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. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    If you're speaking to me, I never thought nor said GP100 will be in mobile.

    Anyway, I think all of these journalists, writers, bloggers, etc. are all hyped up and speculating out their a** without a clue as to what's really happening, and people are eating it up. The only reason they'd use GDDR5X instead of HBM is if they do not have sufficient supply of HBM2 in the market, which is possible. However, there are three manufacturers mass-producing it right now: TSMC, SkHynix and Samsung. Surely NVIDIA has a contact with one of them to deliver on their GPU technology. It is necessary to survive in this market.

    High-bandwidth memory is the foreseeable future of GPU technology. There is no denying that. It will arrive in mobile, we just don't know when.
     
    Last edited: Apr 19, 2016
    J.Dre, Apr 19, 2016
    #1951
    transphasic likes this.
  2. Ashtrix

    Ashtrix ψυχή υπεροχή

    Reputations:
    2,376
    Messages:
    2,081
    Likes Received:
    3,281
    Trophy Points:
    281
    I think they'd first milk the GDDR5X first plus this time the Ti will also be on GP104 and only HBM2 GPU that would be one of 3 GPUs that Nvidia was planning, So maybe 2017 Q2 for HMB2 on mobile considering the full redesign of heatsinks / chassis also maybe plus the MXM redesign due to bandwidth (?) (I just hope there's no BGA madness with this HBM2 due to MXM spec redesign) and we also don't know how AMD will be releasing their polaris GPUs on mobile spectrum, Vega will only be HBM2 from AMDs lineup this time from the leak.

    From that HBM2 will only be reserved for high end GPUs only, but the memory voltage efficiency is much better over GDDR5, no idea about GDDR5X which is almost at HBM1. Wish they really make a good MXM 3.0b specced HBM2 card..

    So first from AMD's response Ngreedia will follow with their own plans this time...Pure speculation.
     
    Last edited: Apr 20, 2016
    Ashtrix, Apr 20, 2016
    #1952
  3. Link4

    Link4 Notebook Deity

    Reputations:
    551
    Messages:
    709
    Likes Received:
    168
    Trophy Points:
    56
    I know people like to treat everything tech websites put out as rumors, and some even just scream fake in the comment sections all the time, but usually there is a decent amount of real information and the GP104 die is one of them. The fact that it already uses GDDR5/X means it only has GDDR5/X memory controller and won't work with HBM2. GP100 is nVidia's first real HBM2 design, I doubt the HBM2 memory controller in it is very die area/transistor efficient, but they can easily get away with it in a 610mm^2 chip. The only time I have seen a chip (not completely enterprise oriented and massive like Intel's Knight's Landing) with two completely different types of memory support was Kaveri APUs with both DDR3 and GDDR5 support (which they never used because Elpida who was going to make GDDR5 modules evaporated) and not counting current CPUs/APUs that have DDR3 and DDR4 support because those two aren't that different. As far as it comes to nVidia they only dedicate the minimum amount of die area to their GPUs, their 200+mm^2 GM206 GPU only has a crippling 128bit bus, GM204 is 400mm^2 and GP104 is 300mm^2 on 16nm (which is comparable to the current flagship GPUs in transistor count) with only 256bit buses, so there is no way they would put two separate memory controllers on a midrange chip. Again if GP102 is real that might have HBM2 but again that won't get into notebook. They would have to redesign another GP104 size GPU with a HBM2 controller instead to get it working with HBM2, and they will have to be willing to take a cut to their margins because HBM2 is expensive and I don't think nVidia is willing to do that in the near future.

    That will take at least another year, at which point it will be too close to Volta if the roadmap is somewhat accurate.
     
