IGN gives new Rev specs (virtually Gamecube 1.5)

[Hamel]

Matt at IGN is pulling your legs again

Insiders stress that Revolution runs on an extension of the Gekko and Flipper architectures that powered GameCube, which is why studios who worked on GCN will have no problem making the transition to the new machine, they say. IBM's "Broadway" CPU is clocked at 729MHz, according to updated Nintendo documentation. By comparison, GameCube's Gekko CPU ran at 485MHz. The original Xbox's CPU was clocked at 733MHz. Meanwhile, Xbox 360 runs three symmetrical cores at 3.2GHz.

Revolution's ATI-provided "Hollywood" GPU clocks in at 243MHz. By comparison, GameCube's GPU ran at 162MHz, while the GPU on the original Xbox was clocked at 233MHz. Sources we spoke with suggest that it is unlikely the GPU will feature any added shaders, as has been speculated.

"The 'Hollywood' is a large-scale integrated chip that includes the GPU, DSP, I/O bridge and 3MBs of texture memory," a studio source told us.

The overall system memory numbers we reported last December have not greatly fluctuated, but new clarifications have surfaced. Revolution will operate using 24MBs of "main" 1T-SRAM. It will additionally boast 64MBs of "external" 1T-SRAM. That brings the total number of system RAM up to 88MBs, not including the 3MB texture buffer on the GPU. By comparison, GameCube featured 40MBs of RAM not counting the GPU's on-board 3MBs. The original Xbox included 64MBs total RAM. Xbox 360 and PlayStation 3 operate on 512MBs of RAM.

It is not known if the 14MBs of extra D-RAM we reported on last December are in the current Revolution specifications.

"The external RAM can be accessed as quickly as the main RAM, which is a nice touch," a developer we spoke with alleged.

Lots of numbers, but what do they all mean? The short answer is that Revolution is exactly as Nintendo has publicly stated: a console whose primary focus is not quadrupling raw horsepower, but rather a potentially gameplay-changing new controller. Nintendo's new hardware supports this innovative new peripheral and not the other way around. Looking back, it makes sense.

That last paragraph is pure wank. His own sources put Rev as being 2-3x Cube's power. However, now it isn't even 2x. These numbers from another forum caught my attention:

485MHz x 1.5 = ~729
162MHz x 1.5 = 243

Gamecube 1.5. It isn't even April 1st yet. WTF. No wonder why Mark Reign won't have UE3 ported to the system.

Without ranting forever about it, I will say that Nintendo can and should do a lot better than this. Betchu this a joke

[Flagg]

Without ranting forever about it, I will say that Nintendo can and should do a lot better than this. Betchu this a joke

Hasn't that been Nintendos hardware M.O. since the N64?

[Hamel]

Without ranting forever about it, I will say that Nintendo can and should do a lot better than this. Betchu this a joke

Hasn't that been Nintendos hardware M.O. since the N64?

Nintendo's M.O. was to do better than this? Maybe you worded your question wrongly considering the point you addressed.

[Flagg]

Without ranting forever about it, I will say that Nintendo can and should do a lot better than this. Betchu this a joke

Hasn't that been Nintendos hardware M.O. since the N64?

Nintendo's M.O. was to do better than this? Maybe you worded your question wrongly considering the point you addressed.

Let me clarify, Ninetendos M.O. has been that they "can and should do better than this".

[SylasGaunt]

If those numbers are accurate I'm not overly shocked. Nintendo has stated outright IIRC that they aren't even trying to compete hardware-wise with the 360 and PS3.

[DarkSilver]

Indeed, Nintendo has never said they were building something tocompete with the PS3 and Xbox360 on thier levels, so these numbers (if they are, indeed true and not idle speculation or guesses) wouldseem approximatly right. On the same level as the original Xbox, as was reported some months back.

[Hamel]

A GPU with the performance and pixel shading power of a 9800 pro would give the system a a good boost in power, and that is an incredibly weak GPU compared to what is in the 360. Less than half as powerful in all likelyhood.

A system with that level of power could output visuals in 480p which would look significantly better than even the best GC games. If Nintendo could put an x1600 level of power into it, all the better. Simply a 1.5x overclock with 64 megs extra RAM and no modern pixel shading power will mean sharper textures and smoother running GC-level games. That's it.

[TheBlackCat]

That's assuming raw clock speeds are sufficient to describe a processor's power. They aren't. The processor architecture and layout have a large part to play as well.

