Panno - sounds great!
Cygnus - are you using readpixelfast and writepixelfast? It may also be a good idea to read the pixels from a bank and write it to an image, so there won't be 2-way bus-transfer-related speed hits. Also make sure to use the fastest graphics buffer access. I don't remember right now, but some buffers are faster (image vs backbuffer, eventually texturebuffer using flag 256).
Cygnus - are you using readpixelfast and writepixelfast? It may also be a good idea to read the pixels from a bank and write it to an image, so there won't be 2-way bus-transfer-related speed hits. Also make sure to use the fastest graphics buffer access. I don't remember right now, but some buffers are faster (image vs backbuffer, eventually texturebuffer using flag 256).