    Last edited: Apr 20, 2016
    Link4, Apr 20, 2016
    #1953
    transphasic and tgipier like this.
  4. Prema

    Prema Your Freedom, Your Choice

    Reputations:
    9,368
    Messages:
    6,297
    Likes Received:
    16,485
    Trophy Points:
    681
    Don't worry, Pascal is coming:

    [​IMG]
     
    Prema, Apr 20, 2016
    #1954
    ExMM, tgipier, Ethrem and 5 others like this.
  5. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    omg, lol

    Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk
     
    jaybee83, Apr 20, 2016
    #1955
    Prema and TomJGX like this.
  6. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    J.Dre, Apr 20, 2016
    #1956
  7. Riddhy916

    Riddhy916 Notebook Deity

    Reputations:
    35
    Messages:
    756
    Likes Received:
    87
    Trophy Points:
    41
    What are your guys predictions on 1080m? will it at stock beat gtx 980 desktop @1250 mhz on the core with maybe 75w to 80w tdp? as nvidia says pascal is 2x perf per watt?
     
    Riddhy916, Apr 20, 2016
    #1957
  8. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    the mobile flagship is always at 100w tdp, 75w will probably be the 1070m. as for performance, ur guess is as good as mine. depends on nvidia's milking strategy.

    Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk
     
    jaybee83, Apr 21, 2016
    #1958
  9. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    880m and 980m are actually about 125w tdp.
     
    HTWingNut, Apr 21, 2016
    #1959
    jaybee83 likes this.
  10. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    ...further confirming my point :)

    Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk
     
    jaybee83, Apr 21, 2016
    #1960
  11. Phase

    Phase Notebook Evangelist

    Reputations:
    7
    Messages:
    483
    Likes Received:
    100
    Trophy Points:
    56
    1070m should outperform the 980m for sure right?
     
    Phase, Apr 21, 2016
    #1961
  12. hfm

    hfm Notebook Prophet

    Reputations:
    2,264
    Messages:
    5,296
    Likes Received:
    3,048
    Trophy Points:
    431
    No one knows anything for sure. Only speculation.

    Sent from a 128th Legion Stormtrooper 6P
     
    hfm, Apr 21, 2016
    #1962
    hmscott likes this.
  13. Georgel

    Georgel Notebook Virtuoso

    Reputations:
    945
    Messages:
    2,378
    Likes Received:
    3,021
    Trophy Points:
    281
    Did this thread stop ?

    Is there no new information ?
     
    Georgel, Apr 21, 2016
    #1963
    hmscott likes this.
  14. hmscott

    hmscott Notebook Nobel Laureate

    Reputations:
    7,110
    Messages:
    20,384
    Likes Received:
    25,139
    Trophy Points:
    931
    Much time to kill between now and the next opportunity for Nvidia to release Pascal Consumer boards...until then...

    And, if you get 2 or 3 tabs of this going at once, it's really quite nice...
     
    Last edited: Apr 21, 2016
    hmscott, Apr 21, 2016
    #1964
  15. Kevin

    Kevin Egregious

    Reputations:
    3,289
    Messages:
    10,780
    Likes Received:
    1,782
    Trophy Points:
    581
    There was no information in the first place.
     
    Kevin, Apr 21, 2016
    #1965
    TomJGX, Georgel and hmscott like this.
  16. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    The only thing we know is that the GTX 1080 is coming at Computex (end of May). There's nothing else we can say with any certainty at this point aside from the fact that we expect AMD to give NVIDIA the middle finger this year.

    Let's just hope they do what they did with the 700 series and unexpectedly launch mobile a week after desktop... I remember back then it was completely unexpected. Everyone thought there wasn't going to be a 780M because they wanted to match the desktop and mobile naming scheme. Or was that the 880M?

    It's all branded crap, who cares. We don't really know anything except aliens. :D
     
    J.Dre, Apr 21, 2016
    #1966
    jaybee83, TomJGX, Ashtrix and 1 other person like this.
  17. Raidriar

    Raidriar ლ(ಠ益ಠლ)

    Reputations:
    1,708
    Messages:
    5,820
    Likes Received:
    4,311
    Trophy Points:
    431
    Raidriar, Apr 22, 2016
    #1967
    Mr Najsman likes this.
  18. Templesa

    Templesa Notebook Deity

    Reputations:
    172
    Messages:
    808
    Likes Received:
    247
    Trophy Points:
    56
    Currently at pax. Will be bothering nvidia about anything and everything.
     