[darthdavid]

That's assuming raw clock speeds are sufficient to describe a processor's power. They aren't. The processor architecture and layout have a large part to play as well.

However it's basically a speed bump on the current processor so clock speed comparisions between the two are a good method of measuring the performance difference. Next time actually read the article asswipe.

[Praxis]

IGN's last article, however, noted that the current dev kits aren't at the final levels and they expect the next revision to be closer to the system's final power.

I wonder how accurate this is.

[TheBlackCat]

However it's basically a speed bump on the current processor so clock speed comparisions between the two are a good method of measuring the performance difference. Next time actually read the article asswipe.

Perhaps. We know that the architectures are similar, but we don't know that they are exactly the same. My point, however, was more regarding clock speed comparisons to the Xbox that are not valid, yet the article uses them repeatedly. They freely admit that this is not a valid comparison, yet they still make it anyway.

[Praxis]

This is all IIRC, please correct me where I'm wrong, as I don't want to repeat the wrong information to others.

The XBox's processor is a Pentium 3 with half the cache disabled, putting it closer to a Celeron in performance. The GameCube's is a PowerPC processor similar to a G3.

Clock for clock I would think that the GameCube's processor is faster than the Pentium 3 (though not by a very significant amount), so this would in fact be faster than the XBox CPU-wise, correct? And that I can freely bash the article's claim that it "falls short of the XBox" on account of the "MHz Myth"?

I hope Nintendo at least increased the amount of cache on the processor.

[Ace Pace]

That's assuming raw clock speeds are sufficient to describe a processor's power. They aren't. The processor architecture and layout have a large part to play as well.

However it's basically a speed bump on the current processor so clock speed comparisions between the two are a good method of measuring the performance difference. Next time actually read the article asswipe.

Read the comparison, they are comparing the Gamecube/Revolution's CPU to the Xbox360, going by the GHZ Myth

[Pezzoni]

If I remember corerctly, the Gamecube CPU uses RISC, whereas the Xbox one is CISC. MHz to MHz comparisons are a bad idea at the best of times (look at an Athlon versus a P4), but when an entirely different instruction set is in use, they are even worse.
Somebody who makes those comparisons in an article such as the above really isn't in all that much of a position to be taken seriously, it just looks like blatent fanboyism.

[Vendetta]

Has no-one been paying any attention to anything Nintendo have actually said about the Revolution up to this point?

They are not going to even try to place themselves in the same market as Microsoft and Sony. They have looked at the market trend for shinier graphics, higher resolutions, and bigger numbers on the chips and said "We can do better than this".

Hence the wand, hence the low price point, hence the virtual console... The obvious decision at Nintendo is that games are not sufficiently improved by the extra shininess to warrant the price they would be paying, and that putting out, essentially, an overclocked Gamecube with more innovation in interactivity than it's competitors had imagined is the way to improve the game experience.

They are doing what they did with the NES. Rather than trying to compete harder and harder for the same market, they have looked for where there might be a new market entirely. Paradigm shift, market-speak, Revolution!.

[Praxis]

You know, if the paradigm shift worked, and people dodn't buy enough XBox 360 and PS3 games and online content to make up for the ENORMOUS losses Sony and Microsoft are taking to put it on the market, they could take a rather serious loss there.

I agree about what you just noted Vendetta. My only complaint; I HAVE been payinbg attention to what Nintendo actually said. One quote:

Reggie: Quite frankly, we'll share more of our technical specs at E3. That said, we believe a stronger CPU and a great graphics card is just the price of entry.

They've also said they don't think there will be a big difference between the other systems.

My opinion has always been it would be in the position of the Dreamcast, halfway between the generations. Since they consider graphics a "price of entry" and not a prime focus, it would be not equal to the other systems but considered the same generation in terms of graphics...

This isn't much more than the XBox. I would have even been happy with the 2-3x figure- a 1 GHz processor & 300-400 MHZ GPU.

That said, I'll buy the console anyway. Just for the Virtual Console and unique experiences with the controller.

[The Dude]

If I remember corerctly, the Gamecube CPU uses RISC, whereas the Xbox one is CISC. MHz to MHz comparisons are a bad idea at the best of times (look at an Athlon versus a P4), but when an entirely different instruction set is in use, they are even worse.
Somebody who makes those comparisons in an article such as the above really isn't in all that much of a position to be taken seriously, it just looks like blatent fanboyism.

As I read it, the point wasn't to directly compare the clock speeds on the 360 or Xbox to the Rev, but to compare the generation increases.

