Topic: IGN Promises Rev Specs: Evening 3/29/06

[KDR_11k]

Yep. Question is why would they choose not to take that approach, instead crippling their hardware's ability to run native software?

I don't get why anyone is complaining about Matt comparing the specs to the Xbox.

Because in the original version of the article he claimed that because the XBox has more MHz it's better. Those Rev numbers clearly outdo the Xbox but they fall hort of anything worth calling "next gen".

[NinGurl69 *huggles]

Quote
Originally posted by: wandering

Quote
Imagine Nintendo's follow up to the N64 was as powerful as the DS.

...except that the ds is less powerful than the n64 in some areas, and almost certainly not twice as powerful as the n64 overall.

What the hell.

N64 games typically ran at 256x224.  DS games are 256x192 for two screens.

Your N64 "favorites" like Zelda and Mario Kart typically ran at 20fps AT MOST.  Only in extremely rare occasions (like, never) were games like Excitebike 64 released that were able to reach a solid 30fps.

Yet I hear about Mario Kart DS (with 3D characters) maintaining 60fps.  That's a SIGNIFICANT jump.  Stictly comparing Mario Karts, DS is 3X as powerful.

[trip1eX]

Quote
Originally posted by: Ian Sane
I don't get why anyone is complaining about Matt comparing the specs to the Xbox.  Most of us aren't very technical.

I emailed the guy about that.  The problem with his comparison is it that it's bs.  His cpu comparison is like you saying my car does more rpms so it's faster  even tho you have a v4 in your car and I have a v8 under my hood.  He just perpetuates the whole Mhz myth.  It's even more irritating because we already knew the Gamecube was on par with the Xbox as it was.  And his comparison makes it sound like the Rev is just now as powerful as the Xbox despite a 50% increase in cpu power.  IT's bs.

I mean look at it another way.  He's saying the 'Cube needed a 50% power increase to equal the Xbox.  Now we know that's bullcrap.  ONe look at RE4 and it's easy to see that.  

I have no illusions that the Rev is going to be a 360 or PS3.  Nintendo has always said it isn't.  They are going there own direction.  But to say a 730 Mhz PPC cpu is somehow the same as a 730 Mhz Celeron processor is bs and really just an outright lie.  

 

[Kairon]

Quote
Originally posted by: KDR_11k
Metriod Prime 3 will be Metroid Prime plus massive environments. Pikmin Rev will be Pikmin plus thousands of enemies on screen. Etc. Would that be so bad?

No, that wouldn't be bad at all but these specs wouldn't even allow THAT.

And you know this...how?

~Carmine M. Red
Kairon@aol.com

[KDR_11k]

Cause the GC didn't get anywhere close to it and I know what to expect from such a small increase in processing power and RAM because I know what difference it makes on my PC. It's a relatively small step, what he's expecting is the machine to perform like 10x as well as the GC. The GC could handle hundred Pikmin with very limited AI and a few enemies, if that number was multiplied by 10 you'd need an appropriate increase in processing power to keep up.

[Kairon]

I'd like to know if anyone tried epic scale games on the GC. A Miyamoto game is not what you should look at for AI or physics levels. His games are so resolutely built around gameplay that he probably didn't care for any advanced model numbers or AI. Those are tools for creating a simulation, which is exactly what what Pikmin is NOT, Pikmin is a puzzle-action game. In fact, he moved in an opposite direction for Pikmin 2, balancing the game for smaller and smaller numbers of Pikmin.

Ah, but then let's look at a launch window game developed in 9 months for the GC, Rogue Leader. The Battle of Endor, especially for a Star Wars fan, is epic. You're looking at about 4-5 capitol ships on the rebel side, maybe 50+ ties and more than a dozen rebel ships, all with models far more complex than in the world of Pikmin, and all moving in action. This was a first-generation title with a 9 month development time and look what it was able to do. And of course, if you're gonna call the Ties dumb, the thing is that AI is a matter of code, not cpu. It only becomes a matter of cpu if you need to do real AI, and then I don't think even our resident supercomputers are good enough for that. Anyways, the Rogue-Leader title wasan action-shooter, which specifically necessitated enemies that you would be able to down by the truckloads.

