The Notebook Review forums were hosted by TechTarget, who shut down them down on January 31, 2022. This static read-only archive was pulled by NBR forum users between January 20 and January 31, 2022, in an effort to make sure that the valuable technical information that had been posted on the forums is preserved. For current discussions, many NBR forum users moved over to NotebookTalk.net after the shutdown.
Problems? See this thread at archive.org.
← Previous pageNext page →

    nVidia 2015 mobile speculation thread

    Discussion in 'Gaming (Software and Graphics Cards)' started by Cloudfire, May 9, 2015.

  1. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    I can see something being released around Computex, but not Q4 2015. That would be so close to Pascal. They've already delayed Pascal twice, I think.
     
  2. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    All of that is understood. The problem I have is believing that a greedy company like nVidia that has no competition at the moment would release such a powerhouse when they still have full GM204 that they can push off on us and then hold back on Pascal performance so they can milk it with two or three rebrands.

    Releasing this monster will mean they will have to really make a performance jump with Pascal.
     
    jaybee83, n=1, triturbo and 3 others like this.
  3. Papusan

    Papusan Jokebook's Sucks! Dont waste your $$$ on Filthy

    Reputations:
    42,712
    Messages:
    29,847
    Likes Received:
    59,655
    Trophy Points:
    931
    What if Nvidia know more than us about what Amd working on and will be ahead?
     
  4. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    That would be even more of reason for nVidia to not show their hand in my mind. Why give AMD a heads up that gives them time to tweak performance further when nVidia can have cards shipping out the door and in consumer hands before AMD?
     
  5. Papusan

    Papusan Jokebook's Sucks! Dont waste your $$$ on Filthy

    Reputations:
    42,712
    Messages:
    29,847
    Likes Received:
    59,655
    Trophy Points:
    931
    A to small perfomance increase over Gtx980m becomes too little I think. Amd has lying so far behind Nvidia that something must happen soon. I believe Nvidia also is aware of this. Amd has promised increased focus on high performance now so who knows?
     
    Last edited: May 12, 2015
    triturbo likes this.
  6. TonytotheB

    TonytotheB Notebook Consultant

    Reputations:
    6
    Messages:
    122
    Likes Received:
    30
    Trophy Points:
    41
    Just why?
     
  7. ratinox

    ratinox Notebook Deity

    Reputations:
    119
    Messages:
    1,047
    Likes Received:
    516
    Trophy Points:
    131
  8. Papusan

    Papusan Jokebook's Sucks! Dont waste your $$$ on Filthy

    Reputations:
    42,712
    Messages:
    29,847
    Likes Received:
    59,655
    Trophy Points:
    931
  9. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    Doubtful they will have one til 2016.
     
  10. ratinox

    ratinox Notebook Deity

    Reputations:
    119
    Messages:
    1,047
    Likes Received:
    516
    Trophy Points:
    131
    That's assuming AMD even has a mobile flagship in the works. And given that the "high" end R5 M375 looks to be between GTX 950M and GTX 960M? I'm not expecting much.
     
  11. be77solo

    be77solo pc's and planes

    Reputations:
    1,460
    Messages:
    2,631
    Likes Received:
    306
    Trophy Points:
    101
    I loved my old mobile HD 5870, that was what, 4-5 years ago? It was relevant for years and kept up well. We just need them to find the magic again and I'll gladly jump ship to keep Nvidia honest.
     
  12. Game7a1

    Game7a1 ?

    Reputations:
    529
    Messages:
    3,159
    Likes Received:
    1,040
    Trophy Points:
    231
    The R9 m375 is actually about on par or better than a 940m thanks to the DDR3 VRAM it has. The R9 m280x is on par with the GTX 850m/950m, so expect the m380x to be competing with the 950m/960m.
     
  13. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    AMD hasn't even responded to NVIDIA Maxwell, yet. That's what they plan to do this June at Computex. Don't expect anything amazing* from AMD this year (in mobile). They fell behind years ago, and have stayed behind NVIDIA, which leads to re-branding on both sides.

