on current gen cards, copyrect has the same speed as render to texture on flag 256 textures
I assume thats because they emulate fixed function pipeline through shader and the driver is intelligent enough to realize what you do and "modifies" it to be a render to texture internally.
I assume thats because they emulate fixed function pipeline through shader and the driver is intelligent enough to realize what you do and "modifies" it to be a render to texture internally.