AMD Polaris 10 GPU To Offer Near 980 Ti Performance For 299 USD?

Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Apr 26, 2016.

  1. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,688
    Likes Received:
    2,282
    GPU:
    Zotac GTX980Ti OC
    Here is a interesting comparison, current scenario as of 16.Feb 2016

    [​IMG]


    moderate factory OC 1216Mhz with real boost up to 1418Mhz. Manual OC is at 1517MHz/8.4GHz ram
    [​IMG]

    https://www.techpowerup.com/reviews/ASUS/GTX_980_Ti_Matrix/26.html

    Everything is left far behind once 980Ti hits 1400mhz (22fps vs stock 980Ti), I quite sure both Polaris10 and GTX1080 won't surpass that either..
     
    Last edited: Apr 29, 2016
  2. Ieldra

    Ieldra Banned

    Messages:
    3,490
    Likes Received:
    0
    GPU:
    GTX 980Ti G1 1500/8000

    8400mhz memory isn't very moderate, i don't know about battlefield 3 but performance seems to scale very linearly with clocks on a maxwell, on average if you're looking at benchmarks run at 1200mhz you count on a 20-25% improvement once overclocked,
     
  3. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,688
    Likes Received:
    2,282
    GPU:
    Zotac GTX980Ti OC
    What, I meant first part by factory OC 1216mhz/ ram 1800mhz or 7200mhz as moderate (Blue line in chart). :)


    I also flashed mine to near Amp!Extreme version freq. 1240MHz and 7220MHz as factory OC, should be enough for a while :nerd:
     
    Last edited: Apr 29, 2016
  4. Ieldra

    Ieldra Banned

    Messages:
    3,490
    Likes Received:
    0
    GPU:
    GTX 980Ti G1 1500/8000
    Yeah but it's the boost that counts :p 1418mhz is a pretty solid overclock :p
     

  5. Picolete

    Picolete Master Guru

    Messages:
    438
    Likes Received:
    208
    GPU:
    R9 290 Sapphire Tri-x
    Total war, now that could really benefit from DX12 and async.
     
  6. Ryu5uzaku

    Ryu5uzaku Ancient Guru

    Messages:
    7,374
    Likes Received:
    472
    GPU:
    6800 XT
    Total war with dx12 holy crap.
     
  7. Ieldra

    Ieldra Banned

    Messages:
    3,490
    Likes Received:
    0
    GPU:
    GTX 980Ti G1 1500/8000
    Hopefully it'll be less of a stuttering mess compared to previous games

    Should put my 6core to good use
     
  8. Clouseau

    Clouseau Ancient Guru

    Messages:
    2,808
    Likes Received:
    486
    GPU:
    ZOTAC AMP RTX 3070
    Does not matter how reliable the benchmarks or theoretical performance is. Am picking one up. Bound to be leaps and bounds better than the current GPU I have. If AMD is around long enough for Navi to hit the market, that will be picked up as well. New rig is in the plans. Only have case so far.
     
  9. ender79

    ender79 Member Guru

    Messages:
    127
    Likes Received:
    0
    GPU:
    Zotac GTX 970 ~ 1455 Mhz
    AMD did not released a new arhitecture from 2011, when moved from 40nm WLIW4 to 28 nm named it GCN . All iterations 1.0, 1.1 ,1.2 are just the same GCN arhitecture with some little improvements from a generation to another.
    Now they are moving to 14nm .... maybe is time for a new arhitecture , because GCN can't compete even with Maxwell . In DX12, GCN may have an edge , but just because the Async Shaders. Meanwhile Pascal have also AsyncShaders in hardware, so I doubt AMD can compete without a fresh new arhitecture, name it Polaris or something else... they need to be competitive again.
     
  10. Ieldra

    Ieldra Banned

    Messages:
    3,490
    Likes Received:
    0
    GPU:
    GTX 980Ti G1 1500/8000
    You've already decided? Without seeing performance metrics, cost, tdp...

    Even navi...

    Yeah but by that token nvidia hasn't changed architecture since g80.

    Any architectural change usually constitutes 'a new architecture', just to distinguish it from the others.

    Pascal is an iterative improvement over maxwell, maxwell is an iterative improvement over kepler etc etc

    AMD's dx12 performance isn't just because of async, the main difference is the removal of cpu bottleneck (for the most part)
     
    Last edited: Apr 30, 2016

  11. leszy

    leszy Master Guru

    Messages:
    330
    Likes Received:
    20
    GPU:
    Sapphire V64 LC
    Core Improvements in Kepler:
    More CUDA cores, up from 512 to 1536
    GPU Boost for automatic overclocking
    Refined instruction pipeline
    Adaptive Vsync
    Low power consumption

    Core improvements in GCN 2.0 (GCN1.1):
    Introduced Shader Engine (SE). A Shader Engine comprises one geometry processor, up to 11 CUs (Hawaii chip), rasterizers, ROPs, and L1 cache.
    Graphics Command Processor, the 8 ACEs, the L2 cache and memory controllers as well as the audio and video accelerators, the display controllers, the 2 DMA controllers and the PCIe interface.
     
  12. leszy

    leszy Master Guru

    Messages:
    330
    Likes Received:
    20
    GPU:
    Sapphire V64 LC
    PS. I would rather say that it was Kepler, Maxwell and Pascal should be treated as pumped and polished Fermi. However, in subsequent iterations of GCN, we see a really big differences in architecture.
     
  13. Clouseau

    Clouseau Ancient Guru

    Messages:
    2,808
    Likes Received:
    486
    GPU:
    ZOTAC AMP RTX 3070
    Nvidia has been my reference point in graphics for a long time. Do not have the time to invest in playing games like I used to. Nvidia would be a waste of money to me. The improvements it has to offer is wasted on me at this point in time. I have no ego to bruise. AMD in the past has come close enough at a cheaper price point. That being said will not cut my nose off to spite my face. Not set in hardened concrete but foundation has been there.
     
  14. Denial

    Denial Ancient Guru

    Messages:
    13,973
    Likes Received:
    3,732
    GPU:
    EVGA RTX 3080

    Rofl

    I think you better go look up Kepler/Maxwell changes again. You missed like all of them.
     
  15. Ieldra

    Ieldra Banned

    Messages:
    3,490
    Likes Received:
    0
    GPU:
    GTX 980Ti G1 1500/8000
    :p goatman to the rescue

    SM structures has changed consistently since Fermi, with Pascal gone full circle and back to 64/sm

    Kepler lost some of its hardware scheduling logic in favor of fixed latency pipeline etc etc

    Big changes

    And Closeau, 'I won't cut off my nose to spite my face' is one of the best sayings I've ever heard.
     

  16. TBPH

    TBPH Active Member

    Messages:
    78
    Likes Received:
    0
    GPU:
    MSI GTX 970 3.5+.5GB
    Polaris 10 will have 8GB, not 6. It's a 256-bit interface.
     

Share This Page