Static Recompilation .. ?!

[tommy87]

@pseudonym
i will ignore your comments about me, if you dont know me than you shouldnt talk like that about me!
And i havent studing computer science and worked on severlal projekts, to have to listen to such sayings.

And now i will explain you, how it works and where are the benefits, because it seems that you dont unterstand them.
Step 1: convert the PS2 code into an assembler code
Step 2: convert the assembler code into a higher language code (like c++) or a pseudo code
Step 3: now you can Manuelly find important places like the handover of the graphic data and backtrack the data, and optimsie the whole process for a PC system.
Step 4: Automate the Optimization for this single Game, that means that you need to built up rules for this, but dont forget: the more oyou optimize, the more difficult it is to automate this process.
Now you have a complete built af game, which is (full) playable.
Step 5: Doing step 1-4 with other games, to expand the automize prozess

Comments:
Step2: the better this step would be done, the easier are the next steps

now im sure you will say that this only works for one game, but if you keep the optimisation as far as possible neutral than it will work with other games too. That means for example that you replace whole code segments to have the same data access for every game.

i know its a long way of doing it this way, but after the first continuous, you have a fully playable game. And now you just have few games, which are nearly playable.

I hope i have nothing forgotten. But please let's bring this conversation to a more friendly level.

[rama]

You must be joking.

[pseudonym]

Since step 2's not feasible for computers and many months or years of a programmer's time if done manually, I can safely stop reading there.

[tommy87]

yes i say it will take a long time doing it. And i know thats it really difficult to transform code into c++, thats why i woukd prefer a simple pseudo code, which just get the meaning of most assembler codes, because manually getteing throught assembler code is a trip throught the hell^^
But i think the advantages would outweigh the disadvantages. In the moment is nearly no game playable at full speed and on most games, i just get 30% and even in FFX i have many times only 70% speed. (on a dual core 2*1,9GHz(real speed)).
If i have the time, i like to start this projekt, i hope i can then count on your help

[Bositman]

In the moment is nearly no game playable at full speed and on most games, i just get 30% and even in FFX i have many times only 70% speed. (on a dual core 2*1,9GHz(real speed)).

Are you serious? Most games are playable, i get ABOVE full speed on hundreds of games, now using a dated dual core processor. I guess your problem is that you need a new PC, not a magical,impractical (and probably impossible) way of making PCSX2 run faster on your slow machine.
You are of course welcome to try and do it, so you can see for yourself it's hopeless

[iakoboss7]

well i dont think after all that way someone would stop his project to start a new one from the scratch! they have worked allot on it to just let it go die at least thats how me as an outsider see all those!

[lightchris]

well i dont think after all that way someone would stop his project to start a new one from the scratch! they have worked allot on it to just let it go die at least thats how me as an outsider see all those!

Not only that, but by the time you would finish such a project as suggested by tommy87 (lets say it would work out), speed wouldn't be an issue anymore. In two years even a mediocre machine will be fast enough for constant fullspeed in most games using PCSX2. Compatability is much more important.

edit: Nice commentary in revision 3000 btw.

[tommy87]

yea maybe you are right its a lot of easier to wait for better PC systems^^
...but it is very interisting to try it...

oh and to say that it wouldnt be a complete new Project, most parts for Step 1 and 2 are aktually done by the PCSX2 team. And imrove the converting into (pseudo) Code can be also a big imrovement for the JITrecompiler.

[rama]

When was the last time you ported a modern PS2 game to PC without having the source code?

[Air]

Split this discussion into its own thread, since it's taken on a life of its own.

(and I might chime in myself soon, if I have time)