AMD Vega GPU

Discussion in 'Video Cards & Monitors' started by Agg, Jan 6, 2017.

  1. power

    power Member

    Joined:
    Apr 20, 2002
    Messages:
    66,090
    Location:
    brisbane
    you realise AMD know how licensing tech works right?
     
  2. rickoles

    rickoles Member

    Joined:
    Aug 25, 2010
    Messages:
    206
    Location:
    Perth 6017
    They also realise it's bad for the consumer.
     
  3. power

    power Member

    Joined:
    Apr 20, 2002
    Messages:
    66,090
    Location:
    brisbane
    nope, it's bad for them because they can't compete.
     
  4. Unframed

    Unframed Member

    Joined:
    Mar 30, 2010
    Messages:
    9,158
    Location:
    Hella south west
    No, I'm completely oblivious.
     
  5. SnooP-WiggleS

    SnooP-WiggleS Member

    Joined:
    Sep 20, 2004
    Messages:
    3,175
    Location:
    Ballarat, Victoria
    I think this article describes the situation well, comparing both an nvidia supplied feature (hairworks with it's absurd tesselation on witcher 3), and amd supplied tressfx (slow on nvidia initially, but patched fairly quickly). Note Polaris and presumably vega have dedicated silicon specifically to cull back absurd tesselation.

    http://wccftech.com/fallout-4-nvidia-gameworks/

    But thankfully gameworks source code is now mostly open source, so at least it's fairly level with AMD provided middleware that was open source from the beginning.

    https://developer.nvidia.com/gameworks-source-github
     
  6. bart5986

    bart5986 Member

    Joined:
    Jan 31, 2006
    Messages:
    4,836
    Location:
    Brisbane
    Not at all.

    Here is a quick pretend example.

    Nvidia spends $1000 per year on tech/drivers/ etc
    AMD Spends $10 per year on tech/drivers etc

    Everything is open source

    Nvidia have to sell their GPU's for $100 more then AMD to pay for staff/research etc.

    AMD can sell for less for the same tech/GPU. AMD wins all marketshare as Nvidia has no benefit.

    Result?

    Nvidia fires most of their staff and just do the basics.

    The result? we all lose.

    Open source works in many situations, but not in the Nvidia/AMD situation.

    Because being able to play Tomb Raider at a higher FPS while having no tech/drivers etc from Nvidia does not mean its worth the same amount. Do you realise how many video games exist?

    And would result in Nvidia firing most of their staff and reducing their budgets to match AMD, which is not enough to make or research any meaningful tech.

    Sounds like AMD's TressFX to me.

    Things like Physx are resource intensive but look good, especially the latest versions are very fair on both systems in terms of performance.

    Having said that its still up to debate as to who wants the effects in games or not, thats not up to me, thats up to game developers and people vote with their money.


    ----------


    Just to make it clear for everyone, the only reason why AMD are into open source is because they do very little research in development for the software side, they don't put a lot of money into it and this is why they can't compete.

    This is normally ok, as AMD normally price accordingly. AMD are the budget GPU company and thats why I loved them for over a decade.
     
    Last edited: Jul 14, 2017
  7. Unframed

    Unframed Member

    Joined:
    Mar 30, 2010
    Messages:
    9,158
    Location:
    Hella south west
    Nvidia might cut back on tech resources but in saying that they wouldn't need to be creating their own tech from the ground up to be equal to/better than their competition. Well not as much. This exploration into an impossible future though doesn't really help our discussion much at all.


    I don't know what to say... "Everything is equal (or better in very few examples) but because it's not Nvidia it's not worth the same money".



    Except TressFX is free software that NV are free to use.


    And Nvidia could do the exact same as AMD.
     
  8. SnooP-WiggleS

    SnooP-WiggleS Member

    Joined:
    Sep 20, 2004
    Messages:
    3,175
    Location:
    Ballarat, Victoria
  9. Unframed

    Unframed Member

    Joined:
    Mar 30, 2010
    Messages:
    9,158
    Location:
    Hella south west
  10. bart5986

    bart5986 Member

    Joined:
    Jan 31, 2006
    Messages:
    4,836
    Location:
    Brisbane
    If Nvidia aren't spending huge $$$ on research and development, who is? Why would either company bother to make anything that won't give them a profit?

    You might not realise this but you don't just send an email out to your development team and ask them to "make technology"

    Nvidia puts crazy amounts of money into research and development.

    This would not happen if it was open source. If Nvidia didn't have to put crazy amounts of money into research and development, why do they?

    Which ran poorly on Nvidia cards and was optimised for the strengths of AMD cards which Nvidia does not have in their cards.

    This is probably why we are seeing AMD making their own versions of Nvidia tech recently.
     