Another near-epic example could be from (the admittedly deeply flawed) Rogue Leader 2 Extreme Survival Modes where players would need to down 100s of Ties on a single life above the endless deathstar towerscape, or the evacuation of Yavin which took place over a Junglescape.

But anyways, here we have this epic game with great graphics, tons of units on the battlefield, developed in 9 months and as a first generation title. Why does the GC seem to not have any epic titles other than this? It isn't for lack of computer power, think what could've been done with a longer dev time, a different game genre and a later generational attempt. No, it's because no one aside from Factor 5 even tried.

Oh, and unless we're hardware geeks who can explain EXACTLY why a PPC core is better than a regular P3, I don't think we have any real authority on how powerful a chipset is. Well, on top of that, IGN just gave us clockspeeds, not anything else on the chipsets, which makes judgements of worth by non-enthusiasts even less factually sound.

~Carmine M. Red
Kairon@aol.com  

[RiskyChris]

Quote
Originally posted by: KDR_11k
Yep. Question is why would they choose not to take that approach, instead crippling their hardware's ability to run native software?

I don't get why anyone is complaining about Matt comparing the specs to the Xbox.

Because in the original version of the article he claimed that because the XBox has more MHz it's better. Those Rev numbers clearly outdo the Xbox but they fall hort of anything worth calling "next gen".

What is next gen?  I think it involves the whole package, not just hardware specs.  The controller is truly next-gen (moreso than any other controller in the past).

The DS is a "next gen" handheld because it combines innovation with functionality.

[Arbok]

Quote
Originally posted by: RiskyChris
What is next gen?  I think it involves the whole package, not just hardware specs.  The controller is truly next-gen (moreso than any other controller in the past).

The DS is a "next gen" handheld because it combines innovation with functionality.

The DS was both, let's admit that. The system is *much* more powerful than the GBA was, its predecessor, while also adding in the extra screen that it has to deal with. Is it as powerful as a PSP? No, but then it's a portable and technology in that sector is much more expensive and there are things to consider like battery life that one doesn't have to put into consideration when making a console system. In other words, the DS was both a huge step in process power and a huge leap in innovation. In terms of the Revolution, it seems like Nintendo just went with the latter and kind of skimped on the former. I hope I'm proven wrong, but it's all speculation without screens or footage at this point at what results it might generate.

[RiskyChris]

I can't see the difference without scrutinizing images between RE4 and many 360 games (not all, of course).  The jump from this gen of consoles to the next gen is so trivial compared to the days of SNES > N64 that next-gen can't solely be decided on horsepower anymore.

I bring up the DS because it sits compared to the PSP similar to how the revolution will be versus the PS3 and 360.  The revolution is unquestionably stronger than the gamecube!  How can you call the DS a next-gen worthy system, when the jump from GBA to DS is very much analogous to the jump from GC to Rev?  Innovation included.

I'll definitely be picking up a PS3 sometime in 2007 (sooner if a killer app demands it), but to question the revolution's place as a valid next-gen system is silly to say the least.

[BlackNMild2k1]

I think everyone should just see these 'next-gen' as if they were handhelds for a second:

Rev would = DS, alternative input method and processing power split between visuals and control method. designed around creating new types of games.

PS3 & X360 = PSP, bumped up specs and 'more of the same' games only in HD, now with lots of non-game functions added.

[Smash_Brother]

Quote
Originally posted by: BlackNMild2k1
I think everyone should just see these 'next-gen' as if they were handhelds for a second:

Rev would = DS, alternative input method and processing power split between visuals and control method. designed around creating new types of games.

PS3 & X360 = PSP, bumped up specs and 'more of the same' games only in HD, now with lots of non-game functions added.

Yes! FFS, people, listen to BNM and understand the comparison.

I'm quite positive that it will be JUST fine.

Plus, did you read the last quote from Matt?

Quote
I am, you could say, fully emerged in the so-called "HD era" of videogames. And yet, I'm promising you that you're going to be impressed with the visuals in some of the Revolution games on the horizon. I guess for now you'll just have to trust me on that point.