    *By "amazing," I mean a new architecture or something better than Maxwell. It's highly unlikely. Amazing in regards to AMD (in general) would be something as measly as proper driver support. :D

    In other words, I agree with @Ethrem : It doesn't make sense. Why hurry? Not like anyone is forcing their hand or competing on their level.
     
  14. Cloudfire

    Cloudfire (Really odd person)

    Reputations:
    7,279
    Messages:
    10,304
    Likes Received:
    2,878
    Trophy Points:
    581
    Mobile cards have always been built around the power metric.
    New cards from a new architecture is pretty much as fast as SLI from the last cards of the last architecture.

    GTX 680M (Kepler) was as fast as GTX 580M SLI (Fermi). Much faster than 480M SLI (Fermi).
    GTX 980M (Maxwell) was almost as fast as GTX 780M SLI (Kepler) and slower than 880M SLI because Nvidia lowballed mobile with Maxwell.

    New architecture like Pascal will be vastly better in efficiency than Maxwell, plus they will introduce stacked VRAM and it will be built on 16nm, trust me, their first Pascal will be much better than 980M SLI.
     
    transphasic likes this.
  15. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    Well there's a difference. From what I understand Pascal won't be here until at least Q3 2016. This fall with 990m will be a Maxwell rebrand and no way that will be 980m SLI performance.
     
    jaybee83 and Prema like this.
  16. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    If I had to guess, assuming the 990M is real and has more cores, it should perform about 30% better than the 980M.

    GTX 980 has 2048 Maxwell cores. The Titan X has 3072 Maxwell cores. The Titan X performs about 30% better than a GTX 980.

    NVIDIA follows trends when it comes to mobile vs desktop graphics. They can easily push another 30% out of Maxwell mobile to still top AMD's response to Maxwell, and that's exactly what I expect of such a company. And not to mention, 30% actually makes perfect sense, considering Pascal is due soon thereafter.
     
    transphasic likes this.
  17. Meaker@Sager

    Meaker@Sager Company Representative

    Reputations:
    9,431
    Messages:
    58,194
    Likes Received:
    17,901
    Trophy Points:
    931
    The titan-x also has more ROPS and a 50% wider memory bus. There is not really much more memory bandwidth room in the 980M, it would all be core horsepower.
     
  18. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    That's true. Didn't consider that. But, nonetheless, 30% is more reasonable than 2x the performance of the 980M, which I'd expect only to come from Pascal.
     
  19. ratinox

    ratinox Notebook Deity

    Reputations:
    119
    Messages:
    1,047
    Likes Received:
    516
    Trophy Points:
    131
    Yep. That's why I figure that if 990M is real then it's more like a better 970M SLI.
     
  20. Prema

    Prema Your Freedom, Your Choice

    Reputations:
    9,368
    Messages:
    6,297
    Likes Received:
    16,485
    Trophy Points:
    681
    Keep your expectations low or you have to be ready for another disappointment...
     
    jaybee83, TomJGX, Papusan and 4 others like this.
  21. D2 Ultima

    D2 Ultima Livestreaming Master

    Reputations:
    4,335
    Messages:
    11,803
    Likes Received:
    9,751
    Trophy Points:
    931
    My only expectation is that more GPUs will come out till nVidia is finished milking us =D.
     
    jaybee83 likes this.
  22. n=1

    n=1 YEAH SCIENCE!

    Reputations:
    2,544
    Messages:
    4,346
    Likes Received:
    2,600
    Trophy Points:
    231
    The main issue I have with all this is that even the 125W 880M proved to be too hot for pretty much any machine to handle. I just can't see how you'd fit all that power into a 150W envelope and not have to drastically redesign your laptop cooling to prevent burning up the card even at stock, nevermind overclocking.

    I guess that crazier thing was the 480M. It happened, it was a disaster, and nVidia nevered bother going that route again.
     