    Last edited: Jul 14, 2017
  11. Unframed

    Unframed Member

    Joined:
    Mar 30, 2010
    Messages:
    9,158
    Location:
    Hella south west
    Fuck, I totally thought that's how it worked. Damn I am sorely mistaken. It couldn't be that they could shift that funding to improving hardware while working on software that has a shared foundation. Either way this is just useless rhetoric at this point.

    Source code is available, Nvidia could adapt it to work as well with their hardware.
     
  12. bart5986

    bart5986 Member

    Joined:
    Jan 31, 2006
    Messages:
    4,836
    Location:
    Brisbane
    two companies collaborating without any funding or staff achieves mostly the same as one company who doesn't make any tech because they have no staff or funding.

    AMD created it to use GCN from their video cards which doesn't exist in Nvidia cards.

    In the same way AMD could adapt Physx to work well with their hardware.

    Lets be realistic here, AMD created it knowing only their cards could run it well. Open source or not the same tricks appear.

    AMD and Nvidia run a business purely for profit. They will do whatever it takes to increase profits.
     
    Last edited: Jul 14, 2017
  13. Unframed

    Unframed Member

    Joined:
    Mar 30, 2010
    Messages:
    9,158
    Location:
    Hella south west
    So they don't make money on hardware?

    They did but being open source means that both companies can add code to implement the tech optimally on their hardware. Nvidia have chosen not to with Tress as AMD have chosen not to with PhysX.

    Thanks for reminding me that they aren't philanthropic. Great contribution.
     
  14. power

    power Member

    Joined:
    Apr 20, 2002
    Messages:
    66,090
    Location:
    brisbane
    after everything that we've explained to you it's clear you are now being deliberately obtuse.
     
  15. bart5986

    bart5986 Member

    Joined:
    Jan 31, 2006
    Messages:
    4,836
    Location:
    Brisbane
    The money made on hardware is used to make future hardware and company profits.

    Why would you pump huge $$$ into making open source software that your competitor will use when you can put that money into getting an edge in hardware?

    Again, AMD and Nvidia only want profits, like everyone else.

    If you were CEO of Nvidia you would send them bankrupt.

    Ok, so we've established that making everything open source is pointless for AMD/Nvidia. There is your proof.

    Lets move on.
     
    Last edited: Jul 14, 2017
  16. Unframed

    Unframed Member

    Joined:
    Mar 30, 2010
    Messages:
    9,158
    Location:
    Hella south west
    No I'm being sarcastic towards unnecessary comments on a line of discussion that runs from a single comment ages ago about a theoretical software approach.

    Aside from the one comment from Snoop explaining that some of Nvidia's tech is open source you guys haven't explained anything.
     
  17. mesaoz

    mesaoz Member

    Joined:
    Jan 15, 2015
    Messages:
    7,809
    Location:
    South East QLD
    You clowns are going to get the thread locked again, read the rules.
     
  18. taldoren

    taldoren Member

    Joined:
    Mar 27, 2005
    Messages:
    360
    How about it has been a lot easier for the last few years for developers to just get nvidia tech working as they actually provide support, vs AMD. I remember one AAA release last year, and the devs replied to a AMA or forum post re-nivdia performance vs AMD and simple said that they asked AMD for support a few times during development and got nothing. Not surprising I guess given that AMD believed that discrete GPU's were dead....

    Only @ Ryzen launch have AMD now started to try and play (again) catchup in regards to providing support for studios etc as per the Ryzen launch AMA.

    End of the day- opensource tech or closed, who cares - studios WILl go with the one that is mos cost effective to implement, e.g Open source is great = but if you have to spend extra x hours or rather, extra dev $ implementing it due to lack of knowledge or support vs paid and done implementation then it's a no brainier.
     
  19. taldoren

    taldoren Member

    Joined:
    Mar 27, 2005
    Messages:
    360
    Do you seriously think AMD is for the consumer ...Lol

    Oh my..that's the best one yet hehe
     
  20. bennyg

    bennyg Member

    Joined:
    Dec 16, 2005
    Messages:
    3,793
    Location:
    Melbourne, Oztraya
    Sorry to break the chain of threadcrap but Gamers Nexus did their Vega FE hybrid mod (sorry would link to YT but it takes like 5 minutes on my phone) and under water its purely power limited, got to a stable 1710Mhz core/1125 mem no downclocking and it was maxing at 400W through the 8pins. It did get a good boost on some things like Doom and at that speed... Watercooled and maxOC... split the 1080 and 1080Ti. But for whatever reason GN didnt put 1080OC results up for comparison.

    I don't think it'll get to 1800mhz stock for RX unless that FE was some bottom bin super leaky silicon.
     
    Last edited: Jul 14, 2017

Share This Page

Advertisement: