What can streaming FPU processors achieve - example here.

[Oda]

The PS2 (and soon to be Cell) and its new style of programming seperated the men from the boys in the development community, and I couldn't be happier because of that. For example:

In early to mid 2000 when forums and game web sites were proclaiming doom as devs were having a hard time with the PS2, I recall an interview with Hideo Kojima. He stated that while his team did have some initial difficulty with the system, they were beginning to get the hang of it, and he applauded Sony for persuing what they believe is the future of game development (it should also be said that he agrees with Sony's approach).

On the other hand, Lorne Lanning of Oddworld Inhabitants wouldn't shut the hell up about how much his team disliked deving on the PS2. He never talked about potentials or weaknesses Sony's approach brought, he merely bitched about how it was different, and how his team couldn't/didn't want to adapt.

Where are these guys now? Kojima now has his own studio within Konami, and MGS2 and MGS3 are some of the most graphically impressive and best games on the PS2.

Lorne Lanning and his team have withdrawn from game development after putting out only two titles on the Xbox, both decidedly mediocre.

I'm with gurgi; let punks like Lanning find new jobs, we've no use for them in this industry.

 

[Frank]

I don't like the way people dismiss ease of development so flippedly. How powerful your platform is is meaningless when you can't tap into that power. If a platform has a easier environment to tap the performance even if that platform is less powerful on paper than the competing platform then you're going to get better games on the platform that's easier to develop on. Untapped power or power so elusive that it requires a prohibitive development cost is useless.

Yes, Microsoft and all developers wanted a single OOO CPU core, with lots of cache, running at 10 GHz. So they can just go on doing what they do. But that isn't going to happen.

 

[jvd]

Oda , I don't know why you think making things harder up front is a good thing .

Why can't they design a chip that is both easy up front and very powerfull and multi threaded ? That is the key. That is what should be worked towards . Not making things more complicated

The ps2 was lucky that it had the biggest installed base so that developers were forced to squeeze out everything .

If the ps2 didn't enjoy the retail success that it did , it wouldn't have enjoyed the programing sucess .

 

[Shifty Geezer]

Why can't they design a chip that is both easy up front and very powerfull and multi threaded ? That is the key. That is what should be worked towards . Not making things more complicated

I'm sure if they could design such a chip, they would.

 

[MechanizedDeath]

Oda , I don't know why you think making things harder up front is a good thing .

Why can't they design a chip that is both easy up front and very powerfull and multi threaded ? That is the key. That is what should be worked towards . Not making things more complicated

The ps2 was lucky that it had the biggest installed base so that developers were forced to squeeze out everything .

If the ps2 didn't enjoy the retail success that it did , it wouldn't have enjoyed the programing sucess .

You can't always get what you want. Compromises must be made. SMP is a compromise due to the power ceiling. When you hit the point of diminishing returns, you can either keep the ever-smaller returns, and fall behind the game, or you find new ways of getting power.

Anandtech is a PC-centric site. I have no doubt that most (if not all) the devs that complained to him are lazy PC devs. The ones who did quicky ports of their PC games to the consoles last gen, and produce forgettable games that sold forgettable amounts of units.

The console industry is used to jumping to new hardware every 5 years. They are used to reinventing the wheel twice a decade, if you will. Will it be easy this time? No. But it wasn't last gen either. To the end-user, ease of development means nothing. I don't want devs complaining, I want devs doing. If they complain, maybe they're just not good enough at their job. Times are changing. Devs either take the heat and get up to speed, or they get the hell out of the kitchen.

BTW, what comes first, the support or the sales success? PS1 earned sales success before completely nailing down the support. The DC got support early, but no retail success. The Xbox had support, but little retail success. I don't know if it's as simple as that. PS3 is guaranteed support again b/c Sony is the market leader. So that support will help feed its success again with the PS3, so it probably doesn't matter how easy the PS3 is to dev. Devs will go where the audience is, regardless of ease of development. Costs would have t be seriously prohibitive to keep devs away. And there's no suggestion that's the case with the PS3. Besides which, 360 dev isn't exactly gonna be a walk in the park either. PEACE.

 

[jvd]

You can't always get what you want. Compromises must be made. SMP is a compromise due to the power ceiling. When you hit the point of diminishing returns, you can either keep the ever-smaller returns, and fall behind the game, or you find new ways of getting power.

OF course you can .

Anandtech is a PC-centric site. I have no doubt that most (if not all) the devs that complained to him are lazy PC devs. The ones who did quicky ports of their PC games to the consoles last gen, and produce forgettable games that sold forgettable amounts of units.