  23. D2 Ultima

    D2 Ultima Livestreaming Master

    Reputations:
    4,335
    Messages:
    11,803
    Likes Received:
    9,751
    Trophy Points:
    931
    Well some of these notebooks can handle it... *cough* P7xxZM *cough*

    Saw a video from J. Dre yesterday. Guy didn't even hit 60c in BF4 at 96% scaling, ultra preset, unlocked FPS. My cards hit over 60c in SLI getting like 50% usage on each and I ain't even at ultra AND I limit my fps to 125.
     
    jaybee83 and Mr Najsman like this.
  24. iaTa

    iaTa Do Not Feed

    Reputations:
    1,328
    Messages:
    2,675
    Likes Received:
    197
    Trophy Points:
    81
    Oh how I've missed the Nvidia GPU speculation, love it.

    Maybe Nvidia fully realise the impact that VR will have towards the end of the year / start of next and are releasing a mobile beast to power it :D
     
  25. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Even if some systems can handle it, I wouldn't want a card that hot, just because... Better off waiting for Pascal. It will be a tiny, little monster that runs cooler than anything today. Either that or it will run the same temps as Maxwell but output twice the performance (due to the shrink in the process).
     
  26. Cakefish

    Cakefish ¯\_(?)_/¯

    Reputations:
    1,643
    Messages:
    3,205
    Likes Received:
    1,469
    Trophy Points:
    231
    Firstly, I think that this '990M' will be called something else (either 1080M or something entirely different), as NVIDIA have never gone over a year without refreshing it's mobile lineup with a new branding/naming scheme. Secondly, I think it's going to have 2048 cores and be a GM204 chip, as that just makes more sense to me with such a hypothetical GPU already offering a sizeable improvement vs 980M and already having a mobile-friendly 256bit memory bus. I think that the story of mobile Kepler will play out again (680M to 780M, both GK104).

    I hope I'm wrong, but I'll be very surprised to see a GM200 chip in laptops.

    Edit: Either way, I don't think I will have enough willpower to resist getting a laptop equipped with one of these and Skylake. Too much temptation. Can't ignore. I'll be kicking myself when it gets blown away by Pascal the following year, but maybe by then I'll have a full time job (one can only hope!).
     
    Last edited: May 13, 2015
  27. Cloudfire

    Cloudfire (Really odd person)

    Reputations:
    7,279
    Messages:
    10,304
    Likes Received:
    2,878
    Trophy Points:
    581
    Meh, worst case should be 2048 core GTX 990M anyway.
    Thats 20-30% better performance. Solid boost that too :)
     
  28. Cakefish

    Cakefish ¯\_(?)_/¯

    Reputations:
    1,643
    Messages:
    3,205
    Likes Received:
    1,469
    Trophy Points:
    231
    I couldn't ignore a boost like that. Seeing as my 980M is hopeless at overclocking, I'm much more open to replacing it quickly than I'd otherwise be.

    I just wonder whether a refreshed P65xSG from Clevo will be able to handle these higher powered chips.
     
  29. Cloudfire

    Cloudfire (Really odd person)

    Reputations:
    7,279
    Messages:
    10,304
    Likes Received:
    2,878
    Trophy Points:
    581
    Its hopeless at overclocking? Didnt Prema release a vbios that makes it very stable or was it Dell exclusive?

    Yeah say 25% boost would be decent. Still won`t convince me to return to notebooks though.
     
  30. Game7a1

    Game7a1 ?

    Reputations:
    529
    Messages:
    3,159
    Likes Received:
    1,040
    Trophy Points:
    231
    It could be called the GTX 985m.
     
  31. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    980MX is more likely, in my opinion. They've already used "MX" once. But the information in the main post suggests 990M. :p

