1. Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

AMD Bulldozer Eng. Sample leaked, benched

Discussion in 'General Hardware' started by twilyth, Jun 10, 2011.

Thread Status:
Not open for further replies.
  1. Velvet Wafer

    Velvet Wafer New Member

    Joined:
    Jun 16, 2009
    Messages:
    5,123 (2.62/day)
    Thanks Received:
    990
    Location:
    North of Germany
    He had to, Nvidia has already implemented countermeasures against the hack the fifth time... also, after i had shot 2 OSes,due to the complex implementation process, i whined so much in the SLI Hack thread, that he made it a one click updater... which about anyone can use;)

    its Anatoly btw, he is a russian :)
     
    theeldest says thanks.
  2. twilyth Guest

    meh, not sure there's anything new here, but I'll post it anyway. Benchmarks are at the end but all the good stuff is blocked out.

     
    Last edited by a moderator: Nov 20, 2013
  3. Damn_Smooth

    Damn_Smooth New Member

    Joined:
    May 16, 2011
    Messages:
    1,435 (1.14/day)
    Thanks Received:
    478
    Location:
    A frozen turdberg.
    I don't see what you're talking about. :confused:
     
    Last edited by a moderator: Nov 20, 2013
  4. repman244

    repman244

    Joined:
    Apr 7, 2011
    Messages:
    1,104 (0.85/day)
    Thanks Received:
    456
    Well he did say that all the good stuff is blocked out :laugh:

    On a more serious note, I don't see anything either.
     
    yogurt_21 and Damn_Smooth say thanks.
  5. Damn_Smooth

    Damn_Smooth New Member

    Joined:
    May 16, 2011
    Messages:
    1,435 (1.14/day)
    Thanks Received:
    478
    Location:
    A frozen turdberg.
    Somewhere in that white square lies the secret. We must decode it.
     
  6. seronx

    seronx

    Joined:
    Jul 10, 2010
    Messages:
    994 (0.63/day)
    Thanks Received:
    219
    Location:
    USA, Arizona
    [yt]-flGRobFzsQ[/yt]

    Solved!
     
    Last edited by a moderator: Nov 20, 2013
    Damn_Smooth says thanks.
  7. Damn_Smooth

    Damn_Smooth New Member

    Joined:
    May 16, 2011
    Messages:
    1,435 (1.14/day)
    Thanks Received:
    478
    Location:
    A frozen turdberg.
    Now I'm dissapointed. I don't trust this guy at all. It was nice seeing those boards though.
     
  8. seronx

    seronx

    Joined:
    Jul 10, 2010
    Messages:
    994 (0.63/day)
    Thanks Received:
    219
    Location:
    USA, Arizona
    Ya, doing the maths it seems that Bulldozer actually has a High IPC and a low clock...


    2.4-3.0 GHz like the Phenom 1s

    Phenom I(4C) to K15(4C) = exactly 50% increase in performance
    1 to 1.5

    and 8C is
    1 to 3

    Stumbled on some old footnotes

    1 AMD CMT core can do(2m = 8t)(4m = 16t)
    2x Integer(1 execution core can do this)
    or
    2x Floating Point(1 execution core can do this)

    1 Nehalem core can do
    1+1x Integer
    or
    1+1x Floating Point

    1 SB Core can do(4 core = 8 threads)
    1+1 x Integer
    or
    1+1 x Floating Point

    To make it easier to understand

    AMD BD:
    1 module can assign
    2 Integer to 1 core(2 Nehalem Threads/ somewhat) while 2 Floating Point to 1 core(2 Nehalem threads/ somewhat)

    Intel Nehalem:
    1 core can assign
    1 Integer to 1 thread while 1 Floating Point to 1 thread

    Sandy bridge is tricky, I can't find any footnotes comparing to it
    I can only assume that it is like AMD BD:
    1 core can assign
    1 Integer to 1 thread while 1 Floating Point to 1 thread

    Bulldozer = Sandy Bridge in IPC

    But actual synthetic core performance is weird

    Bulldozer 8C @ 2.6GHz is equal to Sandy Bridge 8T @ 3.4GHz
    800MHz difference

    Leading me to believe Sandy Bridge is just a faster Nehalem/ somewhat
     
    Last edited: Jun 28, 2011
  9. heky

    heky

    Joined:
    Sep 9, 2009
    Messages:
    888 (0.47/day)
    Thanks Received:
    144
    Location:
    Slovenia, Europe
    If the video is legit, BD will suck. Voltage is 1.4 to 1.5v, the tdp is almost 190W and the memory tests just suck compared to SB.
     
  10. Heavy_MG New Member

    Joined:
    May 18, 2011
    Messages:
    48 (0.04/day)
    Thanks Received:
    10
    The tested chip is a ES.
    An ES chip will use higher voltage,is buggy thus doesn't run as well as it should, and have a higher TDP than the final product released to the consumer.
     
    Velvet Wafer says thanks.
  11. WhiteLotus

    WhiteLotus

    Joined:
    Jul 30, 2007
    Messages:
    6,551 (2.48/day)
    Thanks Received:
    857
    You mean, Bulldozer is going to be streets ahead of everything else that AMD have ever put out, but not quite as good as Intel's current high end/future offerings.

    Honestly I don't see how Bulldozer is going to suck just because it's not going to be the best. That equation does not compute with me. It's still going to be a very powerful chip that will be over kill for MANY MANY MANY systems.
     
    Velvet Wafer says thanks.
  12. Horrux

    Horrux

    Joined:
    Jun 2, 2011
    Messages:
    735 (0.59/day)
    Thanks Received:
    124
    ^^^^
    I certainly hope so...
     
