NVIDIA Titan V Graphics Card Benchmarks

Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Dec 9, 2017.

  1. erm

    erm Master Guru

    Messages:
    396
    Likes Received:
    1
    GPU:
    GeForce GTX 560 1GB
  2. PrMinisterGR

    PrMinisterGR Ancient Guru

    Messages:
    7,868
    Likes Received:
    736
    GPU:
    Inno3D RTX 3090
    40% extra was very much heard off, and at the same price point as the previous top card. Nvidia is raising the price point every generation, and all the dim bulbs find it normal. The only good thing is that this will most likely be the last big consumer GPU they make.

    Yes you can. The original Titan was a similar card. All Vega cards also work with all of their professional hardware enabled.
     
  3. howiec

    howiec Member

    Messages:
    14
    Likes Received:
    2
    GPU:
    GTX 970 HOF
    1) You can get a glimpse of gaming performance but you should NOT compare the price of a workstation or high-end AI compute card with tons of tensor cores and say anything yet about the 1180 (or whatever the name will be) except that it's slightly likely to be more expensive than the previous price point simply because they have little effective competition and because it's a new architecture.

    2) You cannot hate Nvidia for pushing the boundaries of cutting edge technology (AI) because ultimately that is the future and there's a lot of $ to be made. Yes, gaming is still in their hearts too, I presume, so we should expect a decently configured and performing gaming version... eventually.

    3) It's definitely not easy to deliver such significant performance gains WITH lower clocks AND increased power efficiency. We should be somewhat grateful for the fact that Nvidia is developing new tech that doesn't suck and that actually does deliver.
     
    D3Master likes this.
  4. Razoola

    Razoola Member Guru

    Messages:
    125
    Likes Received:
    2
    GPU:
    GTX980 (MSI gaming)
    Anyone know yet what HW video decoding formats these new volta GPUs can handle?
     

  5. Agent-A01

    Agent-A01 Ancient Guru

    Messages:
    11,408
    Likes Received:
    922
    GPU:
    ASUS 3080 Strix H20
    blackguyconfused.jpg
     
  6. PrMinisterGR

    PrMinisterGR Ancient Guru

    Messages:
    7,868
    Likes Received:
    736
    GPU:
    Inno3D RTX 3090
    I can't see it adding any Cuda cores, if anything, it will probably have even less. There is no way NVIDIA is selling gamers that much hardware this gen, for extra CUDA cores it will require a larger die, and I can't see them going over ~650mm2
     
  7. D3Master

    D3Master Member

    Messages:
    21
    Likes Received:
    4
    GPU:
    SLI Gigabyte GTX 1070 G1
    You certainly do not know what it is and what professional cards are when you call it "a money grab from rich fanboys".
     
  8. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,209
    Likes Received:
    1,941
    GPU:
    Zotac GTX980Ti OC
  9. Agent-A01

    Agent-A01 Ancient Guru

    Messages:
    11,408
    Likes Received:
    922
    GPU:
    ASUS 3080 Strix H20
    Clearly you’re lacking technical understanding. This is a compute card with tensor cores that are tied to main clock speed hence GV100 is underclocked vs pascal gaming chips.

    The gaming version will have higher clock speeds and no fp64 or tensor cores included.

    A gv102 will be exactly that. Ampere is pure speculation, core architecture will still be Volta.

    All of this is common sense so sounds like you’re deliberately acting dumb
     
  10. Agent-A01

    Agent-A01 Ancient Guru

    Messages:
    11,408
    Likes Received:
    922
    GPU:
    ASUS 3080 Strix H20
    Now you're just trolling.

    Like vega has anything to do with Volta.

    That hole you're digging is getting bigger and bigger lol.
     
    fantaskarsef likes this.

  11. warlord

    warlord Ancient Guru

    Messages:
    2,761
    Likes Received:
    927
    GPU:
    Null
    Boring. Everyone bashes tech companies nowadays. Are you better/smarter than Jen-Hsun Huang? He does his job right. None of us here could take his place and work. Show some respect and admire their work. If you don't like, build your corporation and throw Nvidia/AMD off the competition.

