook targets, but the AMD failed imho or a new revision if they are thinking about G5X (and HB SLI mobos for mobile ?) but afaik the 3.0b should suffice the bandwidth of G5X.
So maybe we will see the Pascal MXM machines soon...If this goes south I think the whole Clevo user base will go to DT route, I won't surrender to the BGA filth anytime either.Last edited: Jun 1, 2016 -
If nVidia is going with custom boards ala 980 notebook edition, the MXM cards are going to be underpowered in comparison just like they are now with the MSI machines.
-
Me too. I just want my NP9377 to be GPU upgradeable. I guess will find out on August/September.jaybee83 likes this.
-
Maybe they are just going to go with lower TDP desktop chips for the entire line, sans the "M" nomenclature.
GTX 1080, 1070, 1060, 1050; GT 1040, 1030 etc
This is something we've wanted for years.
MSI has a recent statement on upgrade kits,, or are you referring to the original promise they made GT72 owners?jaybee83, Cakefish, Robbo99999 and 1 other person like this. -
I knew that they were offering upgrade kits..Is there any regression now ?Last edited: Jun 1, 2016
-
"According to our contacts, Nvidia is readying the release of its new 10-series chips for notebooks. The kick is, they won't be M versions of desktop GPUs. T hey will be the same chips used on the desktops, just operating at a lower TDP—we're told there will be the same number of shader units, etc."
"With the incredible power efficiency of the new 10-series GPUs, it makes sense for Nvidia to drop its M series GPUs entirely. GTX 1080 has an 180W TDP, uses only one 8-pin power connector, and GTX 1070 only has a 150W TDP. This is a huge drop from GTX 980 Ti's 250W TDP, and we've already seen Nvidia stuff a full GTX 980 (GM204) into notebooks."
This is going to be confusing, the performance will depend on the notebook vendor to squeeze as much performance out as possible given the power / cooling constraints.
If the vendor builds a laptop model with a 1080 that runs at full desktop 1080 performance, and builds a slimmed down laptop with power / cooling and still calls it a 1080, this could cause some disappointed buyers.
It's like the i7 ultra low voltage (ULV) vs i7 full quad core desktop vs i7 quad core laptop CPU's. Everyone wants an i7, but ends up getting a lot less than expected if they aren't paying attention. And, price isn't a factor, I have seen ULV i7's go for more than notebooks with i7 quad core laptop CPU's.
Everyone gets a 1080 with an i7!!
An Nvidia announcement could be forthcoming now...Last edited: Jun 2, 2016mason2smart, jaybee83, AngryPeanut and 4 others like this. -
That was exactly my point. This is just a move to introduce confusion. Larger boost bins that may or may not be achievable, ala Intel.jaybee83, Ashtrix, Papusan and 1 other person like this.
-
Well this could be interesting. Hopefully this means we won't have to wait much longer.
-
FWIW GTX 1080 + [email protected] FireStrike 1.1 at 100% TDP default clocks (1850MHz / 1797MHz Boost) reports about 60% TDP at 1600MHz 0.8V or ~110W. Voltages are programmable for frequency steps (well sort of, actually frequency is programmable for voltage but that may sound confusing).
-
lol...
If they will really put desktop full cards to notebooks...
1070 SLI will be best bang for the bucks... And it will nicely hold 4K. -
This news has made my day!
I can't imagine a better upgrade from my 980M than a full 2560-core'd 1080.
More cores + lower clocks = better power efficiency and thus better performance than fewer cores with higher clocks.
I thought it'd be weird to see them regress on the mobile promise of 80% of desktop performance. This way they can maintain that level of performance in larger laptops.
Awesome!Last edited: Jun 2, 2016 -
How about gaming? What kind of peak power loads can a nice 4k action game with 99% constant GPU usage draw?
I usually use something like Folding@home CPU/GPU full loaded with tasks, 100% CPU/GPU usage at all times, let that sucker run for a while and see what kind of GPU load you see.
Real world load, to the MAX
jaybee83 likes this. -
This is definitely an interesting development... could this mean they have managed to get the full desktop chip in the standard MXM size, rather than the wider P870DM size? In which case, this could be quite revolutionary for regular MXM users. But cooling would be an issue that would prevent the chip from reaching its desktop base frequency probably..