Lazy ? I doubt it . I don't see why you need to fight hardware to be considered not lazy .

Will it be easy this time? No. But it wasn't last gen either

sure it was , on the xbox and gamecube

PS3 is guaranteed support again b/c Sony is the market leader.

ps3 is not the market leader though. There is no market leader . If anything for the first year the x360 will be the market leader and if can keep the lead going into the second or third year you may see alot of dev support on the x360 and see the cell getting the shaft .

As I said it was its retail success that got the devs to push the system . If the retail success wasn't there the devs woulnd't be there and if the market was split between two systems it wouldn't have seen that lvl of support.

 

[seismologist]

The PS2 (and soon to be Cell) and its new style of programming seperated the men from the boys in the development community, and I couldn't be happier because of that. For example:

There's always certain elite developers who will be able to tap into the power. That's not always a good thing for gamers though.

Have you seen the PS2 version of Psychonaughts? Or even Half Life 1 for that matter Most PS2 ports turn out to be a mess unless the game is written from the ground up.

 

[Titanio]

Have you seen the PS2 version of Psychonaughts? Or even Half Life 1 for that matter Most PS2 ports turn out to be a mess unless the game is written from the ground up.

The thing is, the lowest common denominator on PS3, visually at least, is going to be a lot higher than it was on PS2, relatively speaking. RSX alone almost guarantees that. To get to a certain standard on PS3 will be easier than it was on PS2, but there's still a lot of "exotic" headroom there for the more ambitious.

 

[Frank]

A really good way to speed things up would be to use SRAM as the main memory, with a very broad (1024 bits or more) and fast bus. It would probably have to be on the same package as the CPU. That would be like main memory with a better speed than level-2 cache RAM. Who cares about OOO or cache misses when you have that?

But that would be extremely expensive. And that is why Sony uses a bit of SRAM as local memory, and why Microsoft doesn't want a very large level-2 cache.

 

[Fox5]

Have you seen the PS2 version of Psychonaughts? Or even Half Life 1 for that matter Most PS2 ports turn out to be a mess unless the game is written from the ground up.

The thing is, the lowest common denominator on PS3, visually at least, is going to be a lot higher than it was on PS2, relatively speaking. RSX alone almost guarantees that. To get to a certain standard on PS3 will be easier than it was on PS2, but there's still a lot of "exotic" headroom there for the more ambitious.

Wasn't it the graphics chip(s) on the ps2 that were hard to make use of and not the cpu?

 

[Titanio]

Wasn't it the graphics chip(s) on the ps2 that were hard to make use of and not the cpu?

Arguably both, I think, and the line between CPU and GPU on PS2 was a little more blurred. I'm not sure if one was more difficult than the other, someone with PS2 dev experience may be able to answer your question more fully.

 

[Fox5]

Wasn't it the graphics chip(s) on the ps2 that were hard to make use of and not the cpu?

Arguably both, I think, and the line between CPU and GPU on PS2 was a little more blurred. I'm not sure if one was more difficult than the other, someone with PS2 dev experience may be able to answer your question more fully.

Well, I believe the memory had horrible latency as well, but I don't think there was anything about the ps2's cpu that made it difficult to program for like cell, wasn't it pretty much a general purpose cpu, possibly inorder, but didn't require multi threading.

 

[Titanio]

Well, I believe the memory had horrible latency as well, but I don't think there was anything about the ps2's cpu that made it difficult to program for like cell, wasn't it pretty much a general purpose cpu, possibly inorder, but didn't require multi threading.

It's main power was in VU0 and VU1 which required assembly programming, and from what I've heard were generally tricky buggers to work with. Many games just didn't use them at all. I've heard the mips core wasn't exactly setting the world alight either.

Cell brings its own challenges. Most particularly the general scale of parallelism and concurrency that is required, or might be required, at least to extract best performance. But I think early Cell coding should be easier than early EE programming was.

 

[DemoCoder]

You'd use the z-buffer for sorting, not triangle setup. On a G70, you can achieve 10billion compares/s, or 20 billion with double pump.
The challenge is to partition your sort properly and structure the data layout to best take advantage of streaming and involving the CPU as little as possible. It will only really scale well with very large datasets (billions/trillions of elements to be sorted). I sucessfully implemented dynamic programming substring search algorithms on the GPU.