Now, can we STOP the preemptive bitching and wait until we have SOMETHING upon which to make a judgment call before we make judgment calls?

[Ian Sane]

"Rev would = DS, alternative input method and processing power split between visuals and control method. designed around creating new types of games.

PS3 & X360 = PSP, bumped up specs and 'more of the same' games only in HD, now with lots of non-game functions added."

Even if the DS didn't have the touchscreen it would still be seen as next-gen compared to the GBA.  It would be a big deal just because it was a 3D portable with online support.  Mario Kart DS for example doesn't really use the unique DS features but it would still impress because on a portable that level of 3D graphics with that framerate and online play would be a big deal.  Now the PSP kind of kills some of the impact but considering the price for that Sony is kind of overdoing it.

Let's assume Sony and MS are overdoing it with the PS3 and X360.  Let's assume they're going beyond the normal hardware upgrade and thus the price is going to be too high.  If the Rev didn't have the remote would it be considered good enough to be a suitable followup to the Gamecube?  Ignoring the competition, under normal circumstances would the Rev hardware be a big enough jump from the Cube?  I question if it is.  The DS without the touchscreen is still a worthy followup to the GBA.  Ideally the Rev should be good enough that if it were a traditional console we would consider it a worthy followup.

You can say that horsepower isn't all that matters anymore but that doesn't mean it shouldn't be there.  I'd say that perhaps we need more than just horsepower, but that doesn't mean we shouldn't get it as well.  "Not enough anymore" isn't the same thing as "don't need it anymore".

[Dirk Temporo]

Question to someone who understands all this: Would Rev be powerful enough to run games such as Gears of War, and Unreal Tournament 2007? (Note, I am not saying they'll be on Rev, but if push came to shove, could the Rev do it?)

[vudu]

Quote
Originally posted by: Ian Sane
I don't get why anyone is complaining about Matt comparing the specs to the Xbox.  Most of us aren't very technical.  If you just give us a number without any frame of reference we won't know what it means.  So it just makes sense to compare the specs to other consoles to give an idea of where the Rev fits.  If you want to know what the specs are then you want to know how they compare with the Xbox 360.

Except there's one problem.  The comparisons aren't very good.

The United States is approximately 5,983,240 square miles.  Canada is roughly 6,198,886 square miles.  So we can assume they have a similar population size, right?  Nope.  In 2002 the population of the USA was 280,562,489, while the population of Canada was only 31,902,268.  So even though the USA is slightly smaller than Canada in size (96%), it contains almost nine times as many citizens.

Point being, you can't compare raw specs and expect to draw any meaningful conclusion.

Quote
I have a 50" plasma screen hanging on my living room wall. I have a 30" LCD on my dining room wall. And I have an HD CRT in my bedroom.

How much does IGN pay this man?  

[KDR_11k]

And of course, if you're gonna call the Ties dumb, the thing is that AI is a matter of code, not cpu. It only becomes a matter of cpu if you need to do real AI

Nonsense. There are many script-AI algorithms that take more CPU than others without implementing real AI. Pathfinding, for example. Pikmin tend to get stuck on simple obstacles. Raph Koster said that 90% of the CPU load on Everquest servers is just simple pathfinding for the monsters. Especially if you want the AI to dynamically assess a situation and use the terrain to its advantage you'll have to use CPU power for that. Or you could try marking everything in the level with helper entities but the player will exploit that.

You're looking at about 4-5 capitol ships on the rebel side, maybe 50+ ties and more than a dozen rebel ships, all with models far more complex than in the world of Pikmin, and all moving in action.

Don't forget LODs. Do you really think that tie fighter that's drawn as five pixels on your display has a few thousand polygons? Only the immediate surroundings of the player need that much detail, the rest simply uses LOD models* that can go down as far as 1% of the original polycount (the BF2 modder help says 4-5% for vehicles, for example). Now that's not a problem if you can get your points of interest spread out far enough (like the ocean in Windwaker) so you don't have to handle that many of 'em at full detail but in something like Pikmin all hundreds or thousands of them can be on screen at once and taking up significant screen space each.