Kade Storm and hmscott like this. -
Those news are awesome! This is a big step from what we've seen in the past. First of all, those 1070/1080 versions will work for all full-fat gaming laptops like GT73, GT80, P870DM-G, G752 etc. Slim style notebooks should feature M-cards, no doubt about it.
The only thing that worries me is the price tag. If NV continues to charge 1250USD for desktop class 1080 again, then 1070 might as well be at the level of 980M used to sell. Not so great(TomJGX, Kade Storm and hmscott like this. -
You know, maybe this Nvidia plan is more ingeniously simple than I at first thought...
If a 1080 / 1070 is used in every laptop, then we can all see and judge the performance of each laptop pitted against each other, all trying to meet the top performance mark.
Only the best cooled, best powered, highest performing laptops with a 1080 / 1070 will be sought after and purchased.
That's gonna kill the market for all of those slim wimpy overheating notebooks, as they can't possibly keep up performance wise. They may be pretty, but they can't do the heavy cooling required of a full performance 1080 / 1070.
That means computer buyers will finally be hip to the non-performing form factors, and demand bigger thicker laptops with cooling that can actually keep things cool.
That new performance visible trend favoring bigger laptops means more slots for 1080 / 1070's for Nvidia to sell through, ingenious
So much for hiding behind the 'm's, with the 1080 / 1070 the full performance comparison is all out there now for every buyer to see
Last edited: Jun 2, 2016Georgel, Eclipse251, HTWingNut and 6 others like this. -
"Just wait for the AMD show, then we´ll know more".
"Just wait for Computex, then we´ll know more".
Well we still don´t know **** about mobile cards and I´m not surprised.
And when we finally do learn something we´re gonna be unimpressed.
Yes I´m mad bro. -
Finally will a lots of fan boys see their beloved brands laptop models be outperformed by real performing laptops. Slim lightweight <C*******>models will go the safe death!! A dream come true
Georgel, Eclipse251 and hmscott like this. -
moviemarketing Milk Drinker
There isn't a lot of demand for the larger form factor models. Enthusiast laptop segment is much smaller than mid range.Georgel, TomJGX, hmscott and 1 other person like this. -
Thanks for catching that. I didn't mean to say the current trend, I meant that the result of opening up a view into the performance capability of the different form factors would favor bigger better cooling laptops.
Fixed it
"That new performance visible trend favoring bigger laptops means more slots for 1080 / 1070's for Nvidia to sell through, ingenious
"
When people can see what they are getting for their money, performance wise, it will be tough for the laptop makers to charge the same $ for less performance.
Some kind of performance range for each $100 cost would be nice to track, for the 1080 and the 1070, same for the 1060/etc.
There are likely some that would still want a slim laptop, but at least they can see what that form factor is costing them in performance.
Last edited: Jun 2, 2016 -
I will never buy a bulky boat anchor again. It's just not what I want. There's room for both in the market, and I'm sure we'll continue to see both. Obviously I'm sacrificing performance over a laptop with a 980M or notebook 980 or similar. It's not like I don't know that. What is wrong with both existing?
Sent from a 128th Legion Stormtrooper 6Phmscott likes this. -
I don't think this is going to force you into a "bulky boat anchor", but you can now see the advantage of that form factor providing better power and cooling, resulting in a measurable performance advantage.
Along with the disadvantage of the physical form factors that can't provide enough power and cooling to run the 1080 / 1070 at full TDP / clock performance.
It's possible that a down TDP and downclocked 1080/1070 won't be at that much of a performance disadvantage.
It will be fun to find out
transphasic likes this. -
It's just plain physics that a larger laptop has better cooling potential. No one is lost on this fact of physics.
Sent from a 128th Legion Stormtrooper 6P -
Until the 1080 / 1070 mobile distribution the laptops with poorer physics going for them could hide behind the 'm' GPU's, with much lower TDP / power requirements.
Of course the performance difference between the 'm's and the desktop mobile GPU was obvious, but they were considered in different classes of performance.
Now, all the laptops that had 980m's as well as 980's will get the same 1080 GPU, and can all be rated in the same 1080 class.
Same goes for all the 970m laptops, but to a lessor extent because the 970m laptops could all be compared equally.
