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 →

    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.

  1. NuclearLizard

    NuclearLizard Notebook Deity

    Reputations:
    162
    Messages:
    939
    Likes Received:
    728
    Trophy Points:
    106
    I was getting all excited for HBM2.

    Ah, well gotta do something so we dont deflate the desktop users ego's.

    i wonder if its going to retain full clock speeds etc of the desktop cards. they've proved they can do it and supposedly pascal is going to be more efficient.



    Sent from my SGH-M919V using Tapatalk
     
  2. Cloudfire

    Cloudfire (Really odd person)

    Reputations:
    7,279
    Messages:
    10,304
    Likes Received:
    2,878
    Trophy Points:
    581
    HBM2 is too far down the pipeline.

    Hynix began HBM1 production in Q3 2014 and AMD couldnt deliver HBM1 cards until June 2015 and yields was still a mess.
    HBM2 production will begin at the start of Q2 2016 so I would guess HBM2 cards isnt available until around Q4 2016 at earliest with similar yields as AMD had I think.
    [​IMG]

    So Nvidia can`t afford to wait til late 2016 before rolling out Pascal. Thats where GDDR5X comes in.
     
  3. NuclearLizard

    NuclearLizard Notebook Deity

    Reputations:
    162
    Messages:
    939
    Likes Received:
    728
    Trophy Points:
    106
    ahh, i have been hearing that samsung was gonna start cranking out HBM2 modules Q1 2016.

    I suppose it dosent matter much, for I'll probably not have the funds to get anything equipped with pascal till Q3-4 anyways.

    Sent from my SGH-M919V using Tapatalk
     
  4. SR13

    SR13 Notebook Enthusiast

    Reputations:
    0
    Messages:
    26
    Likes Received:
    25
    Trophy Points:
    16
    That's what a ton of tech sites are saying. I am just going to be pissed if they don't refresh mobile until Q3 2016, I really want a new laptop before next summer.
     
  5. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    I really don't think that mobile is going to get HBM for awhile...
     
    jaybee83 likes this.
  6. SR13

    SR13 Notebook Enthusiast

    Reputations:
    0
    Messages:
    26
    Likes Received:
    25
    Trophy Points:
    16
    That may be true, but Micron is saying end of 2016 for GDDR5X, that would be a heck of a wait if the 1080/whatever its name is gets released in March.
     
  7. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    I'm betting HBM2 will be in mobile, but limited to the top end cards, like the 980M replacement.

    Haven't heard a peep about GDDR5X until now and we've known about HBM since 2013-2014. GDDR5X may be for special cards.
     
  8. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    thats why im saying, mobile will most probably lack all the goodies that the desktop cards will be getting. dont get too hyped over HBM, ull just be disappointed in the end

    Sent from my Nexus 5 using Tapatalk
     
  9. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    I hope they're not creating GDDR5X to keep mobile and desktop performance distanced. That's just dumb.

    They are more than capable of creating mobile GPU's with HBM2.
     
  10. NuclearLizard

    NuclearLizard Notebook Deity

    Reputations:
    162
    Messages:
    939
    Likes Received:
    728
    Trophy Points:
    106
    i think it might be more for intergrated systems, from the top of my head tablets, mac books, etc.

    Sent from my SGH-M919V using Tapatalk
     
  11. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    I tend to agree.
     
  12. GTO_PAO11

    GTO_PAO11 Notebook Deity

    Reputations:
    173
    Messages:
    1,309
    Likes Received:
    205
    Trophy Points:
    81
    Will the pascal version upgrade from 970m be more powerful than the 980m? If so, how poweful?

    Sent from my SM-N900V using Tapatalk
     
  13. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    Are you serious?
     
  14. GTO_PAO11

    GTO_PAO11 Notebook Deity

    Reputations:
    173
    Messages:
    1,309
    Likes Received:
    205
    Trophy Points:
    81
    I dunno, breh. You tell me. Or is asking questions against TOS now?

    Sent from my SM-N900V using Tapatalk
     
  15. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    jaybee83 likes this.
  16. deadsmiley

    deadsmiley Notebook Deity

    Reputations:
    1,147
    Messages:
    1,626
    Likes Received:
    702
    Trophy Points:
    131
    Considering that this entire thread is devoted to answering your question, the only reasonable conclusion is that you could not be bothered to read any of it.

    Then when you are called out on it you cop an attitude?

    No soup for you!

    Sent from my SPH-L720 using Tapatalk
     
  17. GTO_PAO11

    GTO_PAO11 Notebook Deity

    Reputations:
    173
    Messages:
    1,309
    Likes Received:
    205
    Trophy Points:
    81
    I prefer toast with peanut butter.



    I'll check back if there are news for this Pascal.

    Sent from my SM-N900V using Tapatalk
     
    TomJGX and deadsmiley like this.
  18. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    Traditionally speaking, the x70 of the new GPU ends up being just slightly faster than the x80 of the old GPU but nVidia has been doing all kinds of things lately so who knows. Its generally not worth the upgrade to the new GPU if you're going from x70 to x80 in the next generation but Pascal could be leaps and bounds better than Maxwell if the heat isn't an issue and there are many cases where the Maxwell 970 actually stomps my 780 Ti so Maxwell didn't really follow that path already.
     
