Thanks for the feedback, fellows, but I'm still none the wiser. I'm definitely setting the VRAM flag and I'm running a GeForce 2 GTS with 32MB memory.
I'm sure I've seen demos running that copy much bigger blocks around with no problems on my machine. How about the slowdown when modifying a few (4) UV coordinates on an entity per loop? I can scale textures in-loop with no slowdown, but modify a couple of coordinates and there goes my frame rate! This didn't happen when I tested it outside my game (i.e: without tweening and all the rest). Surely the GFX card takes care of texturing with the given U,V in real time and all Blitz needs to do is pass the modified values across?
I'm sure I've seen demos running that copy much bigger blocks around with no problems on my machine. How about the slowdown when modifying a few (4) UV coordinates on an entity per loop? I can scale textures in-loop with no slowdown, but modify a couple of coordinates and there goes my frame rate! This didn't happen when I tested it outside my game (i.e: without tweening and all the rest). Surely the GFX card takes care of texturing with the given U,V in real time and all Blitz needs to do is pass the modified values across?