But, the 1070 should open up the performance of the laptops with better power and cooling, pushing their performance well above the performance of the poorer physics form factors.
For me it has been tough watching people spend $2-3K for a laptop that doesn't perform anywhere near as well as the cost indicates.
Same goes for the insanely overpriced enthusiast laptops.
Putting all the 1080 / 1070 wielding laptops on the same performance / dollar graph would be enlightening.Last edited: Jun 2, 2016 -
Another take on the M naming - they'll drop the 1080, or it would be exclusive for the machines where we see the current 980 desktop mobile. Everything else would be 1070 and lower, since 980m was ~= 970, 970m ~= 960. Not exact science, but I hope you get the picture. So my guess is that the best MXM-B would be 1070.
hmscott likes this. -
King of Interns Simply a laptop enthusiast
I reckon it means everything else gets soldered. Simply oversized, hot mxm modules sporting full downclocked 1070 and 1080 cores for huge DR laptops.
Papusan, Ethrem, triturbo and 1 other person like this. -
That would severely disappoint 980m laptop owners. That would be a real ego blow to "trade up" to the "970" replacement.
The 1080 should be tuneable to draw less power and lower clock at the same core count.
Of course the laptops that don't require reducing the maximum TDP and clock will perform better
Georgel, Papusan, jaybee83 and 1 other person like this. -
-
They aren't "hiding" being lower tdp gpus. For instance the MSI ghost, blade. Etc.. They will continue to have 100w and lower tdp gpus just like they always have. Whatever nvidia launches for notebooks with a 100w tdp is what you'll see installed in it. I'm sure they will not just abandon that market. And whatever product exists for the larger notebooks that haver higher tdp gpus, you'll see those there.
The performance of those notebooks will continue to be in line with those tdp restrictions, just like they do now. These different classes of notebooks aren't going to be able to handle a higher tdp gpu than they do now for absolutely no reason.
The state you're trying to explain already exists. The performance gap between different notebook sizes and form factors already exists.
Sent from a 128th Legion Stormtrooper 6Phmscott likes this. -
So still no idea if we will be able to upgrade our 980ms yet?
They haven't announced that these new cards will have Mxm technology yet have they ?hmscott likes this. -
I don't think ALL laptops will have either 1070 or 1080. There will likely be a 1050 an 1060 version that will also go into laptops for the lower/mid-range laptops with 40-60W TDP and I'm certain they will come up with a low voltage version as well that consumes less than 25W. In any case it will be confusing. Joe Blade has a .5" thick laptop with 1070 and compares with Joe Clevo with 1.5" thick laptop with 1070. Joe Clevo's laptop consistently pushed 25-30% more FPS than Joe Blade's. Joe Blade is mad.
-
Not the same.
The 980m's and 970m's all have a TDP limit that is well below the desktop 1080 / 1070. 100w vs 180w/150w.
All the 180w 1080's will be distributed across 980m's, and 980 laptops.
The form factors that barely did 100w, won't be able to tap into that 80w above 100w, while the form factors with better cooling will be able to run at full power.
Is it possible that the 1080 will be offered in fixed TDP MXM's?
Card's at 180w, 150w, 120w, 100w would make sense, or something near those TDP's.
But, they will all be called 1080's, and all comparable as 1080's.
What is different is having such a wide range of power and thermal requirements, all under the same name label.
Same for the 1070, with different TDP breaks. Say, 150w, 125w, 100w, and 75w.
Again, the performance of all TDP 1070's are going to be compared as a 1070.
They could rename them to differentiate, using the TDP as part of the name.
More likely the TDP will be tuneable via the firmware by the vendor, to finely adjust the fit of the GPU to their power and cooling capability.
Any way you look at it, a wide range of laptop form factors will be compared against each other across the 1080 / 1070 / etc labels.
It's very different
Papusan likes this. -
but the 980m is 130w
-
A Pascal MXM was "shown" in the Nvidia Auto Drive demo - it was likely a Maxwell MXM shown, but it's final installation will be a Pascal MXM cards.
There are too many laptops already shipping, and still shipping with MXM form factor GPU's for Nvidia / vendors to stop supporting it with Pascal.
And, there hasn't been any alternative put forth to graduate to beyond the MXM card in consumer laptops.
I wouldn't worry about it
-
The blade will continue to have 90-100w tdp gpu. Whatever has a 980m will continue to have a 120-130w tdp gpu. Those super large notebooks that have the 980 notebook gpu will continue to use a 160-170w gpu.
No matter what label or strategy nvidia uses for this, the tdp restrictions are going to be the same, and the performance gaps will be the same.
Sent from a 128th Legion Stormtrooper 6P -
I think it is still a guess. Pascal 1080 is for example 150watt.TDP but still needs an enormous heatstink because the chip die is so small and it is difficult to spread that heat to lower levels.
Bit similar to the heat issues between Sandy Brdige and Ivy Bridge in combination with overclocking. -
pascal 1080 is 180w, 200w on 120% slider (+/- 10w)
-
This could be possible, not like they aren't doing it already with the 980 mobile. There do seem to be quite a few GP 104 Boards, even accounting for workstation cards:
1B80 GeForce GTX 1080 (GP104-A)
1B81 Graphics Device (GP104-A)
1B82 Graphics Device (GP104-A)
1B83 Graphics Device (GP104-A)
1BB0 Graphics Device (GP104GL-A)
1BB1 Graphics Device (GP104GL-A)
1BB4 Graphics Device (GP104GL-A)
1BC0 Graphics Device (GP104-B)
1BC1 Graphics Device (GP104-B)
1BC2 Graphics Device (GP104-B)
1BC3 Graphics Device (GP104-B)
1BF0 Graphics Device (GP104GL-B)
1BF1 Graphics Device (GP104GL-B)
1BF4 Graphics Device (GP104GL-B)
1BF5 Graphics Device (GP104GL-B)
http://wccftech.com/nvidia-pascal-device-id-gp-102-gpu-confirmed/#ixzz4AQ3KkGt4hmscott likes this. -
By the time the Pascal laptops ship, everyone will be so full of desktop 1080 / 1070 performance game / benchmark results, they will be comparing the laptop they want against those desktop 1080 / 1070 results.
The higher TDP capable laptops will approach and match the desktop 1080 /1070 results.
Those laptops that can't handle the full TDP of the 1080 / 1070 GPU will have performance that falls somewhere well below the full desktop 1080 / 1070 TDP results.
It will become clear that even though they have the *same* Pascal 1080 / 1070 GPU in their laptop, that the designers of the laptop they want to buy have failed to deliver full 1080 / 1070 performance.
I don't think that's going to go very well for those laptops lacking in performance against the full 1080 / 1070 desktop performance.
Those that want the best 4k or VR gaming experience are going to want / need all the performance that they can get.
And they are going to look for the laptops that can deliver that full desktop performance.Last edited: Jun 2, 2016Papusan likes this. -
It would be nice to have a matching list of columns with attributes for each device id.
Nvidia's full Geforce Pascal lineup device hardware IDs revealed
http://www.fudzilla.com/news/graphi...ce-pascal-lineup-device-hardware-ids-revealed
Breaking news: AIDA64 developers confirm Pascal GP102, GP106, GP107 and GP108
http://videocardz.com/60289/breakin...rs-confirm-pascal-gp102-gp106-gp107-and-gp108
AIDA64 developers have just contacted me with stunning information. They released a very long list of almost every Pascal GPU that will be released in the coming weeks and months, even those that were not confirmed by NVIDIA.
NVIDIA Pascal GP102 confirmed?
NVIDIA GP100, GP102
The most important news of the day: AIDA64 confirms GP102 silicon. This is probably the Big Pascal for gamers, that was being speculated ever since GP100 came out.
Judging from the released list GP100 will be computing oriented GPU (as most parts are tagged as GL). Meanwhile GP102 has more non-GL parts (so presumably those are GeForce devices). However both GPUs might be available in GeForce series, because GP100 also has three non-GL parts.
Pascal GP106, GP107, GP108
Moving on to lower end parts, we have GP106, the new GPU that was showcased in Drive PX 2 module. We also have GP107 and GP108, where the latter will probably only be offered in form of mobile graphics cards.
AIDA64 Devs:
Ever wondered how big the new #nVIDIA #Pascal GPU family would turn out to be? Here’s the whole list of Pascal SKUs, with their respective PCI device IDs. Enjoy!
