we could change render target allocation so that we always allocate textures a particular sizes (256, 512, n + 256 or some other interval) and then update the render target cache to go off of texture size. Like @chinmaygarde mentioned a while back, the render pass size doesn't have to match the texture size (the render pass can be smaller).

But have a lot of places that assume render pass size == texture size, so there would be some refactoring.

The goal of this change would be to improve cache efficiency when the saveLayer sizes are unstable, like in the example app in #148496