Backwards compatibility on Xbox 360

[ban25]

IIRC it was the Ceramic White PS2. Which perplexes everyone as it's, in theory, the same damned console with a different colour plastic!

BTW I'm buying a black PS3 to help aid BC

Also, in the case of Tekken 5, I believe the incompatibility is restricted to the old emulated games on the disc, not T5 itself.

 

[see colon]

about the white/"ceramic"/satin silver PS2 and incompatabilities... sony sells those same PS2s with the same incompatabilities in black shels as well. they are just 75000's (the only PS2s currently in production) and not only do they lack compatability with some games, they also lack the built in modem of the 70000's (they've only got a network jack). while that's not such a big deal for most people, it's annoying for a few. i know a few people who have them and have run into problems with a few games. tekken 5 is a big one, because the classic arcade games don't work. tekken 3 is on many of my friend top ten lists.

 

[Fafalada]

Jthe KK press conference specifically mentions that games must conform to the PS2 TRC

TRC = Technical Requirement Check. Failure to comply with TRC is automatic fail to pass the Q&A check - if you're violating the TRC it damn better be in a way that is not possible to detect, or your game will never see print until you fix it.

Which is to say that the games need to have been coded 'by the book' using the SDK.

No, it means nothing of the sort.

In the podcast they seemed to feel that some of the better games may be doing something that dont conform to the TRC in order to squeeze out more performance out of the PS2.

Like others said, I think that's more their 'carte blanche' for any odd title that might experience issues - the list of those on PS2 was incredibly short(IIRC only one of them actually failed to work) but especially press still tried to make a big deal out of it.
Anyway for a software emulator, there's quite a few potential issues I can think of that wouldn't violate the TRC to my knowledge.

 

[expletive]

It certainly seems to me that you were trying to bring down PS3's potential BC capability down to Xbox level and hence this whole PS3 vs Xbox BC discussion.

In the end, since Sony has already announced some sort of hardware support in their implementation, plus based on their experience in implementing BC in this field (hardware + software amalgamation with less than 1% incompatibility in their last implementation), it is only reasonable to expect that they will surpass (understatement ? ) X360 "Backwards compatibility".

I wasnt trying to 'bring down' anything, and honestly its getting old that when anyone has anything potentially negative to say about the PS3 the ninjas are released. My original post was just a rehash of a conversation i heard on a podcast (and i was clear about that), it wasnt related to the 360 at all, just the PS3. I posted it to possibly shed some light on PS3 BC and what it may actually be thats all. Go read my initial post on the subject.

TRC = Technical Requirement Check. Failure to comply with TRC is automatic fail to pass the Q&A check - if you're violating the TRC it damn better be in a way that is not possible to detect, or your game will never see print until you fix it.

I'm just rehashing what was said, and IIRC the number was around 20-30-% of games for the PS2 dont conform to the TRC according to developers they spoke to. If they all HAD to confirm, i dont see why its worth mentioning in the slide at all...

Titianio, what was the non-negligible number referring to?

 

[Fafalada]

I'm just rehashing what was said, and IIRC the number was around 20-30-% of games for the PS2 dont conform to the TRC according to developers they spoke to.

I'd like to see a source for this, while it's possible for some TRC violations to slip under Q&A radar undetected, this number is absurdly high.

If they all HAD to confirm, i dont see why its worth mentioning in the slide at all...

It's called "requirement list" for a reason - it's not "recommendation list", you're required to follow it, and you Will fail Q&A check if violations are found during the process.
Obviously SCE Q&A is not infallible - so there's possibility of TRC violations slipping by, but they are very unlikely to be intentional (developers aren't idiots - you want to do all you can to conform with the TRC, risking delaying title on the hunch that you won't be discovered is just plain stupid).

 

[Phil]

The PS2 has a chip in it to emulate the PS1 IIRC, it seems like this will be an emulator, which is what the 360 uses as well. To me, thats plenty of reason to believe it wont be as good as ps1 emulation on the PS2.

I'm pretty sure I already corrected you on this already, so again, PS1 compatibility is not a 100% hardware solution on PS2. At this point, we really don't know if PS3s backwards-compatibility will be any more software based or not.

 

[expletive]

I'm pretty sure I already corrected you on this already, so again, PS1 compatibility is not a 100% hardware solution on PS2. At this point, we really don't know if PS3s backwards-compatibility will be any more software based or not.

I dont think you did but Titanio has and i've since change my stance based on that info in later posts.

I'd like to see a source for this, while it's possible for some TRC violations to slip under Q&A radar undetected, this number is absurdly high.

