Havok on cross-platform PS3/360 development and challenges

[SL369]

A lot of games are cancelled, if your project is risky you would want to spend as little time and money before reaching a state where the projects viability can be determined. In these cases the X360's easier architecture and better tools would make it the obvious candidate for lead platform.

For a small developer 'failure cost' might be a more important factor than 'success gain'.

Many projects, like movie and sport licenses, also have a time limit. It's better to have a finished 360 game badly ported to PS3 than to have an unfinished PS3 game badly ported to the 360.

There are also good reasons to make quick and dirty ports of games. Oblivion is a good port, but the game is yesterdays news and does not seem to be doing much in the charts. Releasing it while it's hot might be better than releasing it when it's good.

 

[Rolf N]

Studio A develops on the PS3 first and then ports over to the 360.
PS3 dev cost = 15million. Port to 360 cost = 1million

Studio B develops on the 360 first and then ports over to the PS3.
360 dev cost = 10million Port cost to PS3 = 1million

You do realize that for that 5M$ difference you can employ 50 programmers for five months, at a very comfortable 20k$ per month?

 

[ShootMyMonkey]

Until the PS3 development cycle time is financially equivalent or better than the 360, most CFO's will have a hard time justifying developing on it for multiplatform games.

Why wouldn't it be? Programming differences hardly account for a whole lot in terms of monetary cost. And the devkits themselves are not that different in cost, or so I'm told. That only really leaves art assets, which makes up the biggest part of your development costs by far. The two consoles are hardly that radically different in terms of rendering capacity and what kind of content you can push through the pipe when you start comparing things in practice (and not on artificial benchmarking).

While the idea of starting with PS3 as the principal platform is probably the better choice in the long run, you won't really see a good example of this for a while because most multiplatform games right now were likely beginning preproduction and development when the 360 was the only next-gen console out there, so there wasn't really a choice.

Either way, most people assign the word "port" (often times in the pejorative sense) to any game that was developed for multiple platforms, which is rarely ever accurate, since it's far more likely (depending on the time frame) that it was developed in parallel on all platforms simultaneously. And simultaneous multi-platform development is not super-costly in comparison to developing on one platform -- though the case of a platform coming into existence later than others throws a wrinkle into the figures, since it often means not having the staff worrying about one platform's functionality until much later and realizing after the fact that you made a load of mistakes -- since the biggest challenge is not likely to be one for content creators.

 

[RobertR1]

Why wouldn't it be? Programming differences hardly account for a whole lot in terms of monetary cost. And the devkits themselves are not that different in cost, or so I'm told. That only really leaves art assets, which makes up the biggest part of your development costs by far. The two consoles are hardly that radically different in terms of rendering capacity and what kind of content you can push through the pipe when you start comparing things in practice (and not on artificial benchmarking).

I don't know, I'm not a developer. Just going by the echo's of the 360 having better dev tools and the Cell being difficult to program for. This, I equate to more time being spent on development to reach the same level and thus more money. Basically, take two mechanics, hand one a modern computer diagnostics machine and the other a generic laptop with somewhat crude diagnostics software and see which one can routinely find the problem on modern cars first. Both will get the job done but one might be a bit more efficient at it than the other. This efficiency translates to dollars saved. Seeing that you often hear about how expensive development is for these next gen consoles, keeping tabs on such figures would be even more critical.

While the idea of starting with PS3 as the principal platform is probably the better choice in the long run, you won't really see a good example of this for a while because most multiplatform games right now were likely beginning preproduction and development when the 360 was the only next-gen console out there, so there wasn't really a choice.

Either way, most people assign the word "port" (often times in the pejorative sense) to any game that was developed for multiple platforms, which is rarely ever accurate, since it's far more likely (depending on the time frame) that it was developed in parallel on all platforms simultaneously. And simultaneous multi-platform development is not super-costly in comparison to developing on one platform -- though the case of a platform coming into existence later than others throws a wrinkle into the figures, since it often means not having the staff worrying about one platform's functionality until much later and realizing after the fact that you made a load of mistakes -- since the biggest challenge is not likely to be one for content creators.

If a "port" isn't a port but a parallel dev platform, then one system shouldn't be blamed for others shortcomings right? So really, the PS3 ports not performing upto the level of the 360 so far is simply due to the PS3 being late and lack of programming familiarity, not the 360, which many people love to throw the blame on.

 

[Oninotsume]

I'll have to agree with Shoot My Monkey in that, at least as far as I have witnessed, the general trend towards developing on the 360 s a result of the 360 being available on the market 1 year in advance of the PS3.

However, I think it's usually beneficial for companies to release products cross-platform, so developing at least with the PS3 in mind from the outset is preferrable to cramming it in there at the last second for everyone's sake.

I can't help but wonder how shoddy ports will fair when 1st party/good ports begin to set a new standard of quality on the PS3. Perhaps more consumers, knowing what to expect in terms of quality, will be less likely to accept poor ports as time goes on.

 

[Bigus Dickus]

The choice devs face is either 'produce diabolical ports on PS3 at a saving' or 'design for PS3 at a cost but produce impressive results', and it's a choice that's technical in nature, not economic.

One, I think technical = economic in the larger picture. Two, if what you say is really the case then I have to wonder if Sony went the right direction with the PS3.

 

[ShootMyMonkey]

This, I equate to more time being spent on development to reach the same level and thus more money. Basically, take two mechanics, hand one a modern computer diagnostics machine and the other a generic laptop with somewhat crude diagnostics software and see which one can routinely find the problem on modern cars first. Both will get the job done but one might be a bit more efficient at it than the other. This efficiency translates to dollars saved. Seeing that you often hear about how expensive development is for these next gen consoles, keeping tabs on such figures would be even more critical.

