Hybrid Renderer vs VFX Graph

I have some code which uses compute shaders to compile position, normal direction and color data into render textures which then get read by the vfx graph to render. I'm looking to build off of the Unity boids example to render as many objects on screen as possible. My question is, is it worth it for me to take out the ecs rendering in favour of vfx rendering? I don't need any interaction with the physics APIs at all it's purely visual.

The goal is to get as much on screen as possible at 60fps regardless of hardware limitations. 100% efficient CPU usage for simulation and 100% efficient GPU usage for rendering.

I'm new to both ECS and the vfx graph so any input regarding challenges I might be getting myself into or things I'm not considering would be greatly appreciated.