     
  32. D2 Ultima

    D2 Ultima Livestreaming Master

    Reputations:
    4,335
    Messages:
    11,803
    Likes Received:
    9,751
    Trophy Points:
    931
    There is absolutely no basis for this. For all we know it could double maxwell's performance but run as hot as Kepler or Fermi. Maxwell is so unnaturally cool it's unrealistic to simply "expect" any later generation to run as cool or cooler.
    HIS 980M is hopeless, just like my slave 780M needs a ridiculous amount of voltage (I can do 1006/6000 on +37.5 on master GPU, and slave needs +62.5 for the same OC). Not 980Ms themselves (though they're nowhere near as good as 970Ms are). He can't change his 980M either cuz board solder, so it's not as simple as buying a new GPU when another comes out.
    I hear you there. I'm not gonna discard my notebook till I can't use it anymore probably, but I wouldn't jump back into things unless a true amazing notebook comes out. A SLI version of a P770ZM with a 400W-450W brick and skylake and two full GM204 chips WITHOUT low voltage vRAM and WITH proper amounts of VRMs with dual PSU compatibility and even 120Hz screen compatibility? THAT would be something that'd make me feel enthusiast laptops aren't dead.
     
  33. Cloudfire

    Cloudfire (Really odd person)

    Reputations:
    7,279
    Messages:
    10,304
    Likes Received:
    2,878
    Trophy Points:
    581
    *Walks away*

    [​IMG]
     
    Any_Key, Mr Najsman and J.Dre like this.
  34. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Holy crap, they did it. o_O
     
    Last edited: May 13, 2015
  35. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    GM200 mobile... lulz.
     
  36. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    Well they did 880M so I guess we shouldn't be surprised.
     
  37. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Please not an 880M repeat... Just make it a better 980M.
     
  38. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    Dem Photoshop skillz
     
  39. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    Why is the vendor ID 10DS? NVidia is 10DE.
     
  40. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    350.12 don't contain INF entry for no stinkin' GTX 980MX.

    Why use a 2048-shader cut-down GM200 instead of a full fat GM204?

    Why is all the info filled in despite the GPU not being in GPU-Z's database?

    Smells fishy, eh? :p

    Can't believe Cloudfire has pulled some people's leg for 14 pages and counting. :D
     
  41. Game7a1

    Game7a1 ?

    Reputations:
    529
    Messages:
    3,159
    Likes Received:
    1,040
    Trophy Points:
    231
    The BIOS revision number also looks weird.
    I wonder if Cloudfire's walk is done...
     
  42. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    Release date predates Titan X ROFL.
     
    Last edited by a moderator: May 14, 2015
  43. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    That was the first thing I noticed, then the shaders, plus GPU-Z actually showing all the info but there is no 10DS vendor, it isn't possible. There's no S in hexadecimal. That's what sealed it for me.
     
  44. D2 Ultima

    D2 Ultima Livestreaming Master

    Reputations:
    4,335
    Messages:
    11,803
    Likes Received:
    9,751
    Trophy Points:
    931
    Only thing I could think of would be improved number of TMUs? *shrugs*
     
  45. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    Those get disabled as well. 128 TMUs/64 ROPs is the same as GTX 980.
     
    D2 Ultima likes this.
  46. Any_Key

    Any_Key Notebook Evangelist

    Reputations:
    514
    Messages:
    684
    Likes Received:
    316
    Trophy Points:
    76
    I... wanted... to... believe... why couldn't you let me be happy for more than a moment!
     
  47. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    Our condolences
     
  48. n=1

    n=1 YEAH SCIENCE!

    Reputations:
    2,544
    Messages:
    4,346
    Likes Received:
    2,600
    Trophy Points:
    231
    Honestly the first 2 things I saw were "980MX" and "2048 shaders". I immediately thought to myself yep a full GM204 that makes sense. :)

    Then I saw the die size and release date and was like :rolleyes: "I wonder how many will fall for this :D"
     
  49. Game7a1

    Game7a1 ?

    Reputations:
    529
    Messages:
    3,159
    Likes Received:
    1,040
    Trophy Points:
    231
    You know, I just went back to the "source", and no where does Cloudfire's phone pic (not saying it is his phone) info, the SLI bit, show up anywhere in the source. Not even the time.
    I have to ask, did anybody look back at the original source, or did people just blindly believe this?
     
    Last edited: May 13, 2015
  50. James D

    James D Notebook Prophet

    Reputations:
    2,314
    Messages:
    4,901
    Likes Received:
    1,132
    Trophy Points:
    231
    Am I the only one who sees something wrong in numbers here?
     
← Previous pageNext page →