Bone的UpdateWorldTransform方法消耗巨大,有没有可能用Burst加速?另外spine的每帧计算结果都是固定的是否考虑缓存计算结果?

  • Harald hat auf diesen Beitrag geantwortet.
    Related Discussions
    ...

    linky Bone的UpdateWorldTransform方法消耗巨大,有没有可能用Burst加速?

    Parallelization is covered in this issue ticket. We won't be using burst due to the imposed constraints not going well with the basic runtime architecture. Nevertheless, basically all parts of Update and LateUpdate which can be parallelized are off-loaded to worker threads. Most of the work has been done already, however some cleanup remains until we will officially release the feature.

    linky 另外spine的每帧计算结果都是固定的是否考虑缓存计算结果?

    What do you mean exactly by the first of these two sentences? Machine translation was unfortunately not very good.

    • linky hat auf diesen Beitrag geantwortet.

      Harald 谢谢回复。不知道正式发布有没有一个大概的时间。This mean “cache calculation results”

      • Harald hat auf diesen Beitrag geantwortet.

        linky This mean “cache calculation results”

        Thanks for traying to explain that again. I'm not sure what calculation results you want to cache, as there is nothing that remains constant when animating. What does your animation look like that you think parts of the results could be cached?

        • linky hat auf diesen Beitrag geantwortet.

          Harald 一个动画在任意一帧所有的骨骼数据应该都是确定的,如果这个数据能缓存下来,对一场景里有许多同个spine的情况性能会提升很多。相比这个问题我更想知道并行化工作有没有预计的发布时间

          • Harald hat auf diesen Beitrag geantwortet.

            @linky I don't quite understand your question. All data changes each frame during animation. No computation can really be cached. What exactly do you think remains constant and should be cached?

            linky 相比这个问题我更想知道并行化工作有没有预计的发布时间

            We can't provide dates on this one, sorry, it's released when it's done and fulfils our quality criteria.