 

[Megadrive1988]

"the SPE array ends up being fairly useless in the majority of situations, making it little more than a waste of die space"

The article's gone now, so let's not debate it any more. Just pointing out that anyone thinking GP is essential, there's evidence to the contrary (though as you say, devs need to get a handle on it. Though I'm not happy to think of PS4 out in 3 years!!!)

hehe, you have nothing to worry about. PS4 probably won't be out until 7 years from now. 6 years after PS3 comes out

 

[Oda]

Lazy ? I doubt it . I don't see why you need to fight hardware to be considered not lazy.

It's not about 'fighting' the hardware, it's about adapting to its changes; changes that are made so that games can get better and better. The age of bumping up the clock speed to solve all our problems is gone. The devs that realize this are making the games I want to play. The devs that don't will quickly be following Lorne Lanning into nowheresville.

sure it was , on the xbox and gamecube

And yet look at Gran Turismo 4, God of War, Metal Gear Solid 3. Ten years ago a system that was 18 months older would not have been able to create visuals that matched or in some cases surpassed the more powerful machines. MS has realized this and has changed their architecture with the 360, and Sony is continuing on with the Cell, the next step of the EE.

The level of ease that the Cube and Xbox provided is no longer relevant in my opinion.

ps3 is not the market leader though. There is no market leader . If anything for the first year the x360 will be the market leader and if can keep the lead going into the second or third year you may see alot of dev support on the x360 and see the cell getting the shaft .

As I said it was its retail success that got the devs to push the system . If the retail success wasn't there the devs woulnd't be there and if the market was split between two systems it wouldn't have seen that lvl of support.

jvd, this is what you simply don't get. Devs don't have the luxury to wait around to see sales numbers, and they didn't wait around to see the PS2's sales numbers. When do you think Konami started MGS2, or Square started Final Fantasy X, or Rockstar started Grand Theft Auto 3, or Capcom started Devil May Cry? Many of the big name games need dev cycles of 1-3 years. Decisions have to be made now.

The first 2 years of this gen (and every gen from now on) was locked up before the machines even hit retail.

Devs know two things:

1) The PlayStation name is the strongest brand in videogaming right now, and it alone guarantee's more sales than any other factor for any other console. There are individuals who buy the PlayStation expecting to play certain companies games, but they do not follow whether those companies are actually developing those titles for the PS3. They just know the brand, and expect the games to be there.

2) Devs are fully aware they sell the hardware. Sony and MS can do their part with hype and what have you, but devs know they sell hardware. So what is the smarter bet right now? Go with the 360, a console that has a respectable brand name, but in only one region? Or go with the PS3, a console that is guaranteed to provide you with a solid audience from the get-go based on brand name alone?

The 360 right now is a gamble, and I can tell you straight up devs don't care about the headstart. No dev is looking to fulfill their bottom line on next gen consoles either this year or next, so to put off their title 6-12 months to benefit from the guaranteed install base of the PS3 is simply the smarter financial choice. I'm not saying some devs won't take that gamble, but the majority will not.

The brand name gets the sales which gets the games which gets more sales which gets more games. Until MS can finish a gen establishing the Xbox as a stronger brand in all territories than the PlayStation, it will not get the lionshare of early support. Sony has created a cyclical effect that benefits them each and every time so long as they finish out each gen on top.

The gaming industry is too big for gambles nowadays; if you want to survive you simply have to go with the best bet.

Sorry for the long post. 8)

 

[jvd]

It's not about 'fighting' the hardware, it's about adapting to its changes; changes that are made so that games can get better and better. The age of bumping up the clock speed to solve all our problems is gone. The devs that realize this are making the games I want to play. The devs that don't will quickly be following Lorne Lanning into nowheresville.

From what i've heard about ps2 development it wasn't a walk in the park and alot of times they had to fight the hardware to get it to do things they wanted and to come up with unique ways of doing it that just aren't needed on other hardware .

And yet look at Gran Turismo 4, God of War, Metal Gear Solid 3. Ten years ago a system that was 18 months older would not have been able to create visuals that matched or in some cases surpassed the more powerful machines. MS has realized this and has changed their architecture with the 360, and Sony is continuing on with the Cell, the next step of the EE.

And those games wouldn't have come out if the ps2 didn't enjoy the retail success it would never have gotten games like that . If the ps3 doesn't enjoy the sucess the ps2 did you wont see devs squeezing the same out of the cell .

The level of ease that the Cube and Xbox provided is no longer relevant in my opinion.

because none of them reached the sales success .

