Index
> Projects and Ideas > fEPU - fasm Emulated Processor Unit |
| Author |
|
) ])
|
revolution 01 Feb 2009, 23:36
Are you currently coding this or are you suggesting someone else code this? Do you have some test code to post?
Which OS has been written in byte code? Will this project also require writing the OS along with the emulator? What is the "instruction set" of the byte code emulator? Has this been defined? |
|||
01 Feb 2009, 23:36 |
|
|
edfed 02 Feb 2009, 00:11
about an emulated processor, fool concept is able to do it because it can be without any node, or only nodes like jne, jge, etc etc...
for example: Code: 00030304300234 44324344300030 23424000043200 these random numbers can become a code if you build a structure over. like a big set of LUTs first byte is the instruction (1/256) second byte can be an operand or anything else etc etc encoding on 35 bytes is possible too. after, you build a set of macros just to write easyly the code. Code: macro if op1,op2,op3,op4 { } if a=b that that: but for a new language edition it is better to build a dedicated editor parser, compiler oriented for the new language. |
|||
|
02 Feb 2009, 00:11 |
|
|
narada 02 Feb 2009, 10:59
2revolution: yes, I have some code, but it is very-very "alpha" for now... (first of all I want to finish os kernel)
An idea becouse it's a big problem to compile OS's for 10 or more platforms (as Linux or UX)... I posted this idea to know - did this problem actual for now? |
|||
|
02 Feb 2009, 10:59 |
|
|
revolution 02 Feb 2009, 11:54
I've never seen another virtual OS before. Java would be the closest I've seen for a platform independent system but that is not an OS.
FWIW: I don't see how this will save you any coding time over the standard C OSes. You'll still need the low level stuff in CPU specific code. Last edited by revolution on 02 Feb 2009, 16:36; edited 1 time in total |
|||
|
02 Feb 2009, 11:54 |
|
|
f0dder 02 Feb 2009, 16:25
I don't really see the point, to be honest.
First of all, we already have java and dotnet. While those aren't perfect, it's still going to take you quite a while to reach their level of performance - and you have to spend time writing standard library stuff as well. And then you have to spend time porting the VM to various systems - which is going to take a while even if you don't want to do it as a full OS but "just" as a usermode application. And when you're done with all this, you have a system that performs worse than native code and only lets you use a subset of CPU features. |
|||
|
02 Feb 2009, 16:25 |
|
|
bitRAKE 02 Feb 2009, 16:51
Might want to look at the L4/x86 micro kernel - it would be ideal to build a VM on top of, imho.
|
|||
|
02 Feb 2009, 16:51 |
|
< Last Thread | Next Thread > |
Forum Rules:
|