Hi all
I have a general question for a design of DSP at floating point.
I use a DSP for embeded system.
For my application I know approximatively the type and the number of operations.
multiplier : 2000ops per 100µs
addition : 2000ops per 100µs
Cosinus and Sinus : 200ops per 100µs
exp sqrt ... : 300ops per 100µs
and test (if then else.. ) : 2000ops per 100µs
So with a big margin I use a DSP > 500 MIPS approximatively?
This value is correct ?
And now I need to design a size of memory
For the store application I need ( 4000 word x 32bits) = 64kBytes + for application ( 32000 word x 32bits ) 320Kbytes
So with a 512kBytes it's ok ...
Now with these criteria
What is the better DSP or microcontroller can be use for this project ?
Thank you for advance ..
Olive