I used to use PXA255/270 as embedded computing platform. I did finger prints pattern searching from about ~1000 sampled finger prints library and conducted a bench of calculation like correlation calculation. it usually took 30 sec to get the complete results.
Now I am thinking about BF53X as the computing platform. I checked benchmarks. But they are mixed. it's hard to tell which platform is better.
Looked at BF53X it only supports a few MB async sram ? Is it easy to use nand flash + sdram up to 128MB with BF53X? My library is about 32MB with 1000 sampled finger prints in it. Every time I powered up I need to load my library from NAND flash to sdram. or any better idea?
I guess the BF53X would be much faster? need answer from other people with experience.
I am also thinking about TI's C64XX, they are low cost ics, too.