I thought I've always understood why Java was portable, until I took Computer Organization.
This is my interpretation of a C program from start to finish:
C program --> compiler --> assembly --> machine code --> ISA --> micro architecture (how the computer interprets ISA) --> logic gate --> circuit --> device
Where the compiler must have knowledge of the ISA. Also, the assembly and machine code will vary based on ISA.
Java is as such: (inside JVM): Java program --> compiler --> bytecode
bytecode is the ISA for a JVM.
So, I am guessing the JVM also has it's own micro architecture to interpret it's ISA (bytecode).
Is this accurate?
So the overall cycle would be: (inside JVM) Java program --> bytecode(ISA) --> JVM micro architecture --> host hardware
Also, if the micro architecture is implemented in the processor, does the JVM need to use one of the hosts' processors?
It's more like (following your layering for C):
Note that there are essentially two compilers. That's how Java achieves portability.