Confusion about IJVM

Question

I was reading the chapter 4 of the book 'Computer Architecture' by Tanenbaum (since my copy is not in English maybe the title of the English version is not exactly that one), however it is the chapter about the microarchitecture level. I understood the overall mechanism of the IJVM: one ISA instruction (or Java bytecode) is read from the method area (RAM) and the relative microprogram is launched (which can consist of some datapath cycle).

This is what happens if I program my device with Java, but the mechanism is different if I program it with C for example? I think so since C is compiled while Java is interpreted at the microarchitecture level. But if the mechanism is different (and so the microarchitecture) I can't figure out how my computer can handle a program written both in Java or in C. I mean there must a unique method through which the code setup the datapath control signals.

Probably the answer is trivial but I'm not an expert in this field.

Answer 1

The point to remember is that the JVM is simply another ISA. The java source is compiled to jvm bytecode whereas your C source is usually compiled to the native ISA, you could compile to jvm bytecode as that is simply another ISA.

You could say the native ISA code is also interpreted by the microarchitecture just like the jvm bytecode (jvm bytecode tends to be more involved).

For a given microarchitecture to support a ‘native’ and jvm ISA there needs to be a mechanism to switch between them. Probably the best example is the ARM Jazelle extension. Wikipedia has a rundown of this.