    Volta is great, no, it's huge. Gaming wise you couldn't except all this kind of billions transistors help towards to it. It is a professional chip. Nvidia already put TITAN away from GTX series at the drivers section. Enjoy life and wait for x80 to criticize this gen. It will not be another fiasco like VEGA FE vs VEGA 64.
     
  12. Denial

    Denial Ancient Guru

    Messages:
    13,509
    Likes Received:
    3,044
    GPU:
    EVGA RTX 3080
    It's not a die shrink, it's a optimized variant of 16nmFF. Also it's running at nearly double perf/w compared to Pascal.. I don't really understand how anyone can say that's disappointing.

    This card was clearly launched because they had a number of failed GV100's where they lost a stack of HBM2. Like the previous Titan will most likely be the best selling deep learning card. It's expensive but it's offering up to 3.7x the performance for the same power consumption as the previous GP102 Titan.
     
    yasamoka and Thunk_It like this.
  13. Agent-A01

    Agent-A01 Ancient Guru

    Messages:
    11,408
    Likes Received:
    922
    GPU:
    ASUS 3080 Strix H20
    This isn't a gaming variant of a card.
    Seriously you're either trolling/being deliberately stupid or have a reading comprehension.

    This card is for professional work and is much cheaper than the last tesla while doing the same workloads.
     
    Andrew LB likes this.
  14. Thunk_It

    Thunk_It Master Guru

    Messages:
    288
    Likes Received:
    51
    GPU:
    Asus 2080ti Turbo
    Though the "someone" would no doubt be Hilbert, I too (if it is possible) would be interested in this comparison.
     
  15. Aura89

    Aura89 Ancient Guru

    Messages:
    8,183
    Likes Received:
    1,285
    GPU:
    -
    When that "disappointment" is unjustified.
     

  16. Neo Cyrus

    Neo Cyrus Ancient Guru

    Messages:
    9,927
    Likes Received:
    715
    GPU:
    Asus TUF 3080 OC
    So... any predictions on when nVidia is going to release this with the tensor cores removed, and everything else such as shaders shaders cut by a crap ton, call it a GTX xx80, and sell it for $1K+ USD?

    My GTX 970 has been begging me to euthanize it for months now. But I can't bring myself to buy a 1080 or 1080 Ti at the current retarded miner/monopoly prices considering how old they are. Feels bad man. 30 fps here I come.
     
  17. tensai28

    tensai28 Maha Guru

    Messages:
    1,416
    Likes Received:
    366
    GPU:
    2080ti MSI X TRIO
    Where's the 4k benchmarks?
     
  18. GxCx

    GxCx Member

    Messages:
    17
    Likes Received:
    0
    GPU:
    burned
  19. fantaskarsef

    fantaskarsef Ancient Guru

    Messages:
    12,422
    Likes Received:
    4,696
    GPU:
    2080Ti @h2o
    With the TitanV looking like this, I only wonder about three things:

    How much impact does HBM have on it's performance?
    When the Ti of such a card will be released?
    When the damn waterblocks are available?

    :D
     
  20. Ryu5uzaku

    Ryu5uzaku Ancient Guru

    Messages:
    7,080
    Likes Received:
    284
    GPU:
    6800 XT
    In timespy it beats aftermarket 1080 ti by 20%. It is kind of all over the place. But then again it's only bump in SMs regarding gaming really. And this most likely will be the fastest chip till 7nm from nvidia. Tensor cores do not make the chip big tho. Bump of 30% in SMs makes the die nearly 30% bigger then gp100, since the 12nm node isn't really that much smaller.

    2080 ti or whatever should be more in align with pricing when they cut out fp64 and TC.

    https://www.3dmark.com/3dm/23867564 most likely a driver issue comparing this to this http://www.guru3d.com/articles_pages/msi_gtx_1080_ti_gaming_x_review,28.html

    Or it just does not perform in some tasks any better.

    But anyways from 20 to 40% faster then aftermarket 1080 ti.
     
    Last edited: Dec 11, 2017

Share This Page