Page 10 of 432 FirstFirst ... 891011122060110 ... LastLast
Results 91 to 100 of 4311

Thread: Polaris 10 size / performance estimation

  1. #91
    Senior Member
    Join Date
    Jan 2010
    Location
    London
    Posts
    3,310
    That Ethereum mining set up- is it the same pain in the a*** to set up as it was with Bitcoins? I think I still have around ?80 from bitcoin mining in now bankrupt MTGox.

  2. #92
    Senior Member
    Join Date
    Jan 2014
    Posts
    355
    Quote Originally Posted by muziqaz View Post
    That Ethereum mining set up- is it the same pain in the a*** to set up as it was with Bitcoins? I think I still have around ?80 from bitcoin mining in now bankrupt MTGox.
    It is definitely harder to use, which is why the hashrate is lower.

    Granted, it's pretty easy once you've tinkered for an hour or so.

    You'd need to download Geth (Go Ethereum client), setup a wallet with passphrase, run it in RPC mode (so that the miner can reach it), then download Ethereum C++ client, install, and run ethminer -G from command line. You can select specific opencl devices as well, to easily use multiple GPUs. It certainly isn't as pretty or simple to use as the Litecoin client was.

  3. #93
    Senior Member
    Join Date
    Jan 2014
    Posts
    355
    I mentioned Ethereum was at $3.70/coin this morning. It is at $4.25 as of right now.

  4. #94
    Senior Member
    Join Date
    Dec 2010
    Location
    Switzerland Geneva Cern
    Posts
    2,951
    Recently i saw some rumors about a possible 232mm2 amd gpu.

    Such a gpu would end up having around ~4.6 - 5 billion transistors so it would probably end up having similar performance to 285 - 380 - 380x.

    According to amd claims about at least 2x efficiency increase tdp should be around ~120W.

    Do you guys believe that this is polaris 11 gpu?
    Current rig: i5 3570 @3.8GHz, Asus 7750, Dell 24'' 2560X1440, Corsair Vengence 16GB Ram DDR3 @1600MHz, Chieftec 550W PSU, 480GB OCZ SSD + 500GB Seagate + 1TB WD hd, Windows 10 64-bit

  5. #95
    Senior Member
    Join Date
    Jan 2010
    Posts
    3,622
    Yeah, I think the 232mm^2 rumour is pointing to the Polaris 11 GPU.

    I posted on a different thread that this probably means it will have 3072 shaders, give or take 512 depending on various design aspects, etc.

  6. #96
    Senior Member
    Join Date
    Dec 2010
    Location
    Switzerland Geneva Cern
    Posts
    2,951
    Quote Originally Posted by JeeBee View Post
    Yeah, I think the 232mm^2 rumour is pointing to the Polaris 11 GPU.

    I posted on a different thread that this probably means it will have 3072 shaders, give or take 512 depending on various design aspects, etc.
    There is no way its going to have more than 3072 shaders. Even this number is quite large. 290x has over 6 billion transistors and has less than 3000 cores.
    Current rig: i5 3570 @3.8GHz, Asus 7750, Dell 24'' 2560X1440, Corsair Vengence 16GB Ram DDR3 @1600MHz, Chieftec 550W PSU, 480GB OCZ SSD + 500GB Seagate + 1TB WD hd, Windows 10 64-bit

  7. #97
    Senior Member
    Join Date
    Feb 2010
    Location
    Planet Earth, Solar System, Milky Way
    Posts
    3,521
    Quote Originally Posted by Nerdmaster View Post
    Recently i saw some rumors about a possible 232mm2 amd gpu.

    Such a gpu would end up having around ~4.6 - 5 billion transistors so it would probably end up having similar performance to 285 - 380 - 380x.

    According to amd claims about at least 2x efficiency increase tdp should be around ~120W.

    Do you guys believe that this is polaris 11 gpu?
    Yes, I believe but your performance expectation/estimation is extremely low.
    If that 14nm 232 sq mm chip equals a 28nm 366 sq mm, then I would call this a uber fail.