Well, on top of that, IGN just gave us clockspeeds, not anything else on the chipsets, which makes judgements of worth by non-enthusiasts even less factually sound.

No, Matt said quite clearly that this thing is just a slightly upgraded GC CPU. The architecture is ancient, if you wanted to make a somewhat modern chip out of it you'd throw it away and build a new one from scratch (adding some legacy operations to make it backwards compatible). You can tune a Trabbi all you want, it's not going to become a Mercedes.

I can't see the difference without scrutinizing images between RE4 and many 360 games (not all, of course). The jump from this gen of consoles to the next gen is so trivial compared to the days of SNES > N64 that next-gen can't solely be decided on horsepower anymore.

RE4 had very limited player freedom and as such could use a lot more trickery than, say, Oblivion. On the XCircle you could do a game on the level of RE4 without nearly as much effort. Once you do put in the effort it becomes a lot more.

I think everyone should just see these 'next-gen' as if they were handhelds for a second:

Rev would = DS, alternative input method and processing power split between visuals and control method. designed around creating new types of games.

That comparison would be more valid if the GBA was a portable N64 in terms of power.

*= In Doshin The Giant it's very obvious, you can see villagers change from being sprites with two frames of animation to 3d models with fluid animation.

[animecyberrat]

I dont know about everybody else but I do know that when i buy a newxt generation video game system I ALWAYS have been impressed with the graphical improvements, even 360 has shown that itis an improvement over xbox, sure its going to take some time to get it to its full potential but right now its already better looking than normal xbox, if Revolution comes out and it lookx exactly like GC with just more RAM and a few fancy efects I am seriously going to reconsider it as my main system.

I am not alone if the next gen system doesnt look next gen then its not as good as the others. Games do make the system and if the games dont look good that does hurt thier overal appeal.



Hearing that GC did Resident Evil 4 does not comfort me much, because as good as it looks, 360 already has games that look better, and they are just the start. I want Revolution to look BETTEr than resident Evil 4 not as good as. Every GC game could have looked like that if they had put in the effort. Less RAM is going to hurt a lot also, and I dotn care about undercloking and crap like that, the CPUs speed has little to do with how well it performs, same with GPU, its about what images it can display on screan and what effect etc it can do with those images. now I am not a graphics whore, but has not Nintendo fanboys always claimed Nintendos superior graphiocs every generation as a reason for liking their system, now all of a sudden REV is going to have weaker graphics and now its graphics dont matter, grow up.



Graphics DO matter, they are all that makes a great game but they DO HELP make a great game. I dont care about HD much cuz I dont see hwo its going to benefit games at all, but smother animations, mroe detailed characters and more lifelike enviromnets always improve games. I dont want ugly games that dont wow me, GC over N64 was a huge wow, but GC to REv soudns more like comparing GC to Xbox. Now that doesn not change myf aith in Nintendo because I do think Rev will look decent enough but if tehre are gorgeous games on the competitors systems that arent possible on rev than yes its going to make a differnce.  

[jasonditz]

Quote
Originally posted by: Dirk Temporo
Question to someone who understands all this: Would Rev be powerful enough to run games such as Gears of War, and Unreal Tournament 2007? (Note, I am not saying they'll be on Rev, but if push came to shove, could the Rev do it?)

The short answer is, probably not.

Mark Rein has already gone on record as saying the Rev is probably not going to be a target for Unreal Engine 3 (which is what both of those games are running on).

I don't know that this means it necessarily can't theoretically run it at all, but if the specs from IGN hold, I would guess it wouldn't be able to in any meaningful way.

OTOH, since the Rev isn't support HD at all, it's still very likely that a game designed to take advantage of the hardware could do everything that games like that can do on an Xbox 360 in 480p mode.  

[jasonditz]

Quote
Originally posted by: animecyberrat
dont want ugly games that dont wow me, GC over N64 was a huge wow, but GC to REv soudns more like comparing GC to Xbox.

So what's the game on the Xbox that looks as good as RE4?