  19. octiceps

    octiceps Nimrod

    Reputations:
    3,147
    Messages:
    9,944
    Likes Received:
    4,194
    Trophy Points:
    431
    [​IMG]
     
  20. hfm

    hfm Notebook Prophet

    Reputations:
    2,264
    Messages:
    5,296
    Likes Received:
    3,048
    Trophy Points:
    431
    Well.. they're just straight up putting desktop GPU's into boat anchor notebooks these days.. Why not HBM.
     
  21. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    Because HBM would require a modification to MXM to support the larger die size. Unless you want everything to be soldered. Besides that, nVidia's real money is in desktop cards and HBM2 is going to be in short supply for awhile. I'm not saying we will never see it but look how long it took to get the 980 in laptops...
     
    jaybee83 and i_pk_pjers_i like this.
  22. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    yupyup, people tend to go all kinds of crazy now that we got the 980 in mobile format. its DOESNT mean were going full desktop from now on, i DOESNT mean that mxm 3.0b is dead and it ALSO, MOST OF ALL, doesnt mean that mobile chips will be equal in specs to desktop cards from now on. the 980 is an absolute exception from the rule!

    its like saying: oh look, theres the bugatti veyron with 1100+ HP, that means from now on every prius is gonna sport insane engines like that! mainstream powah for all!!!

    nope, not gonna happen, so dont get your panties wet yet...

    Sent from my Nexus 5 using Tapatalk
     
    TomJGX and moviemarketing like this.
  23. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    I wouldn't be all doom and gloom just yet, fellas. We really don't know what's going to happen. :p GDDR5X could just be for lower end cards, like the replacement for the 940M. HBM2 will be in mass production by the end of this year. This was confirmed pages back.
     
  24. Ethrem

    Ethrem Notebook Prophet

    Reputations:
    1,404
    Messages:
    6,706
    Likes Received:
    4,735
    Trophy Points:
    431
    Why on earth would they throw a more expensive memory chip on a card that doesn't even max GDDR5 as it is?
     
  25. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Pascal may utilize it.
     
  26. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    well, its not like the 980M doesnt profit from more bandwidth, it actually does so pretty significantly, especially when oced. its just that it would be too good to be true if ngreedia went HBM and/or GDDR5X on mobile right away with pascal...
     
  27. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    This article makes a good argument for NVIDIA not using GDDR5X as a standard for their 2016 lineup of GPU's, but rather for a limited few.

    It also explains a bit more about GDDR5X, for those who have no idea what the hell is going on. :D
     
    jaybee83 likes this.
  28. deadsmiley

    deadsmiley Notebook Deity

    Reputations:
    1,147
    Messages:
    1,626
    Likes Received:
    702
    Trophy Points:
    131
    What!? Six days and no updates?

    It must have been cancelled.

    Sent from my SPH-L720 using Tapatalk
     
    sa7ina, hmscott, TomJGX and 1 other person like this.
  29. PrimeTimeAction

    PrimeTimeAction Notebook Evangelist

    Reputations:
    250
    Messages:
    542
    Likes Received:
    1,138
    Trophy Points:
    156
    No news is good news. :)
     
  30. Zymphad

    Zymphad Zymphad

    Reputations:
    2,321
    Messages:
    4,165
    Likes Received:
    355
    Trophy Points:
    151
    Well most of these pages aren't actually news or what we know, but speculations.
     
  31. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
  32. DataShell

    DataShell Notebook Deity

    Reputations:
    47
    Messages:
    777
    Likes Received:
    354
    Trophy Points:
    76
  33. Cakefish

    Cakefish ¯\_(?)_/¯

    Reputations:
    1,643
    Messages:
    3,205
    Likes Received:
    1,469
    Trophy Points:
    231
    I remember when this thread was created and Pascal was but a distant dream. Now it seems like it might as well be launching tomorrow. It feels close.


    Sent from my iPad using Tapatalk
     
  34. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Yeah, it probably isn't a 990. That's pretty cheap, unless that's the "invoice" price and not MSRP. It could also be a mobile Pascal GPU. Top end mobile cards are around $1000. Who knows... All speculation. All we know for sure is that NVIDIA has begun shipping them to laboratories for final testing before mass production.

    @Cakefish: Yeah, it's not going to be too long now before we see some legit leaks. ;)
     
  35. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,110
    Messages:
    20,384
    Likes Received:
    25,139
    Trophy Points:
    931
    Nvidia GPU daughter card.jpg
    I don't see a connector evident, but I do see a number of fastening holes, to me it looks like a daughter card to attach to another host board.

    Could be an in place GPU upgrade for a new form factor delivery system - IDK what connector is on the bottom of that card.

    Maybe a daughter card for a new high density compute platform; add a bunch of those to a host board that would be used for a compute cluster.
     
    TomJGX and jaybee83 like this.
  36. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    Pascal (for mobile especially) is still at least six months off... we have a lot of waiting (and more speculation) to do.
     
