Post by: )Dark-LInk( on February 12, 2003, 12:59:23 PM
i think they should ally AN AMERICAN COMPANY(WEIRD) the great NVIDIA! then theyl have the best GRAPHICS for sure! what do ya think?(hopefully NVIDIA wont go MAD like with MS for the reduction cuz then THERE IS A LAWSUIT for IT!)
Post by: Bill Aurion on February 12, 2003, 01:12:27 PM
Post by: Perfect Cell on February 15, 2003, 03:07:25 PM
Post by: Darc Requiem on February 20, 2003, 05:38:37 AM
Darc Requiem
Post by: Grey Ninja on February 21, 2003, 09:50:48 PM
Anyways, I would just like to say that ATI wasted nVidia recently, and I am hoping for more of the same out of them in the future.
But that being said, ATI did NOT make the GPU in the GameCube. It was ArtX. ATI bought ArtX right as Flipper was being taped out.
Post by: mrbojangos on February 22, 2003, 01:23:52 PM
Post by: Berto2K on February 24, 2003, 10:49:37 PM
Post by: Darc Requiem on February 27, 2003, 02:25:16 PM
Darc Requiem
Post by: Riotamus on February 27, 2003, 02:57:33 PM
Post by: Pilotwings on March 04, 2003, 09:10:22 PM
the CPU will probably be co-designed by NEC-Cray and possibly Fujitsu. maybe as a joint effort to counter the Sony-IBM-Toshiba designed CELL for PS3. an NEC-Cray-Fujutsu partnership is not much better than wild speculation though. perhaps just NEC or IBM. but the GPU is probably a lock with ATI if you have seen the latest news on any of the Nintendo websites or www.ATI.com > press releases.
ATI will probably take their in-development R500 (West Coast team with many ArtX engineers) and modify that for Nintendo, incorporating parts to make it not only backwards compat with Flipper, but more powerful than the standard PC version of the R500 which could be out as early as fall 2004 using the .09 micron process. maybe ATI will use a R500 refresh, R550, as the basis for the Nintendo GPU with the additions i mentioned above. the newer the GPU the better, since this next Nintendo will be coming out in late 2005 or late 2006 and have to last until 2010-2011. HDTV resolution @ 60fps with exellent FSAA will be a must for this console. as will plenty of memory this time. at least 512 MB if not more.
Post by: BlkPaladin on March 05, 2003, 07:38:43 AM
NEC will more than likely be producing the chips again. They did it for the N64, Gameboy, Gameboy Advance, and the Gamecube so I think that partnership is in no danger of collapsing anytime soon.
Post by: manunited4eva22 on March 05, 2003, 10:42:59 AM
Post by: [o]CUBE on March 06, 2003, 03:10:25 AM
Post by: manunited4eva22 on March 06, 2003, 10:20:14 AM
Post by: )Dark-LInk( on March 06, 2003, 12:29:05 PM
ALSO they should get NVIDIA and pay them ALOT OF CASH to make a GRAPHICS chip at least 3x better then the GEFORCE4 ti 4200!then N wil have a very strong console!
Post by: manunited4eva22 on March 06, 2003, 03:26:44 PM
Post by: )Dark-LInk( on March 06, 2003, 04:24:11 PM
well if they cant they should at least make it 5x better(2.5 ghz low but alright....)
Post by: BlkPaladin on March 06, 2003, 08:30:49 PM
Post by: Tael on March 06, 2003, 10:26:50 PM
QuoteWhy not get ATi to make a chip three times faster than the Radeon 9800 Pro?
Originally posted by: )Dark-LInk(
ALSO they should get NVIDIA and pay them ALOT OF CASH to make a GRAPHICS chip at least 3x better then the GEFORCE4 ti 4200!then N wil have a very strong console!
Post by: BlkPaladin on March 07, 2003, 07:40:15 AM
Post by: NickNiteQ93 on March 07, 2003, 01:32:25 PM
Having a portable PS2-type console wouldn't bother me one bit.
Post by: Grey Ninja on March 07, 2003, 02:27:01 PM
Xbox2: we can do 600 billion polys per second!
PS3: we can do 700 billion polys per second!
GCN2: we can realisticaly do about 700 million polys per second
Which console will have the best looking graphics?
Nintendo's of course.
GameCube is far from a budget machine, it's just treated as such because Nintendo tells the truth about their specs.