I agree and the impression i got from the podcast was that Sony allows certain games to pass QA without conforming strictly to the TRC becuase from a testing standpoint they are bug-free. However, the code needs to talk directly to the hardware in ways that boost performance and also compromise the TRC. I wish i could remember which one (podcast) it was, i subscribe to about 10 and this was some weeks ago now so tracking it down would just be too time consuming. I was hoping that mentioning what i remembered would trigger some developers on the forum to confirm or summarily reject the entire notion. I would say your comments do the latter but the TRC slide is the one thing that doesnt add up in the equation, to me at least.

 

[one]

I'm just rehashing what was said, and IIRC the number was around 20-30-% of games for the PS2 dont conform to the TRC according to developers they spoke to.

So who are 'they'?

 

[Qroach]

Sony will waive TRC issues for the larger publishers with more clout, and also internally developed titles. Companies like Konami, square, EA, capcom, etc....

Also TRC's are different from region to region.

 

[NavNucST3]

Well, I'm not sure if it started with Major Nelson or not, but apparently in show #167 he starts to talk about TRC @ around the 4:00 min mark.
The number he gives is 50%-80%, then there are a couple websites that state 50%-85% edit for clarity: those are the numbers that PASS, obviously unofficially, and the only proof offered is anecdotal "...according to the developers I talked to..."
wikipedia
PS3Portal

DONT KILL THE MESSENGER, none of this came out of my mouth.

 

[patsu]

I don't want to comment on Major Nelson since he is obviously not a credible source and the survey he mentioned is "Ask a few devs". We will probably see a confirmation of Sony's B/C effectiveness based on their "incompatible games" list in November. I would be disappointed if it's 50%.

What is the difference when Ken said Sony will do B/C from the get-go vs Microsoft's B/C approach ? What is Xbox 360's B/C % today ? If an Xbox game passed MS's TCR, will it automatically be Xbox 360 friendly ?

 

[Legend]

as for PS3's b/c precentage, I would expect something like 95% if not more since PS2 did it so well. some may argue that we are dealing with different hardwares here, but still, b/c was not a moment's thought for Sony. they had it since the beginning (I presume).

for what its worth, I remember the list of PS1 games (or was it precentage?) that do not work on PS2 and the number was very small.

 

[ERP]

TRC = Technical Requirement Check. Failure to comply with TRC is automatic fail to pass the Q&A check - if you're violating the TRC it damn better be in a way that is not possible to detect, or your game will never see print until you fix it.

Not strictly true, I know many products that have shipped eith upwards of 20 violations, it's all negotiable to some extent.
But most TRC's aren't going to impact backward compatability in any way.

 

[DemoCoder]

Someone once described a trick which to me will seem hard to emulate efficiently on the RSX. They claimed they would forcefully flush the texture cache mid-render and use the results during the render, or something like that, effectively processing a render target without waiting for the full frame to finish. I guess with enough rendering power, you could swallow an inefficient implementation.

 

[Fafalada]

Someone once described a trick which to me will seem hard to emulate efficiently on the RSX. They claimed they would forcefully flush the texture cache mid-render and use the results during the render, or something like that, effectively processing a render target without waiting for the full frame to finish.

Why would anyone need to wait for a frame to finish to process various render targets? We have a dozen targets floating around midframe, sometimes moving the addresses of frontbuffer from frame to frame as well.
Anyway, flushing texture page buffers(equivalent of texture cache on PC chips) frequently, is a common operation on all PS2 games, and afaik a literate emulation of that behaviour would have performance problems for any modern PC chip.
But if you want to talk really difficult examples - some games(well at least one I know of) render stencil data by setting texture and frame buffers to the same address, and then flushing texture page buffer on every rendered primitive to ensure coherency between reads and writes (ask nAo if you want more details).
It's kinda nasty, but on GS the performance is almost unaffected by things like that.

And that's just one example - there are other GS 'friendly' things that directly translate badly to other GPUs (eg. our renderer typically switches rendertargets a couple of hundred times per frame).

Not strictly true, I know many products that have shipped eith upwards of 20 violations, it's all negotiable to some extent. But most TRC's aren't going to impact backward compatability in any way.

Indeed I was only referring to TRCs that could impact compatibility in some way. The other stuff is 90% legal mumbo jumbo that has absolutely no effect on emulation.

However, the code needs to talk directly to the hardware in ways that boost performance and also compromise the TRC.

IMO that's a load of BS.
I listed two examples to democoder above that "talk directly to the hardware in ways that boost performance" and neither of them violates anything in TRC. In fact I would be very hard pressed to find things in PS2 TRC that restrict me from doing funny things to boost performance.