The reason why I'd like to experiment two flash memory chip is to improve the number of contemporary players. I've seen that reading 128 samples from flash chip requires about 100 microseconds; in a very simplistic calculation: if we have 10 players the total read time is 1000 microseconds. Since the interval between 2 consecutive update() calls is about 2900 microseconds, and 1000 are spent for reading, than 1900 microsecond are left as computation time for all 10 players: 190 microseconds per player.
With my application, a player requires not less than 180-200 microseconds for processing data; actually, I can use up to 7 players, each reading up to 256 samples:
200 microseconds each read --> 1400 microsecond to perform 7 readings --> 1500 microsecond for whole sound processing --> 214 microseconds per player
If I could access "contemporary" 2 flash chips, I could share readings and should reach a limit of 10 players:
200 microseconds each read --> 1000 microseconds to perform 10 readings (5 readings per flash chip) --> 1900 microsecond for whole sound processing --> 190 microseconds per player