1. 运行spine-flutter/Example中的Simple Animation（和原始Demo中的动画资源和骨骼一致），同样采用Flutter DevTools进行查看，主要的性能瓶颈集中在以下几个方面

    1. Spine 动画的渲染和更新：
      SkeletonDrawable.update（占用大量CPU资源，用于更新骨骼动画状态）
      AnimationState.apply（用于应用动画状态）
      SpineFlutterBindings.spine_animation_state_apply（与Spine动画状态绑定的FFI调用开销较大）
      SkeletonDrawable.renderToCanvas
      _SpineRenderObject.paint(主要性能瓶颈，用于渲染Spine动画)
    2. FFI调用：（FFI相关调用都显著占用CPU时间）
      SpineFlutterBindings.spine_animation_state_apply
      SpineFlutterBindings.spine_skeleton_drawable_render

  2. 为了对比FFI在其中的影响，对比运行spine-ios/Example中的Simple Animation，性能上有明显差距

    1. 动画运行期间CPU平均使用率30%-40%（对比flutter的50%-60%）
    2. 同样动画内存占有50M-60M（对比Flutter的250M-300M）
    3. 对比真机Webview 版本 cpu 5% 内存25.4M https://zh.esotericsoftware.com/files/runtimes/spine-ts/examples/canvas/

为什么spine-flutter对比纯ios的内存开销会差别如此之大？

Related Discussions
...

The memory overhead you see does not stem from spine-flutter, or to be more precise, the underlying C++ objects. That is the same as for spine-ios, which also uses spine-cpp under the hood. What you see there is simply whatever Flutter's engine needs to run. Also make sure you are using a release build not a debug build for profiling. I'm pretty sure Flutter in release mode does not use 250-300M.

Regarding CPU usage/performance, Flutter's general purpose FFI can simply not be as efficient as Swift/ObjC code directly calling into C/C++. There is absolutely nothing we can do about on the spine-runtimes side.

The only alternative to this would be to port the Spine core APIs to Dart. I can guarantee you that the result would be worse, Dart's AOT compiler not withstanding.

You can also not compare spine-ts Canvas backend with the Simple Animation example in spine-flutter/spine-ios. spine-ts Canvas can not render meshes, only rectangular image regions.