GTX 980 launched for notebooks

haha seeing all your posts concerning 980 desktop mobile in combination with your sig just makes me LOL every time @RIP mobile gaming

 

LOL we are indeed leaving the "mobile" world it seems
Oh well, can`t wait and see what these Clevo machines can do with the desktop (or was it mobile) GTX 980. The bigger MXM cards will be interesting to see too

 

If this 980 will be able to fit in my P370SM I'm gonna be one very happy man.

 

some form of it definitely will, just not the 200W TDP model

 

The main reason for the larger pcb is the 1750 mhz ram along side the extra power circuitry. I beefed up my vrms and it started impacting the memory clocks on the 980m.

 

i remember that franken-980M of yours

 

Because vrm stands for VROOOOOM!

 

The thing is with a regular 980M MXM design you could get fairly close clock wise with a fully unlocked vbios, being able to run two of those cards with fully unlocked cores will do better than even the most extreme clocked 980.

 

purely jumping into this with no info, are these cards upgradeable? It would be a huge waste if not.

 

They're modified MXM so I don't see why not. Provided that nVidia and manufacturers find value in doing so anyway.

 

It's not MXM 3.0b, so it depends on if future mobile cards keep the same form factor and this becomes a new standard.

 

All of the new Clevo "DM" models support MXM 3.0b. I'm 98% certain it will stick around through Pascal.

What we're seeing here is an experimental change.

 

Or maybe this is the beginning of the end for MXM

 

HBM takes away the requirement for the larger PCB though, the board space is left for VRM circuitry.

 

Next year... GTX 1080 in iPhone 7s!

 

People seriously think this is a "one-off" GPU?

You can't go back after Pandora's Box is opened, and this just flung it wide the heck open.

Nvidia can't just return to dinky little 100W, cut down 'M' flagships anymore peeps, not with the expectation now being that the TOTL mobile GPU will be a reflection on the desktop side. That just won't cut it any longer.

The Pascal 1080M won't be a cut down hybrid like the 980M, because once you give the market steak, they won't be able to then get excited by peanut butter sandwiches.

 

I don't agree. Just because they release the Titan X or whatever Titan flavor of the week, doesn't mean the subordinates aren't still eaten up like spilled M&M's at a Chuck E Cheese's.

 

You have to consider cost. That's why I think it's more likely to be a one time thing.* It's definitely enthusiast focused, and that's not where the big money is. I highly doubt the entire market will shift to this type of structure. I have no doubt it will stick around. Maybe once per year, like NVIDIA releasing a mobile Titan we've always wanted. It's definitely possible now that HBM2 is in the game.

We are entering a new era of mobile gaming, gentlemen. Can't wait to see what's ahead.

*"One time" meaning a rare breed, like the 690 or Titan Z. Not launched annually.

 

Kevin, it's likely going to be a dual track, normal MXM + custom MXM to get the most performance. Both have a place moving forward.

I do hope someone comes out with the "maxxed out" 980 MXM as an SLI for mobile.

If not MSI in the GT80, then Clevo in the 870DM or a "970DM" w / 18", if not them then... Alienware 18 2016?

One of the 980 announcement videos/articles mentioned Asus was going to put in an SLI into the GX700 water-cooling frame. That's an outside shot, but still possible. Asus likes to do custom MXM per laptop, so it plays to their expertise.

Either way, as long as the "desktop" mobile 980's sell well, I think we will see the makers doing this again for the next generation too

 

Like I say it's mostly for the fast GDDR5 which goes away with HBM....

 

Meaker, the size of the 980 MXM also seems to be varying due to the power handling, 200w, and 125-150w sizes.

Even with HBM memory, there may be real estate requirements for power delivery beyond available via the standard size MXM.

Unless you think Pascal desktop GPU won't have higher power requirements than mobile Pascal'm?

 

With HBM you should be able to double the number of available phases from 3 to 6 which is what this is advertising

 

even without HBM meaker has shown that the MXM 3.0 B form factor is able to handle more than 3 power phases easily

thus far, it seems like the 200W and 180W TDP models of the 980 are non-mxm standard. lets see what other tdp versions will pop up / get specified and where the TDP limit of the regular sized MXM boards will be at "stock"

 

__________________________________________________________

I wonder if upon the release of this via news coverage that the anticipated 990m is now DOA for 2015?
I can't see Nvidia pushing out both products simultaneously this fall.

_____________________________________________________

What a nice thought, but highly unlikely, as the architecture is too different to house it inside, as well as the cooling limitations, according to what I have been told by some large Notebook resellers.
It would be nice if this could be backwards compatible on my Haswell machine, and save me the cost of buying a new machine, but it's not gonna happen. It's probably not even cost efficient anyways to do so.

 

The 990M is a figment of imagination

 

980 = "990M"

if / when any 980 standard mxm 3.0 B cards arrive, they will be power and performance restricted, so expect less performance than a fullfledged 980 (980M +35%)

 

Quite surprised and pleased for the "full 980" announcement, and the new path that mobile gaming is embracing.....
buuuut, im afraid about one issue, and its the next:
What happens if i purchase a P770DM right now (in fact, that is what i was thinking to do... ) and the new 980´s custom MXM type set as the new standard? ( bringing up what the fellows Jr Dre, Kevin or Hmscott, were discussing)
How could i will be able to update my system?...
Would they leave me in the lurch?,
Will "they" (Nvidia or industry, or whatever) say something on this matter ?, or
the only way we have to know it, its just wait for Pascal one year?...
These are dangerous times; they are very dangerous times indeed... (Xd)

^^

 

Both Asus and Clevo will use the bigger MXM module with more power phases. Here are the details. Notice the memory clock difference between Asus and Clevo.

Nvidia say that GTX 980 goes from 1064MHz. All OEMs are free to increase clocks from that to suit their notebook. The 1064MHz is probably the normal MXM with 4 power phases instead of the bigger MXM with 6-8 power phases.

Asus GTX 980 (GX700)



Clevo GTX 980 (P870DM)

 

it wont become the new standard, no worries.

even if it WOULD, the socket would still be compatible to mxm 3.0 B that custom board has the exact same socket, the board itself is just a little wider, thats it!

@Cloudfire: hilarious how clevo manages to beat asus in their 980 specs WITHOUT watercooling

 

" even if it WOULD, the socket would still be compatible to mxm 3.0 B"
More or less, that is what i wanted to hear, haha

 

Ya, Clevo is gunning for the sickest 980 it seems. To quote one reseller explaining the difference between P870DM with GTX 980M SLI and the one with GTX 980:

So it will be interesting to see how succesful Clevo is at taming this beast lol.

Too bad for current notebook users though. Since the notebooks are carefully designed around the exact same size of current MXM, they cant use the wider MXM with better power phases.

 

Yeah well you could technically say the same thing about Asus cards. Still wouldn't fit in your M15x (or P770DM).

 

yeah, still waiting on those internal pics of the P870DM with the 980, wanna see that gargantuan 6 pipe heatsink prema mentioned i can imagine it stretching the entire length of the space normally covered by two MXM gpus in sli, thus making use of both of their gpu fans

@octiceps: with the difference that the clevo machines are able to take both the wider format 980 and the regular size mxm cards. asus machines wouldnt be able to do that... so its not as customized as is the case with asus mxmish designs

 

You sound worried

By extension, the left mounting hole is in a different location, and it would also require a different heatsink

That's exactly why it's 200W TDP, it's using the cooling of 2x100W normal MXM cards

Only the P870DM and P775DM will take the wide 980, in 200W and 180W flavors respectively. The other -DM and -ZM models will not support it as they're different internal designs.

Asus board designs exist in another dimension, but the end result would be the same for people who want to use the wide 980 and don't have the P870DM/P775DM.

 

So do you think the P775DM its the safe bet (and cheapest, he) for the future?

 

Yeah it looks that way. It'll support both MXM 3.0b and this new board design. If this new design becomes the standard, it would be more futureproof.

 

nah, not worried

nope, the boards are just a little wider, but the left mounting holes have the same locations as regular sized mxm boards

yup, that was exactly my reasoning@2x100W. thx for spelling it out

yup, thats correct@P870DM / P775DM being special machines

yes, but the user's question just now was pointing towards upgradeability of P870DM/P775DM machines with regular mxm 3.0 B boards, and that is a given (as you yourself just posted, while i was still writing my comment here )

Sent from my Nexus 5 using Tapatalk

 

I'm not talking about the heatsink mounting holes. I'm talking about the left standoff that secures the card to the mobo. The hole for it is on the extended section of PCB.

Regular MXM-B:

Spoiler

980:

Spoiler

 

So this card will take the same space as 2x 980M cards in SLI inside laptops that support it?

