Hi, microcontroller project). Because I did not find any benchmark results in the archive I would like to ask if someone on this list could give me some numbers. I am especially interested in the duration for the +-*/ operations
I think the best approach would be to implement one of your core algorithms and then look at the generated assembly language. You then just have to count cycles. This has the big advantage of including your control structures which you certainly need in the considerations. The reason I say this is that from what I have seen the pure arithmetic operations are done quite quickly for almost any uc compiler simply because they are assembler written. Control flow is where you really loose speed with bad compilers and imo mspgcc is a very good one. /Roland