Note: The list contains a few Maxwell IDs as well, in order to clarify the difference between GMxxx-A and GMxxx-B PCI device regions.
Note #2: GM200-B may have been a second iteration, an optimized variant of the original GM200, but never reached the market. Could have been a plan B in case Pascal slips to late 2016 or 2017.
Pascal GP100
What to expect: TESLA P100 (released), multiple QUADROs, TESLAs, GRIDs, possible GeForce cards in 2017.
15F0 Graphics Device GP100GL-A
15F1 Graphics Device GP100GL-A
15F8 Graphics Device GP100GL-A
15F9 Graphics Device GP100GL-A
15FA Graphics Device GP100GL-A
15FB Graphics Device GP100GL-A
15FC Graphics Device GP100GL-A
15FD Graphics Device GP100GL-A
15FE Graphics Device GP100GL-A
1725 Graphics Device GP100-B
172E Graphics Device GP100-B
172F Graphics Device GP100GL-B
1731 Graphics Device GP100GL-B
1738 Graphics Device GP100GL-B
1739 Graphics Device GP100GL-B
173A Graphics Device GP100GL-B
173B Graphics Device GP100GL-B
173C Graphics Device GP100GL-B
173D Graphics Device GP100GL-B
Pascal GP102
What to expect: GeForce GTX 1080 Ti, new TITAN, few QUADROs/TESLAs.
1B00 Graphics Device GP102-A
1B01 Graphics Device GP102-A
1B30 Graphics Device GP102GL-A
1B38 Graphics Device GP102GL-A
1B3E Graphics Device GP102GL-A
1B40 Graphics Device GP102-B
1B41 Graphics Device GP102-B
1B6E Graphics Device GP102-B
1B6F Graphics Device GP102-B
1B70 Graphics Device GP102GL-B
1B78 Graphics Device GP102GL-B
Pascal GP104
What to expect: NVIDIA GTX 1080 (released), GTX 1070 (released), GTX 1080M and GTX 1070, plus few QUADROs.
1B80 GeForce GTX 1080 GP104-A
1B81 Graphics Device GP104-A
1B82 Graphics Device GP104-A
1B83 Graphics Device GP104-A
1BB0 Graphics Device GP104GL-A
1BB1 Graphics Device GP104GL-A
1BB4 Graphics Device GP104GL-A
1BC0 Graphics Device GP104-B
1BC1 Graphics Device GP104-B
1BC2 Graphics Device GP104-B
1BC3 Graphics Device GP104-B
1BF0 Graphics Device GP104GL-B
1BF1 Graphics Device GP104GL-B
1BF4 Graphics Device GP104GL-B
1BF5 Graphics Device GP104GL-B
Pascal GP106
What to expect: GeForce GTX 1060 and GTX 1060M, also one QUADRO card.
1C00 Graphics Device GP106-A
1C01 Graphics Device GP106-A
1C02 Graphics Device GP106-A
1C03 Graphics Device GP106-A
1C30 Graphics Device GP106GL-A
1C41 Graphics Device GP106-B
1C42 Graphics Device GP106-B
1C43 Graphics Device GP106-B
1C70 Graphics Device GP106GL-B
Pascal GP107
What to expect: GeForce GTX 1050 and GTX 1050M series. Also three QUADRO cards.
1C80 Graphics Device GP107-A
1C81 Graphics Device GP107-A
1C82 Graphics Device GP107-A
1CA7 Graphics Device GP107GL-A
1CA8 Graphics Device GP107GL-A
1CAA Graphics Device GP107GL-A
1CC2 Graphics Device GP107-B
Pascal GP108
What to expect: GeForce GTX 1040/1040M, basically minesweeper-capable graphics.
1D01 Graphics Device GP108-ALast edited: Jun 2, 2016 -
Usually when we start seeing a large mass of hardware ID's like that, the products are in testing phase and the first batch of shipments is out of mass production, being prepared to be shipped to manufacturers (which can still take a couple months but at least we know there is progress being made).
jaybee83, VoodooChild and hmscott like this.
Pascal: What do we know? Discussion, Latest News & Updates: 1000M Series GPU's
Discussion in 'Gaming (Software and Graphics Cards)' started by J.Dre, Oct 11, 2014.