This right here is exactly why I'm annoyed at the way IGN leaked these. I'm fine with them reporting the specs, but Matt's ignorant Apples to Oranges comparison of CPU clock speeds is spilling over into tons of people's views of the system.

The fact of the matter is, RE4 looked great, and even the worst case scenario of the Rev is more than a little better in every conceivable way than the Cube, so a game made with a comparable amount of care and attention to detail is going to look more than a little bettter. I'm going to go out on a limb here and say at some point, there is going to be a Rev game that visually slaughters the average title on the 360 and the PS3. That's not to say the average sloppy third party Rev title won't look worse, but when Nintendo or one of their key 3rd parties like Capcom or Namco sets their mind to it, they're going to blow us away with what this system to capable of.

[NinGurl69 *huggles]

I think we've seen enough.

[Jensen]

Quote
Originally posted by: KDR_11k
Yep. Question is why would they choose not to take that approach, instead crippling their hardware's ability to run native software?

For the same reason the Game Boy Advance Micro and the DS don't run Game Boy Color Games... Size and Cost.

[Kairon]

Quote
Originally posted by: KDR_11k

Quote
Nonsense. There are many script-AI algorithms that take more CPU than others without implementing real AI. Pathfinding, for example. Pikmin tend to get stuck on simple obstacles. Raph Koster said that 90% of the CPU load on Everquest servers is just simple pathfinding for the monsters. Especially if you want the AI to dynamically assess a situation and use the terrain to its advantage you'll have to use CPU power for that. Or you could try marking everything in the level with helper entities but the player will exploit that.


Don't forget LODs. Do you really think that tie fighter that's drawn as five pixels on your display has a few thousand polygons? Only the immediate surroundings of the player need that much detail, the rest simply uses LOD models* that can go down as far as 1% of the original polycount (the BF2 modder help says 4-5% for vehicles, for example). Now that's not a problem if you can get your points of interest spread out far enough (like the ocean in Windwaker) so you don't have to handle that many of 'em at full detail but in something like Pikmin all hundreds or thousands of them can be on screen at once and taking up significant screen space each.

How can you relate an MMO to what we're talking about here? Not only does an MMO have to simulate an enormous simultaneous persistant world, but it does all this OFF of your computer. Sure, 90% of the server load is pathfinding, but when you can run WoW on an 800 mhz Pentium 3 computer, obviously none of that has to do with the cpu in the client side machine. And I doubt any client side machine will ever need to simulate as much as an MMORPGs server needs to.

You've convinced me that AI routines are not independent of the cpu, but we still have no idea how cpu intensive they are, using terrain as cover or not, (because, all the MMORPG example does is show that MMORPG servers have a tough job), ESPECIALLY as they relate to the as-of-yet unmeasured strength of the Rev's cpu. Probably if the X360 has a hundred different characters it has to keep intelligent simultaneously, then I'd agree that the Rev couldn't do it. But even something like Oblivion, with it's supposed so-many-npc-s-have-real-lives thing going on strikes me as little more than Zelda:MM daily npc routines with more spit and polish.

And this brings us to the epicness of battles. I guess you'd call a game like Ninety-Nine Nights epic right? Huge fields of thousands of warriors. Well, you see what good that is when you plow through hundreds upon hundreds of them at a time with the application of a few button presses. There is a definite limit to how much of anything you want on the screen before they either be trivialized through gameplay or players learn to handle the visual bombardment by developing multiple complex eyes and enlarged occipital lobes.

~Carmine M. Red
Kairon@aol.com

[BigJim]

Comparing the specs isn't a bad thing for those that don't understand, Matt just used a bad comparison.

Ignore specs for a second. Most developers will tell you that GameCube = Xbox. If Xbox can do it, GameCube can do it. (It was the PS2 that was the bitch to work with because it needed optimizations that the others didn't.)

If GameCube = Xbox,  then GameCube + 50% = Significantly better than Xbox.

We can't put a quantitative value on it yet, but we know it's better. I know we're used to 2x, 10x, whateverx, performance increases each generation, but that amount of performance boost isn't necessary anymore when you're reaching a point of diminishing returns (especially if you're not doing HD.) There would be some nice things they could do, like add more shaders, but whose to say what the final hardware will add yet, if anything. Matt says no, but we'll see.