IF the xbox was the market leader do you really think half of those games would have come out on the ps2 or look as good ?

jvd, this is what you simply don't get. Devs don't have the luxury to wait around to see sales numbers, and they didn't wait around to see the PS2's sales numbers. When do you think Konami started MGS2, or Square started Final Fantasy X, or Rockstar started Grand Theft Auto 3, or Capcom started Devil May Cry? Many of the big name games need dev cycles of 1-3 years. Decisions have to be made now.

And ?

What is developed right now for each console wont matter 2 years into the race . What will happen is they will see sales data by then and support the system that is selling well .

That is what happens .

Developers went to the ps2 because they had a choice of the dc or ps2 and after the failure that was the saturn , 32x and sega cd people weren't all that quick to support it and ms was a new entry .

This time around ms will be out for a period before the ps3 adn will have a holiday alone , as you can already see alot of devs are more than willing to switch to the xbox or become multiplatform .


This article talks about how xna makes programing for the difficult cpu in the xbox 360 very easy but if u want to get full power you have to dive in . Puts in alot better position than the ps2 was in

 

[Frank]

jvd, rest assured. The xbox360 will probably do very well.



Just like the PS3. Give them both a chance, all right? Let's judge them on the games they run.

 

[JasonLD]

It's not about 'fighting' the hardware, it's about adapting to its changes; changes that are made so that games can get better and better. The age of bumping up the clock speed to solve all our problems is gone. The devs that realize this are making the games I want to play. The devs that don't will quickly be following Lorne Lanning into nowheresville.

From what i've heard about ps2 development it wasn't a walk in the park and alot of times they had to fight the hardware to get it to do things they wanted and to come up with unique ways of doing it that just aren't needed on other hardware .

And yet look at Gran Turismo 4, God of War, Metal Gear Solid 3. Ten years ago a system that was 18 months older would not have been able to create visuals that matched or in some cases surpassed the more powerful machines. MS has realized this and has changed their architecture with the 360, and Sony is continuing on with the Cell, the next step of the EE.

And those games wouldn't have come out if the ps2 didn't enjoy the retail success it would never have gotten games like that . If the ps3 doesn't enjoy the sucess the ps2 did you wont see devs squeezing the same out of the cell .

The level of ease that the Cube and Xbox provided is no longer relevant in my opinion.

because none of them reached the sales success .

IF the xbox was the market leader do you really think half of those games would have come out on the ps2 or look as good ?

jvd, this is what you simply don't get. Devs don't have the luxury to wait around to see sales numbers, and they didn't wait around to see the PS2's sales numbers. When do you think Konami started MGS2, or Square started Final Fantasy X, or Rockstar started Grand Theft Auto 3, or Capcom started Devil May Cry? Many of the big name games need dev cycles of 1-3 years. Decisions have to be made now.

And ?

What is developed right now for each console wont matter 2 years into the race . What will happen is they will see sales data by then and support the system that is selling well .

That is what happens .

Developers went to the ps2 because they had a choice of the dc or ps2 and after the failure that was the saturn , 32x and sega cd people weren't all that quick to support it and ms was a new entry .

This time around ms will be out for a period before the ps3 adn will have a holiday alone , as you can already see alot of devs are more than willing to switch to the xbox or become multiplatform .


This article talks about how xna makes programing for the difficult cpu in the xbox 360 very easy but if u want to get full power you have to dive in . Puts in alot better position than the ps2 was in

While it is not 100% certain that PS3 will be market leader, but with Sony coming from two straight generation of market leadership, PS3 will likely have most support from developers. Sony is proven leader, Microsoft is not up to that level yet. It sound like you are completely neglecting the importance of previous generation but it has major influence on next generations as well. If your theory is true, then first to the market will always be the market leader, but that never happened. X360 is in kind of position like Genesis back in 1991, but that didn't stop SNES from outselling Genesis did it? Unless Sony makes major mistake with PS3, X360 won't likely take over market leader position from Sony. Sony benefitted from Sega's major hiccups with Saturn and Nintendo's decision with catridge..and that propelled them into the market leadership. We are not seeing mistakes comparable to those from Sony(yet) and they are pretty good at handling transition as well. I said it before and I will say it again, market leadership won't change unless leader makes major mistake. I think industry is matured enough that we won't see situations like Sega ever again, but I don't know.

Damn...these are completely off topic.

 

[randycat99]

Aw, damn! Now you've just given jvd the bright idea to focus on more topics about how Sony is making all kinds of "major mistakes". (Trust me...it's not hard to see his gameplan)