OpenGL Successor

Discussion in 'Planetary Annihilation General Discussion' started by blightedmythos, February 6, 2015.

  1. termathor

    termathor New Member

    Messages:
    11
    Likes Received:
    12
    There is really no point in spending a lot of effort rebuilding an game engine with a completely new API, which is totally needed however.

    It's just lost effort.
  2. exterminans

    exterminans Post Master General

    Messages:
    1,881
    Likes Received:
    986
    It's not that easy, really. There are cases where you can actually profit from more recent hardware generations, like Blizzard did in WoW when they swapped out the old low poly models for high poly models, but dumping the entire render pipeline?

    That's basically rewriting it from scratch. It's not just a simple upgrade, they would need to rethink every single decision and optimization they made in the past, every single compromise they made to adapt to the bottlenecks of the old engine.

    But it's not only that... The good thing about Mantle and OpenGL-Next is, that it is much closer to how the hardware actually works, nowadays. The bad thing about that is ... basically nothing is given to you "for free" any more. No structured pipeline, no strictly typed API, nothing. It can give you tremendous performance gains if you manage to write your own engine which entirely discards the old command and data flows, but it also means that you need the knowledge and manpower to develop such an engine of your own.

    OpenGL-Next isn't exactly going to leverage the entry level for newcomers or smaller studios. It's only levering the abstraction level of what is considered part of the fixed graphics API, and what is freely programmable. It is, however, going to promote a whole bunch of new libraries which are going to sit on top of the "low level" API, pampering a limited set of use cases each.

    Until these libraries are in place - and there is one which caters your specific needs -, it's no use to move to the new API as an independent studio. Too much work, too much which can go horribly wrong, and just not enough manpower or money to tackle the new issues on your own.
    doud likes this.
  3. xanoxis

    xanoxis Active Member

    Messages:
    459
    Likes Received:
    238
    Im not saying to rewrite everything but to upgrade what there already is. And I would like to hear how it really looks, from Uber engineer, than speculations.
  4. crizmess

    crizmess Well-Known Member

    Messages:
    434
    Likes Received:
    317
    Wasn't that answered like 4 posts above yours? The timeframe we are talking about here are years.

Share This Page