    Templesa, Apr 22, 2016
    #1968
    triturbo, TomJGX, Ashtrix and 6 others like this.
  19. Raidriar

    Raidriar ლ(ಠ益ಠლ)

    Reputations:
    1,708
    Messages:
    5,820
    Likes Received:
    4,311
    Trophy Points:
    431
    Give em a big middle finger on behalf of Fermi and Kepler users
     
    Raidriar, Apr 22, 2016
    #1969
    triturbo, TomJGX, Ashtrix and 2 others like this.
  20. Georgel

    Georgel Notebook Virtuoso

    Reputations:
    945
    Messages:
    2,378
    Likes Received:
    3,021
    Trophy Points:
    281
    Georgel, Apr 22, 2016
    #1970
  21. nick81

    nick81 Notebook Evangelist

    Reputations:
    136
    Messages:
    386
    Likes Received:
    101
    Trophy Points:
    56
    nick81, Apr 23, 2016
    #1971
    hmscott and jaybee83 like this.
  22. Zero989

    Zero989 Notebook Virtuoso

    Reputations:
    910
    Messages:
    2,836
    Likes Received:
    583
    Trophy Points:
    131
    225Watt card.

    Both cards say 980 on them.... lmao.

    This is our 1080 leak...

    [​IMG]
     
    Zero989, Apr 23, 2016
    #1972
    i_pk_pjers_i, hmscott, tgipier and 1 other person like this.
  23. tgipier

    tgipier Notebook Deity

    Reputations:
    203
    Messages:
    1,603
    Likes Received:
    1,578
    Trophy Points:
    181
    ^GDDR5x
     
    tgipier, Apr 23, 2016
    #1973
  24. Omnomberry

    Omnomberry Notebook Evangelist

    Reputations:
    45
    Messages:
    556
    Likes Received:
    497
    Trophy Points:
    76
    They are just overlaying the leaked yellow image on 980 PCB while trying to get the size right according to the numbered arrows.
    As far as I understand its something along "the layout fits MSI pcbs"
     
    Last edited: Apr 23, 2016
    Omnomberry, Apr 23, 2016
    #1974
    jaybee83, Ashtrix and hmscott like this.
  25. Georgel

    Georgel Notebook Virtuoso

    Reputations:
    945
    Messages:
    2,378
    Likes Received:
    3,021
    Trophy Points:
    281
    Come on. They just need to show the actual GPU. Or at least something relevant. Until benches, I could not care less for it.
     
    Georgel, Apr 23, 2016
    #1975
    TomJGX and HTWingNut like this.
  26. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    Right. That yellow image means nothing. Show us actual PCB and benchmarks.
     
    HTWingNut, Apr 23, 2016
    #1976
    Georgel likes this.
  27. Zero989

    Zero989 Notebook Virtuoso

    Reputations:
    910
    Messages:
    2,836
    Likes Received:
    583
    Trophy Points:
    131
    It is indeed. I doubt we will see that one at launch though, or I'd be really surprised.
     
    Zero989, Apr 23, 2016
    #1977
  28. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Damn them. So much re-branding... Now with VRAM.

    Next year GDDR5Z.
     
    J.Dre, Apr 23, 2016
    #1978
  29. Mr Najsman

    Mr Najsman Notebook Deity

    Reputations:
    600
    Messages:
    931
    Likes Received:
    697
    Trophy Points:
    106
    Mr Najsman, Apr 25, 2016
    #1979
    deadsmiley, Robbo99999 and jaybee83 like this.
  30. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    right, so the only difference between 1080 and 1080 Ti will be GDDR5X? and probably 10 mhz more on the core :p
     
    jaybee83, Apr 25, 2016
    #1980
  31. Mr Najsman

    Mr Najsman Notebook Deity

    Reputations:
    600
    Messages:
    931
    Likes Received:
    697
    Trophy Points:
    106
    More and more leaks coming out. Wanna start a poll when the first benchmark comes? :D
    I say May 12th.
     
    Mr Najsman, Apr 25, 2016
    #1981
  32. Georgel

    Georgel Notebook Virtuoso

    Reputations:
    945
    Messages:
    2,378
    Likes Received:
    3,021
    Trophy Points:
    281
    Even if they would only fix the drivers, it could mean a 40% increase in benchmarks, as the latest drivers are all bad. So very bad.

    And we have no bench data yet. They change, add make, but they forget to bench.

    Who cares about new technology, about new power consumption, about how many transistors are in there, or about other technical specifications. These are for marketinfools. We need benchmarks and performance indicators, not them to tell us "You know we doubled the number of cuda cores, but it would probably be the same performance with less electricity". I need more performance. I don't even understand why they don't make the main chip a bit larger, to consume a bit more electricity, and to head a bit less (It should in these conditions, and should allow much better OC)
     
    Last edited: Apr 25, 2016
    Georgel, Apr 25, 2016
    #1982
  33. HTWingNut

    HTWingNut Potato

    Reputations:
    21,580
    Messages:
    35,370
    Likes Received:
    9,877
    Trophy Points:
    931
    So GTX 980 has GDDR5X as well then? Both show 6GA77 Z9TXT RAM on 980 and 1080. Since when did 980 get GDDR5X?
     
    HTWingNut, Apr 25, 2016
    #1983
    Georgel likes this.
  34. Georgel

    Georgel Notebook Virtuoso

    Reputations:
    945
    Messages:
    2,378
    Likes Received:
    3,021
    Trophy Points:
    281
    They might have photoshopped something wrong there. :D
     
    Georgel, Apr 25, 2016
    #1984
    TomJGX and Prema like this.
  35. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    http://www.computerbase.de/2016-04/...scal-gp104-400-mit-8-gbyte-gddr5x-von-micron/

    so, according to this gp104-400 will be the 1080 and gp104-200 the 1070. i wonder if the gp104-150 is the 1080M.... ;)

    as for memory, id wager:
    980-1080 => 8gbs - 10gbs going from gddr5 to gddr5x(see above article)
    970-1070 => 7gbs - 8gbs, both gddr5 (see above article)
    980M-1080M => 5gbs - 7gbs ? that would be nice...

    Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk
     
    jaybee83, Apr 25, 2016
    #1985
  36. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    It's all photoshopped. Same thing happened with Maxwell. Attention whores. :p
     
    J.Dre, Apr 25, 2016
    #1986
  37. tgipier

    tgipier Notebook Deity

    Reputations:
    203
    Messages:
    1,603
    Likes Received:
    1,578
    Trophy Points:
    181
    Heh? Thats a pic of pascal including the mem on a 980 pcb. Not sure whats the point of even comparing them but whatever. Its rumor mill though...

    We see the performance when the pricing leaks.
     
    tgipier, Apr 25, 2016
    #1987
  38. Ashtrix

    Ashtrix ψυχή υπεροχή

    Reputations:
    2,376
    Messages:
    2,081
    Likes Received:
    3,281
    Trophy Points:
    281
    I guess most of us will be left wowing if the perf is really up to mark and will skip the gen 1 release, shelling out $$$ again for the big green milk machine is not optimistic...The other side is ngreedia gimping cards on purpose to force the crap out of us is inevitable, hope though that will not be the case.
     
    Ashtrix, Apr 25, 2016
    #1988
    TBoneSan and Georgel like this.
  39. PrimeTimeAction

    PrimeTimeAction Notebook Evangelist

    Reputations:
    250
    Messages:
    542
    Likes Received:
    1,138
    Trophy Points:
    156
    So any updates/confirmation/speculations on the rumored Nvidia pascal launch event next week?
     
    PrimeTimeAction, Apr 27, 2016
    #1989
  40. nick81

    nick81 Notebook Evangelist

    Reputations:
    136
    Messages:
    386
    Likes Received:
    101
    Trophy Points:
    56
    It's not rumored, it's confirmed. Can't find any date/time for it though.
    Invites were distributed during GTC to the press and it's estimated to happen "a couple" weeks before Computex so best guess is mid-May
     
    nick81, Apr 28, 2016
    #1990
  41. HaloGod2012

    HaloGod2012 Notebook Virtuoso

    Reputations:
    766
    Messages:
    2,066
    Likes Received:
    1,725
    Trophy Points:
    181
    This launch is just for desktop parts right ? Or is mobile also part of this event ?
     
    HaloGod2012, Apr 28, 2016
    #1991
  42. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    Don't know about mobile. Desktop for sure.

    They launched 780 at the same event and the 780M a week later, so never know. If they do launch mobile this early, it'll likely have GDDR5X.
     
    J.Dre, Apr 28, 2016
    #1992
  43. aarpcard

    aarpcard Notebook Deity

    Reputations:
    606
    Messages:
    1,129
    Likes Received:
    284
    Trophy Points:
    101
    How soon after launch will cards be available? Everything I've been reading has been saying that no one will have production quantities of GDDR5X memory chips until well into the summer. . . .
     
    aarpcard, Apr 28, 2016
    #1993
  44. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    theres not gonna be any nextgen vram for mobile cards, forget it. its all reserved for desktop gpus!

    Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk
     
    jaybee83, Apr 28, 2016
    #1994
    Ashtrix likes this.
  45. Kevin

    Kevin Egregious

    Reputations:
    3,289
    Messages:
    10,780
    Likes Received:
    1,782
    Trophy Points:
    581
    We'll see HBM2 mobile chips in 2017. My GTX 980M will last me at least that long, I'm good.
     
    Kevin, Apr 28, 2016
    #1995
    jaybee83 and hmscott like this.
  46. jaybee83

    jaybee83 Biotech-Doc

    Reputations:
    4,125
    Messages:
    11,571
    Likes Received:
    9,149
    Trophy Points:
    931
    sounds about right :)

    Sent from my Huawei Mate 8 NXT-AL10 using Tapatalk
     
    jaybee83, Apr 29, 2016
    #1996
  47. nick81

    nick81 Notebook Evangelist

    Reputations:
    136
    Messages:
    386
    Likes Received:
    101
    Trophy Points:
    56
    nick81, Apr 29, 2016
    #1997
    Cloudfire, hmscott, i_pk_pjers_i and 2 others like this.
  48. Robbo99999

    Robbo99999 Notebook Prophet

    Reputations:
    4,346
    Messages:
    6,824
    Likes Received:
    6,112
    Trophy Points:
    681
    Robbo99999, Apr 29, 2016
    #1998
  49. HaloGod2012

    HaloGod2012 Notebook Virtuoso

    Reputations:
    766
    Messages:
    2,066
    Likes Received:
    1,725
    Trophy Points:
    181
    So MSI guarantees two upgrade cycles. I wonder if Sager does something similar? I really hope I can upgrade to mobile pascal in 2017 even though I have no need for more power :)
     
    HaloGod2012, Apr 29, 2016
    #1999
    hmscott likes this.
  50. J.Dre

    J.Dre Notebook Nobel Laureate

    Reputations:
    3,700
    Messages:
    8,323
    Likes Received:
    3,820
    Trophy Points:
    431
    I'm pretty much tired of speculating lol. Just waiting for the events now to know what's going on. There's so much misinformation.
     
    J.Dre, Apr 29, 2016
    #2000
    jaybee83, Ashtrix, Kade Storm and 2 others like this.
← Previous pageNext page →