Indirect jump and STM8S core instruction pipeline

fggnrc · ‎2009-07-27

Posted on July 28, 2009 at 06:43

luca239955_st · ‎2011-05-17

Posted on May 17, 2011 at 15:03

Hello,

this will probably not answer your question, but, generally speaking, the first format has the big advantage (at least from a compiler point of view) of not ''breaking'' X: no need to save and restore it when it contains a meaningful value.

Regards,

Luca

fggnrc · ‎2011-05-17

Posted on May 17, 2011 at 15:03

Hello!

the indirect jump instruction usefulness isn't clear to me.

There are indeed two ways for code an indirect jump.

The first one is:

Code:

JP [Vector.w]

while the other one is:

Code:

LDW X,Vector

JP (X)

According to the STM8 Core Programming Manual, the former instruction is 4 bytes long and it takes 5 clock cyckes to execute.

The latter two instructions take also 4 (= 3 + 1) bytes and take 4 (= 2 + 2) clock cycles to execute.

I therefore see no advantage in using the indirect addressing mode because it can be emulated with index registers...

How the STM8 core instruction pipeline affect my reasoning?

Where can I learn more about it, since there are almost no references in the STM8 Core Programming Manual?

Regards

EtaPhi