What is the point of that? SLI 980M would beat a desktop 980 hands down in many titles. Even 970M SLI could match and beat it if mildly OCed and at the same price probably.

So yeah basically what is the point of this card and why are people excited? Old non SLI laptops can't fit it and SLI laptops could run SLI 970M at same or less cost with oft better performance......

 

Because SLI is problematic, troublesome, and much more expensive than a single 980.

 

I know that is true. I thought SLI was improving with time though? I guess when you look at it that way it is a good 2nd option for dual mxm slot owners that don't want SLI and instead have a working top end single GPU.

Pretty small market segment though

 

Yes, in the case of the P870DM

It's being marketed as the best mobile solution for VR, where single GPU is still preferable and laptop 980 is the first mobile GPU to meet or exceed Oculus' recommended spec (970). And what @HTWingNut said.

 

LOL

 

Single 980 in a laptop reports roughly 11500 firestrike score with desktop cpu and a 980m sli laptop with mobile cpu reports roughly 13900 at a lower cpu clock speed so I am not sold on a single gpu yet either. Its really cool but I will just have to wait for the 980 SLI GT80 to throw my dollars at

BTW SLI is not troublesome (more expensive of course), you can always revert to a single GPU if the game or driver doesn't support it which is in very rare cases especially with new titles.

 

http://www.geforce.com/hardware/notebook-gpus/geforce-gtx-980/specifications

Minimum 1064 MHz + Boost

Well looks like even the MXM models will have at least a 1064Mhz core clock, looking good for SLI.

 

Honestly, nvidia slapping a full 980 on a laptop seems to me an answer to AMD touting how they got the R9 nano a full fury in a tiny envelope. Now nvidia answered back with "Well, I can fit my desktop GPUs to a laptop, how about that?"

 

I swear, the more I look at that mobile 980 the more I think it'll fit in the slave GPU slot of the P370SM. Heatsink won't fit (the fan will get in the way) but the card itself should be capable of being put there size-wise.

Or you could wait for the SLI mobile GTX 980 in the P870DM which will inevitably appear. Better than 980M, worse than full ~200W mobile 980, but dem_extra_GPU_powerz_mayne

It's not that SLI is "troublesome" (to someone of ballpark expertise to myself and/or many of the users currently commenting here), but that SUPPORT is dwindling. Engines that support it, games that support it on launch (or for months after), games that scale well (I had a scaling issue with MGSV: Ground Zeroes; I could not go above ~80% max scaling (and hovered near 70% most of the time) no matter what I did, and was getting FPS drops to the mid 40s somewhat often... OCing my cards from 850/5000 to 1006/6000 gave me constant 60fps at the same scaling levels, indicating an engine-level problem. I didn't even close the game, I simply alt-tabbed and double-clicked my NVI OC .exe files), technologies (Maxwell can no longer use CSAA, and cannot use MFAA in SLI, effectively making Maxwell SLI a flat out downgrade to equivalent Kepler cards in SLI... and SLI + Gsync + DSR does not work, far less the mismatched clockspeeds and voltages in SLI that Maxwell cards have by design which is unstable compared to previous gens) and plain old random issues which you see lots of glorious fixes for (and open issue comments) in nVidia's latest set of drivers for the last few months.

In other words, SLI is now considered by me to be a "wow, this game works well with SLI! Cool!" instead of a "wow, this game doesn't support SLI? That sucks, like everything else does; lazy-ass devs!".

I had a nice laugh at that one too =D. But sadly even at its worst, the Fury Nano should still beat a GTX 980. It'd be funny if AMD made similar-style MXM-ish cards from the Nanos. But this is AMD we're talking about so...

 

Will be real fun to see pictures of the new? die.
Old 980 GM204 where A1 stepping. For example Fermi where A3.

Either Nvidia are cherry picking cores for GTX 980 for mobile (and the upcoming dual card) or they have had time to do A2 or even a base layer change on the chip to B1. New stepping would explain lower energy requirements and save some money for Nvidia by not having to cherry pick chips.

Nvidia had some incentive to do a re-spinn since they use dual GTX 980 in their compute lines of cards (where the money is, not gaming)

 

SLI will however scale great in VR since each card render its own screen. It will be almost perfect scaling. So just hook up a VR set and SLI is solved.

 

If only this were to usher in a new standard of GPU's being offered by vendor's. Fully upgradable laptops... The way it should be.