So just ignore the paper stats and keep the above equation in mind.

"Would Rev be powerful enough to run games such as Gears of War, and Unreal Tournament 2007? (Note, I am not saying they'll be on Rev, but if push came to shove, could the Rev do it?)"

As mentioned above, probably not. UE3 is geared towards high-end equipment and multi-core environments. But I think he did say the Unreal Engine 2 would work for Revolution. Unreal Engine 2 was used for UT2004, Splinter Cell, XIII, Americas Army, etc. It's not bad by any means.

Here's a UE2 page:  Unreal Engine 2

If you want to see what's possible on the next-gen systems, compare the UE2.x and UE3 visuals on that site.  Sure UE3 is a lot better, but I'm not crying over UE 2.x. I think this is the best way we have right now to compare their capabilities.    

[KDR_11k]

How can you relate an MMO to what we're talking about here? Not only does an MMO have to simulate an enormous simultaneous persistant world, but it does all this OFF of your computer. Sure, 90% of the server load is pathfinding, but when you can run WoW on an 800 mhz Pentium 3 computer, obviously none of that has to do with the cpu in the client side machine. And I doubt any client side machine will ever need to simulate as much as an MMORPGs server needs to.

Your home console doesn't have to simulate as much but it'll have to do it on much less power and using more complex AI scripts (MMO monsters are DUMB). I've heard devs talk about AI scripts that didn't make it into the final game because the system couldn't handle them. Less CPU means dumber entities, especially when we're talking about the numbers seen in Pikmin and the like.

But even something like Oblivion, with it's supposed so-many-npc-s-have-real-lives thing going on strikes me as little more than Zelda:MM daily npc routines with more spit and polish.

Zelda had a less dynamic environment, you couldn't e.g. kill a mayor and see the town plunge into disarray. Okay, so I'm not sure if Oblivion does that either. But take X. That simulates an entire universe on your local system. It has to use simplified simulations for things far away that greatly alter the behaviour.

My favourite example for system power is crowds. Ten people on a street aren't a crowd. Having to push through a crowd chasing someone isn't possible on current-gen consoles. Worst case would be a game where you have to pick out robots or aliens (Frog blast the vent core!) looking like humans out of a crowd (think Bladerunner). The robots would have minor differences to the humans, perhaps even limited to the shaders used on the skin. That'd require being able to render MANY people at once AND giving them enough details for such minor imperfections to become visible.

Crowds change the game dynamic a lot. You could scatter the crowd by firing your weapon into the air but you might alert enemies to your presence or not be able to pick out the guy you're chasing between all the running people.

Big Jim: I don't think that Revolution could even handle Unreal Tournament 2004. Don't point at Unreal Championship, that's downscaled.

[Caliban]

dedededede de dedededede de lalalalalalalalalalalala la ao fim do mundo lalalalalalalalalala nanananananananana

[UltimatePartyBear]

Large groups of entities don't require separate AI for each individual.  It's easier to model schooling or mob behavior until something breaks up the group, and modeling panic behavior when that happens is much simpler still.

You can see the concept at work thanks to a flaw in Rogue Leader.  TIE fighters fly in formation, with all the members simply following the leader.  Unless you fire on the leader, they won't react to you, so you can learn which member is the leader and pick off all his wingmen without much effort (unless you're a particularly bad shot, in which case the one you're shooting at will eventually break off when the script gets around to making a check for that, or whatever causes it).  Fire on the leader, however, and he and the group will start trying to evade you.

The pikmin didn't have any noticeable pathfinding at all.  Even rudimentary pathfinding would have worked better than their simplistic beeline approach, which suggests that there wasn't any effort put into the task for us to judge.  The pikmin had no self-preservation tendencies, either.  In fact, the only evidence of AI I can think of is their single-minded pursuit of nectar.  I'd say the almost total lack of AI for the pikmin was pretty clearly an intentional part of the game design.  Maybe that was because the AI would have been too expensive, but I doubt that.