    TomJGX and hmscott like this.
  37. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    yay, let the fun / speculating continue :D
     
  38. Cakefish

    Cakefish ¯\_(?)_/¯

    Reputations:
    1,643
    Messages:
    3,205
    Likes Received:
    1,469
    Trophy Points:
    231
    Currently, six months feels like a blink of an eye!
     
  39. Zymphad

    Zymphad Zymphad

    Reputations:
    2,321
    Messages:
    4,165
    Likes Received:
    355
    Trophy Points:
    151
    No worries then. 980M is performing well with 1080p 60 FPS DX11 games. This should arrive in time for DX12 gaming.
     
  40. Zymphad

    Zymphad Zymphad

    Reputations:
    2,321
    Messages:
    4,165
    Likes Received:
    355
    Trophy Points:
    151
    Are those HBM on the chip? And my guess would be it's for server/super computing.
     
    hmscott likes this.
  41. NuclearLizard

    NuclearLizard Notebook Deity

    Reputations:
    162
    Messages:
    939
    Likes Received:
    728
    Trophy Points:
    106
    Does anyone else think they might hold off HBM2 on mobile? Instead using GDDR5X for ram?

    also i wonder how well the mobile will stack up with the desktop. as from what i have been seeing its supposed to be a very sizable jump in power savings.

    Sent from my SGH-M919V using Tapatalk
     
  42. King of Interns

    King of Interns Simply a laptop enthusiast

    Reputations:
    1,329
    Messages:
    5,418
    Likes Received:
    1,096
    Trophy Points:
    331
    Considering GDDR5X will double bandwidth and that mobile parts usually lack power vs desktop parts it will not matter if they do use it...then again mobile cards are expensive. Perhaps they will use HBM 2.
     
  43. NuclearLizard

    NuclearLizard Notebook Deity

    Reputations:
    162
    Messages:
    939
    Likes Received:
    728
    Trophy Points:
    106
    Yeah, though i wonder how far pascal mobile is going to lag behind the desktop, its been creeping up for some time. discounting the missing 20nm architecture i'd think they'd be rather close by now. (there is a graph for this)

    As well as they have demenstrated they can stuff a suposedly more power hungry chip into a laptop and get away with it.

    My take would be that desktop will see HBM2 mobile will get to have GDDR5X and we might get to enjoy closer levels of chip power.

    Sent from my SGH-M919V using Tapatalk
     
  44. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    I think HBM2 will be for the refresh of Pascal (2017). For mobile, that is. I'm sure desktop Pascal will have HBM2 by next summer.

    Makes sense. Hope that's not the case but it probably is. GDDR5X is still more than twice as fast as GDDR5. :cool:
     
  45. NuclearLizard

    NuclearLizard Notebook Deity

    Reputations:
    162
    Messages:
    939
    Likes Received:
    728
    Trophy Points:
    106
    thats what im thinking, that or a "1080 for notebook" line. lol

    Sent from my SGH-M919V using Tapatalk
     
  46. n=1

    n=1 YEAH SCIENCE!

    Reputations:
    2,544
    Messages:
    4,346
    Likes Received:
    2,600
    Trophy Points:
    231
    I really don't know how much GDDR5X will improve things, given that current cards are not really memory bandwidth limited, especially at 1080p. Even at 1440p, overclocking the memory from 7 to 8GHz on my 980 Ti nets me a paltry 5-7% gain.
     
  47. PrimeTimeAction

    PrimeTimeAction Notebook Evangelist

    Reputations:
    250
    Messages:
    542
    Likes Received:
    1,138
    Trophy Points:
    156
    In my humble opinion, it is not likely that they will go that route. Their recent marketing campaigns for mobile GPUs boast about reducing gap between mobile and desktop. This is why we have things like GTX 980.It will not be inline with their strategy to keep HBM exclusive to desktops. However it is quite possible to have a mixed bag lineup similar to 800m series.
     
  48. Robbo99999

    Robbo99999 Notebook Prophet

    Reputations:
    4,346
    Messages:
    6,824
    Likes Received:
    6,112
    Trophy Points:
    681
    Well, if Pascal doubles core performance in comparison to Maxwell, then we could say we'll need close to double the bandwidth to ensure we don't remain bandwidth starved - this fits in quite nicely with the information out there that GDDR5X is twice as fast as GDDR5.
     
  49. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Pretty much exactly what I was going to say. :) I'm sure they will fully utilize GDDR5X. HBM2 has a fairly long lifespan.

    GDDR5 would be a bottleneck for Pascal.
     
    Last edited: Nov 4, 2015
    TomJGX, i_pk_pjers_i and Robbo99999 like this.
  50. n=1

    n=1 YEAH SCIENCE!

    Reputations:
    2,544
    Messages:
    4,346
    Likes Received:
    2,600
    Trophy Points:
    231
    But GDDR5 still doesn't bottleneck even a Titan X though, meaning current GPUs just aren't bandwidth starved. While I get the point about Pascal being 2x more powerful, improving upon something that isn't a current bottleneck is unlikely to yield significant improvements. That's all I'm trying to say.
     
    jaybee83 and i_pk_pjers_i like this.
← Previous pageNext page →