  13. xenocide

    xenocide

    Joined:
    Mar 24, 2011
    Messages:
    2,149 (1.64/day)
    Thanks Received:
    463
    Location:
    Burlington, VT
    The point he was getting at is people who are buying a $300 CPU expect it to at least be on par with the competition's offering around the same price range. Sure, it would be great for many applications, but if that were the case why not save money and just go for a last-gen option?
     
    heky says thanks.
  14. Heavy_MG New Member

    Joined:
    May 18, 2011
    Messages:
    48 (0.04/day)
    Thanks Received:
    10
    I agree,while BD might not blow SB away,it will be a lot faster than anything AMD has out today. AMD's goal isn't to be the best,but to provide a competitive product that will satisfy most users needs. However,for $300 it should at least be on par with a 2500K,if not,users will still buy from Intel.
     
  15. WhiteLotus

    WhiteLotus

    Joined:
    Jul 30, 2007
    Messages:
    6,551 (2.48/day)
    Thanks Received:
    857
    Then take that up with the sales team. Not the chip itself. Instead of saying, omg this chip is going to suck, say hmm AMD seem to be charging too much.
     
  16. heky

    heky

    Joined:
    Sep 9, 2009
    Messages:
    888 (0.47/day)
    Thanks Received:
    144
    Location:
    Slovenia, Europe
    I am not saying BD will suck becouse it will not be the most powerfull, but becouse if the video is legit, it will use 2x the power of SB and be inferior or on par with SB. For the same 32nm and the same price, that is just unacceptable!
     
  17. repman244

    repman244

    Joined:
    Apr 7, 2011
    Messages:
    1,104 (0.85/day)
    Thanks Received:
    456
    How do you know how much power does it consume :confused: the 186W in CPU-Z is a bug, HWiNFO says ~125W TDP (and AFAIK BD will be able to shut down a whole module if it doesn't use + it's on 32nm). The voltage doesn't say anything (don't compare it to Intel).

    And don't forget, you are paying for 8 cores (and there will be 4 and 6 cores versions, which will be cheaper)
     
    Last edited: Jun 27, 2011
  18. heky

    heky

    Joined:
    Sep 9, 2009
    Messages:
    888 (0.47/day)
    Thanks Received:
    144
    Location:
    Slovenia, Europe
    Voltage x amperage = wattage so unless BD draws minimal amounts of current, it will consume heaps of energy!

    And the fact you state you pay for 8 cores, only makes it suck even more, becouse it takes 8 AMD cores(even though they are not real cores) to compete with 4 Intel cores + HT.
     
  19. WhiteLotus

    WhiteLotus

    Joined:
    Jul 30, 2007
    Messages:
    6,551 (2.48/day)
    Thanks Received:
    857
    Using what as evidence?

    A chip that might not even be a real BD, some benchmarks that could be false, hell, the chip itself might be a very early development chip, like a Beta chip or something... a work in progress?

    Stop dismissing the chip before anything official comes out, and until it actually gets released so the public can bench and compare without being biased.

    :shadedshu
     
    Heavy_MG says thanks.
  20. repman244

    repman244

    Joined:
    Apr 7, 2011
    Messages:
    1,104 (0.85/day)
    Thanks Received:
    456
    I know how to get wattage, and we still don't know the current. And don't forget AMD is using SOI which can take more voltage than bulk. And of course it will consume more power when at 100%, it's 8 cores.

    If you ask me SB will be torn appart in multithread (and that's where you need the cores), singlethread it maybe comes close.

    You also forgot I mentioned 4 and 6 core versions which I believe will be priced competitively against Intel's lineup .
     
  21. Horrux

    Horrux

    Joined:
    Jun 2, 2011
    Messages:
    735 (0.59/day)
    Thanks Received:
    124
    Aren't you guys confusing current drawn and thermal dissipation? They are different... If you are comparing an estimate of BD's current draw to SB's TDP, of course you will have a huge disparity. They aren't the same thing at all. Although both are measured in watts.
     
    Heavy_MG, Velvet Wafer and cadaveca say thanks.
  22. theeldest

    theeldest

    Joined:
    Feb 7, 2006
    Messages:
    652 (0.20/day)
    Thanks Received:
    140
    Location:
    Austin, TX
    Also, the chip is overclocked...

    If I recall correctly, overclocking increases current draw...
     
  23. repman244

    repman244

    Joined:
    Apr 7, 2011
    Messages:
    1,104 (0.85/day)
    Thanks Received:
    456
    That's why I used bold text for TDP.

    And like you said TDP isn't very accurate for measuring power consumption. And also, Intel's TDP is not the same as AMD's TDP, AFAIK.

    ACP should be more accurate in representing the power usage of the chip, but so far I only saw that in the HP Proliant server specification
     
  24. Horrux

    Horrux

    Joined:
    Jun 2, 2011
    Messages:
    735 (0.59/day)
    Thanks Received:
    124
    Yeah I can't remember which one uses what, but one is "Thermal Design Power" and one is "Typical Dissipated Power" or somesuch? Bleh, I knew this stuff years ago, now it's all a bunch of fuzz... Can someone clarify?
     
  25. heky

    heky

    Joined:
    Sep 9, 2009
    Messages:
    888 (0.47/day)
    Thanks Received:
    144
    Location:
    Slovenia, Europe
    What is wrong with you? Cant you read? I said IF the video is legit!!! Oh and believe me, there is a reason why the BD line got delayed. The frequency of the produced chips was too low to compete with current intel offerings, so AMD is preparing a new stepping!
    Mark my words, if BD doesnt beat SB, AMD is in truble.

    I am not trying to diss the chip, i was a AMD supporter once(socket 939), but for 2 gens now, they just cant find the right path.IMHO
     

Currently Active Users Viewing This Thread: 1 (0 members and 1 guest)

Thread Status:
Not open for further replies.

Share This Page