[Scummvm-devel] AGOS slowness

Kostas Nakos knakos at phys.uoa.gr
Wed Jul 16 14:04:03 CEST 2008


On 07/16/2008 12:50 PM, Bertrand Augereau wrote:
> You will notice that the huge winner is Normal2X, just ignore it, it is 
> only VRAM-write bound (verified by outputting only 0) on my machine and 
> not representative at all of the CPU load on other devices.
> 
> Maybe the blitting functions need a few optimizations, but maybe there 
> is something else involved.

This *could* mean that the scaler proc is called too often (indicating 
unnecessary repaints) and/or the updates handle a lot of data (e.g. 
screenfulls).

K.




More information about the Scummvm-devel mailing list