The only real problem with that analogy is the fact that it's on the scale of individuals. You have to think of it more in terms of the cost and productivity of an entire shop full of mechanics rather than one mechanic against another. Say you have two shops each with 20 mechanics. In each of them, 15 or so mechanics are typically doing more basic stuff like swapping fluids, changing tires, replacing brakes, doing alignments -- comparatively simple stuff, but it accounts for the vast majority of your business. The other 5 or so are the ones doing complex diagnostics and finding more esoteric issues. Now if one shop has better diagnostic equipment than the other they might be able to give them a 20% productivity gain than the other for those tasks. But those tasks occupy the workload of only a quarter of their mechanics, so that 20% advantage on a quarter of their staff only becomes a 5% advantage overall -- Now I'm assuming that the staff is apportioned this way because that's about the fraction of people who bring in their car for something major. In a game studio, it's probably more extreme, though not because of staff proportions. Content occupies so much more of the end product, and the biggest thing that people on the tools and technologies end of a particular studio have to worry about is making things accessible and easy for content creators.

If a "port" isn't a port but a parallel dev platform, then one system shouldn't be blamed for others shortcomings right? So really, the PS3 ports not performing upto the level of the 360 so far is simply due to the PS3 being late and lack of programming familiarity, not the 360, which many people love to throw the blame on.

Yep, that's about the size of it. The PS3 is the new kid on the block and it's just a sudden shift. That said, a lot of things that work out nicely on the PS3 work out nicely on the 360. It's just that the PS3 is less forgiving than the 360. A good implementation runs like stink, and a bad implementation just plain stinks.

I should say that the things that do qualify as "ports" almost all the time, for example, might be something like a title that exists on 360 as well as PSP since the content and code differences are too disparate to have the two versions to be developed by the same team. 360 vs. PS3 at this point now that both are available probably wouldn't fit the bill excluding those games that may come out on PS3 that were already out on 360 before the PS3 existed (and I doubt there will be many of these [if any]). There were a few examples of games early on in the 360s life where there was parallel development across generations (i.e. Xbox1+Xbox360 or something), and that's a perfect example where a port might actually have been better because the older machine being the weaker machine means you have to work around its weaknesses and any case where you play to the strengths of the new hardware would be non-intrusive matters that don't account for too big of a difference.

However, I think it's usually beneficial for companies to release products cross-platform, so developing at least with the PS3 in mind from the outset is preferrable to cramming it in there at the last second for everyone's sake.

That would be nice... if only it actually worked out that way right now.

 

[chachi]

A: You’ll have a hard time if you port without having a PS 3 game in mind when you created the 360 version. That is where a lot of complaints are coming from. They created the 360 engine with a unified memory architecture in mind, with the embedded frame buffer with its advantages and disadvantages, and not thinking too much in early stages about multicore. If you try to get that over to the PS 3, you’re in for a bad surprise. The PS 3 is all about streamlining [and] about the two different memory pools. They are separate. You don’t have to do tiling because you don’t have an embedded frame buffer. All of these advantages of the PS 3 turn into disadvantages if you don’t start making your game on the PS 3. Hence the griping. If you create first on the PS 3, it is pretty easy to port it to the 360. A lot of companies coming on board now will probably start on the PS 3 and move to the 360. The lucky thing for us is we didn’t have to think about the 360 at all.

The problem I have with this is he isn't saying that starting with the PS3 and porting to the 360 will result in a better game. The PS3 having a problem with framerates isn't because it started life as a 360 game, it's because the PS3 apparently has a hard time pushing the same amount of polygons as the 360 does (and it's not like the 360 is a polygon crushing monster). If the game started off on the PS3 and they scaled back models so the PS3 didn't have a problem maintaining the framerate, is that better for gamers? I don't want to see them cut art assets any more than I would want them to scale back the physics across the board because the 360 can't handle as many calculations as the PS3.

I think for multiplatform Capcom has it right, with the bonus of actually getting three platforms for little more than the price of two.

 

[Fafalada]

If the game started off on the PS3 and they scaled back models so the PS3 didn't have a problem maintaining the framerate, is that better for gamers?

If game performance could be narrowed down to something as trivial and specific as "model complexity", framerate problems wouldn't exist in the first place.

 

[Arwin]

The problem I have with this is he isn't saying that starting with the PS3 and porting to the 360 will result in a better game. The PS3 having a problem with framerates isn't because it started life as a 360 game, it's because the PS3 apparently has a hard time pushing the same amount of polygons as the 360 does (and it's not like the 360 is a polygon crushing monster). If the game started off on the PS3 and they scaled back models so the PS3 didn't have a problem maintaining the framerate, is that better for gamers? I don't want to see them cut art assets any more than I would want them to scale back the physics across the board because the 360 can't handle as many calculations as the PS3.

I think for multiplatform Capcom has it right, with the bonus of actually getting three platforms for little more than the price of two.

But everyone is pointing to the unified memory and lack of streaming in many different areas of current (often PC based) game engines as the main culprits. So it's not about the polygons.

I actually think that if games are more designed with streaming in mind, the 360 will benefit as well.

 

[Panajev2001a]

If game performance could be narrowed down to something as trivial and specific as "model complexity", framerate problems wouldn't exist in the first place.

You underestimate me and my ability to create bottlenecks!

 

[pipo]

You underestimate me and my ability to create bottlenecks!

 

[AlNom]

I actually think that if games are more designed with streaming in mind, the 360 will benefit as well.

But... but... isn't that what the Core unit is all about? "Secretly" forcing devs to implement streaming Conspiracy!