But you're right Sony developers should know more then some 'amateur'homebrew developers.. But then again, see what 'amateur'homebrew developers can wring out of old hardware which no-one deemed possible (like the Philips Odyssee or C64)..
Homebrew coders are definitely amateurs compared to real professional developers - ESPECIALLY the people in this scene. Trying to deny that is being in denial. There's nothing a homebrew coder can do that a seasoned industry professional couldn't. If they can, they get bought out by companies.
That's why you will find that most of the really good 'emulators' written are often done by people who either
a) work for the big-name companies directly but keep themselves anonymous (like the Quake2x port for Xbox 1, like pSX done by PSXAuthor).
b) eventually get to work for those companies (like Sardu, like Mooglyguy, like zsKnight, and the list goes on, on, and on).
3a. You mean YOU aren't going to write PPU/SPU code because of the limited userbase
For the record, the 'PPU' is just like every other bogstandard processor - there is nothing special about writing 'PPU' code - it's the main CPU and any application you write for PS3 is 'PPU code'.
Second, SPUs are best thought of as specialized vector processors - it all sounds cool that you have 6 SPUs to work with, until you find out you can't really use them for scalar code - until you find out that every time you use them, there's IPC overhead to deal with.
You think you can just wave a magic wand, sprinkle some 'SPU code' around (whatever you think that means) and everything will just run all fine and suddenly slow code magically transforms into fast code? You need to actually have a PURPOSE in your code to be able to offload stuff to SPUs - otherwise you're just going to shoot yourself in the foot and end up with an even slower program.
Originally Posted by stevorkz
Dont even waste your time with this guy. It seems whenever he posts hes running someone down letting them know just how stupid they are if they not up to his "brand" of thinking.
Unfortunately that's where we are at on this forum yeah - I don't blame you guys for being end-users, but I do blame you guys for not having the willingness to educate yourself.
I repeat - thinking that somebody is going to make a PCSX2 port for PS3 that runs at fullspeed is going full retard - period.
The chances someone will even try to port the code is slim, but who are you to tell others not to ask or discuss this..
Who am I? I'm a guy who has dedicated more than two years of his time to keep this 'homebrew scene' from totally dying out - that's who.
And I'm telling YOU - and your kind - to stop 'hoping' that somebody is going to do this - because that day is never going to happen - and when it does, it will probably run like **** and be abandoned within a few months.
So stop deluding yourself and stop being 'unrealistic' as to what can be achieved.
PlayStation 3's Cell CPU achieves a theoretical maximum of 100 GFLOPS in double precision calculations.
six-core PC processor reaches 109 GFLOPS (Intel Core i7 980 XE) in double precision calculations.
I asked mudlord if I should even dignify this crap with a response - he told me not to - but I guess I'm stupid like that - I will try to debate and get you people to become realistic in some way even though it's a waste of time - other people will just call you 'idiotic end-users' behind your back and leave it at that.
If you people want to 'debate' this **** with me, here is what you do -
Get off your ass, stop being a lazy 'know-it-all' end-user, stop playing 'backup games', and start PROGRAMMING for PS3 - that's right, PROGRAM for a change. Start with some ports, see for yourself how well your precious PS3 runs code. Then find out that there's no easy way to run dynarecs at this stage for CFW PS3s - then find out that an LHS costs you 500/600 cycles - then find out that your 3GHz CPU with 'a theoretical 100 GFLOPS' runs more like a Pentium 4 2.4GHz CPU (for the record - this is a CPU dating back to 2001/2002).
It's not about how many double/float calculations can be done per second, it's about how fast your Cell CPU can push all those instructions through its deep, deep pipelines. It's about how long your Cell CPU takes to start resuming work after it trips over an LHS. And to give you a short answer to all that - the Cell CPU runs like absolute crap and ends up barely outperforming a Pentium 4 because of that.
Good luck BTW trying to find yourself a run-time profiler to be able to produce 'optimized code' well enough so PCSX2 will run at fullspeed on PS3 - note - you will have to go buy yourself a DECR to be able to use something like SN Tuner. Thats the thing Math just bought a few months ago - that will set you back a good few thousand dollars at least - and it's illegal to buy for an end-user or for a developer to sell just so people know.