The 360 has 4X faster CPUs (and three of them) and 8X more RAM than the Xbox. Conversely, the Revo has a 50% faster CPU and 2X the RAM as compared to the Gamecube.

Anyway, if the Revolution doesn't do HD, does it really need to be much more powerful than the Xbox?

[Ace Pace]

If I remember corerctly, the Gamecube CPU uses RISC, whereas the Xbox one is CISC. MHz to MHz comparisons are a bad idea at the best of times (look at an Athlon versus a P4), but when an entirely different instruction set is in use, they are even worse.
Somebody who makes those comparisons in an article such as the above really isn't in all that much of a position to be taken seriously, it just looks like blatent fanboyism.

As I read it, the point wasn't to directly compare the clock speeds on the 360 or Xbox to the Rev, but to compare the generation increases.

The 360 has 4X faster CPUs (and three of them) and 8X more RAM than the Xbox. Conversely, the Revo has a 50% faster CPU and 2X the RAM as compared to the Gamecube.

Anyway, if the Revolution doesn't do HD, does it really need to be much more powerful than the Xbox?

Ah, but the problem is, that 4x faster CPU is only about 2x better, since its an in order chip. Getting maximum preformance out of that is a very differant task then out of order design.

[phongn]

Clock for clock I would think that the GameCube's processor is faster than the Pentium 3 (though not by a very significant amount), so this would in fact be faster than the XBox CPU-wise, correct? And that I can freely bash the article's claim that it "falls short of the XBox" on account of the "MHz Myth"?

Actually, I'd expect the Xbox's CPU to be faster - late-model P3s were pretty good, even if the Xbox's varient had inferior cache.

If I remember corerctly, the Gamecube CPU uses RISC, whereas the Xbox one is CISC. MHz to MHz comparisons are a bad idea at the best of times (look at an Athlon versus a P4), but when an entirely different instruction set is in use, they are even worse.

This is true, but the P3 was pretty good for its time (if not quite the equal of K7) and IMHO better than the G3-class PPC in the GameCube.

[Admiral Valdemar]

So, it comes to this. Two of the Big 3 are going for bigger, brasher systems that utilise cutting edge technology at a price to give you super shiny 3D games in HD. The other party is going for maximising the potential of what we already have and giving innovation the floor rather than somewhat ridiculous graphical additions to the same old tat.

We shall see who wins. Given the way PC games have been going in many areas, it is shiny FPS games using gameplay from the early nineties that are winning over games that look decent but are amazingly playable and innovative.

[Ace Pace]

So, it comes to this. Two of the Big 3 are going for bigger, brasher systems that utilise cutting edge technology at a price to give you super shiny 3D games in HD. The other party is going for maximising the potential of what we already have and giving innovation the floor rather than somewhat ridiculous graphical additions to the same old tat.

We shall see who wins. Given the way PC games have been going in many areas, it is shiny FPS games using gameplay from the early nineties that are winning over games that look decent but are amazingly playable and innovative.

The sucess of GalCiv2 is saying something else.

[Hamel]

Anyway, if the Revolution doesn't do HD, does it really need to be much more powerful than the Xbox?

To present an immediately noticable improvement in visuals from a GC or Xbox, yes, it would have to be much more powerful than IGN says it is.

I don't understand why people have to masturbate while talking about how Nintendo is trying to find a different audience, and is not going to compete head to head with expensive hardware. We know that, but there doesn't have to be an extreme dillema of unique gameplay vs HD "make you lol" graphics. They are targetting 480p, so something like a 9800pro for the GPU isn't asking for the world. I mean, come on, no fucking pixel shaders.

[Praxis]

I wish it was the "2-3x" figure given before Hopefully it is faster than it sounds- maybe the cache and bus speed are faster.

[Ace Pace]

I wish it was the "2-3x" figure given before Hopefully it is faster than it sounds- maybe the cache and bus speed are faster.

Depending on the chip design, it could be. Honestly, I was thinking they would need some kind of special chip just to handle the control interface, but that wouldn't affect the CPU.

[TheBlackCat]

We shall see who wins. Given the way PC games have been going in many areas, it is shiny FPS games using gameplay from the early nineties that are winning over games that look decent but are amazingly playable and innovative.

You can't compare the PC gaming industry to the console industry like that. FPS and RTS games are really the only genres where the PC has a clear advantage over consoles, so one would expect them to do better than games that tend to be better on consoles.