Mill Computing, Inc. Forums The Mill Architecture Pipelining Reply To: Pipelining

Post count: 2

Hello, I have a question about the prologue of the software pipeline. Specifically the example you gave for the retire operation where normally in the prologue there would be nothing there you replaced with operations which would return “none” because they are calculating something before the loop.
What is the point exactly of having those operations which do not calculate anything, would that not waste cache space instead of just having nothing there? I don’t know much at all about compilers and I mainly program in assembly for different architectures and perhaps that is a compiler optimization which I don’t know about.
Secondly you have stated (and as I found out by experimenting a little) the belt is terrible to program with assembly. I know its would be rather silly to optimize for assembly but will there be a tool (Perhaps in the debugger) that allows you to look at the current status of the belt when writing the program, It would make the assembly much nicer to program in.