    According to the information we have about the 14nm LPP process, and based on transistor density increase, a 232mm? GPU would be roughly equivalent to a 464mm? 28nm processor ? at the same TDP levels. Since we already know that AMD is going to be focusing not just on performance but power efficiency as well ? this number could be be much higher, in fact we will discuss the number AMD is using below.

    A 232mm? die multiplied by 2.5 times gives us the performance range of a 28nm 580mm? chip (at the same TDP level) which would be pretty ************ impressive.
    Read more: http://wccftech.com/amd-polaris-gpu-...#ixzz40FjidpR0

  8. #98
    Senior Member
    Join Date
    Dec 2010
    Location
    Switzerland Geneva Cern
    Posts
    2,951
    Quote Originally Posted by GeorgiD View Post
    Yes, I believe but your performance expectation/estimation is extremely low.
    If that 14nm 232 sq mm chip equals a 28nm 366 sq mm, then I would call this a uber fail.



    Read more: http://wccftech.com/amd-polaris-gpu-...#ixzz40FjidpR0
    The 366 sq mm 28nm chip was the last generation of 28nm. First 28nm gpus offered way higher density because engineers improved drastically transistor density.

    Evergreen (40nm first gen) : 6.4 million transistors / mm2
    Cayman (40nm last gen) : 6.8 million transistors / mm2
    Tahity : 11 million transistors / mm2
    Pircairn : 13.2 million transistors / mm2
    Tonga : 13.9 million transistors / mm2
    Fury : 14.93 million transistors / mm2

    My hypothesis on density increase is based on doubling tahity transistor density which was the first arch developed for 28nm. I believe that if they manage to make a card with 22 million transistors / mm2 this wouldnt be a failure at all. You can compare this trend on wikipedia which has all transistor counts for multiple generations of gpus. You can read my first post on this thread for more details.

    If you double cayman transistor density you end up with 13.6 million transistors / mm2 which is significantly higher than tahiti achieved.
    Current rig: i5 3570 @3.8GHz, Asus 7750, Dell 24'' 2560X1440, Corsair Vengence 16GB Ram DDR3 @1600MHz, Chieftec 550W PSU, 480GB OCZ SSD + 500GB Seagate + 1TB WD hd, Windows 10 64-bit

  9. #99
    Senior Member
    Join Date
    Jan 2010
    Posts
    3,622
    Well you can do the maths yourself.

    Polaris 10 is between 100 and 120 mm^2. Estimates put it between 1024 and 1280 shaders. A large portion of the chip is "unshader", 40-50% looking at 28nm chips with similar shader counts.

    So scale up the shader area whilst keeping the unshader the same to this 232mm^2. How many shaders do you get?

    Also, it has been estimated that 14nm has around twice the transistor density of last-generation 28nm.

    More? From https://www.chipworks.com/about-chip...ess-technology
    Earlier this year, we completed a limited analysis of the high density SRAM on the AMD RadeonTM HD 7970 215-0821060 graphics processor, which was fabricated with TSMC’s HP process. Our TEM analysis confirmed the 215-0821060 transistor structure was identical to that seen in the Altera Stratix V device, as would be expected since both are based on the TSMC 28 nm HP process. The 215-0821060 features a 0.16 ?m2 6T-SRAM with the transistors arranged in a uniaxial layout.
    TSMC 28nm SRAM Cell: 0.16 ?m^2 in 2012 (so this may have got smaller over time)
    Samsung 14nm SRAM Cell: 0.064 ?m^2 in 2014 (LPE) and around 0.05?m^2 in 2015 (LPP?) (http://www.clivemaxfield.com/area51/...ance-x-800.png)

  10. #100
    Quote Originally Posted by Nerdmaster View Post
    Recently i saw some rumors about a possible 232mm2 amd gpu.

    Such a gpu would end up having around ~4.6 - 5 billion transistors so it would probably end up having similar performance to 285 - 380 - 380x.

    According to amd claims about at least 2x efficiency increase tdp should be around ~120W.

    Do you guys believe that this is polaris 11 gpu?

    Supposedly clocks could be a good deal higher on 14nm...

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
WordPress Appliance - Powered by TurnKey Linux