Post by: Christberg on March 07, 2003, 02:29:49 PM
Quote
Originally posted by: NickNiteQ93
Personally, I hope Nintendo plops down as much, or more, money than Microsoft to ATi for the GPU. I'm sure MS is happy to call their console "the most powerful console available", so it wouldn't suprise me if/when ATi decides to make GPU's for the Xbox Next that they'll make a huge donation to ensure superiority. And even if the next Nintendo system does have powerful specs, it still won't be the most powerful. The Big N, as stated earlier, is money concious, and money concious doesn't equal extreme power. Get your head out of the clouds. Unless Nintendo changes its tune about its money, then I'm looking for the next console from them to be the budget one, as always. But that's fine with me. And if it is the most powerful thing available, then rock on. I'll only buy Nintendo no matter what, so it doesn't matter. Personally, I think ATi should make the GPU for the next Game Boy.
Well, thing is both MS and Sony's next systems are most likely going to be "convergence devices" meaning that they'll have a bunch of extra stuff packed in there that has nothing to do with gaming. Because of this, their systems are going to not only be more expensive to produce, but very likely also less durable than Nintendo's. That said, Nintendo can pack more potent gaming hardware into their next system than the competition because they won't be blowing cash on things like DVD licensing fees, ports for digital cameras and recorders, TV tuners for TiVo-type support, and really large hard drives to store digital video on.
All that stuff is really expensive, especially the big hard drive. Granted, Sony and MS will be willing to lose more money than Nintendo on their next systems, but that doesn't mean that this isn't a golden opportunity for Nintendo to launch at the same time with more powerful hardware if MS and Sony really do intend to do convergence devices, which seems very likely. The chances of Nintendo having the most powerful hardware next generation is looking pretty good to me, maybe 50-50 right now, even despite their cash conscious ways.
Look at what Nintendo did with the GC and think about that level of hardware efficiency versus a convergence device and ask yourself "who wins?"
Getting back on topic though, Nintendo should probably go with IBM again and have them make them a modified Power5 processor with the same instructions Gekko had, and maybe a few more little gaming related tweaks. It makes for easy backward compatibility and also easier development for developers returning to the system.
Post by: )Dark-LInk( on March 07, 2003, 06:17:23 PM
(lets just say 500mhz for gekko since ts SO NEAR)500X10=5000 MHZ(5GHZ) which should be powerful enough to compete hehe
Post by: NickNiteQ93 on March 08, 2003, 11:47:32 AM
Post by: rodtod on March 09, 2003, 07:16:07 PM
'course maybe they went out of business, not sure.
Post by: BlkPaladin on March 09, 2003, 07:27:02 PM
Post by: manunited4eva22 on March 10, 2003, 12:56:55 PM
Post by: Christberg on March 12, 2003, 09:01:05 AM
Quote
Originally posted by: manunited4eva22
Dark link you have to understand that there is more to architectures than wishing something to be 10 times more powerful than it is right now. To gain that much power out of the Gekko (or for that point a hybrid of even the latest powerpc chips, it would require a lot of development time that IBM will want a lot of money for. The cell is taking up a lot of the team that has worked on the powerpc architecture I am sure, so you would be working with an already limited core of developers. IBM will probablly not be the next nintendo CPU producer, PERIOD.
This is totally incorrect. IBM owns several processor design teams. They're finishing work on Cell this year AND launching single and dual core versions of their brand spankin' new Power5 processors which are supposedly 50% faster clock for clock than a Power4 and should be shipping this Fall in time to go head to head with Athlon64. Did I mention IBM also owns Motorola? They can easily design chips for both companies.
Post by: BlkPaladin on March 12, 2003, 09:18:46 AM
Post by: manunited4eva22 on March 12, 2003, 09:59:45 AM
Post by: Christberg on March 12, 2003, 11:38:08 AM
Quote
Originally posted by: manunited4eva22
Thank you for taking my message out of context. He said GEKKO. As in the current powerpc used in gamecube, which is also basically the same used in the CELL. The message above applied to that basic team. As for IBM being a massive company with many different devloping teams, no crap.
You know Gekko is just a PowerPC chip with a few added instructions right? And that it took IBM not more than probably a few months to R&D it because they just added a few special shader based instructions to it? Literally?
You also know that the Power5 processor is basically just a 5th generation redesign of the same chip right?
You also know that CELL has a completely different core type designed from the ground up and has nothing to do with that technology too, right?
Post by: manunited4eva22 on March 12, 2003, 12:21:14 PM
Post by: Christberg on March 12, 2003, 12:51:59 PM
Quote
Originally posted by: manunited4eva22
You also know they are based on different cores right, and that they require different designs right? And that by your logic all the P4 are slight redesigns of each other right?
Which is why Intel can crank out a redesign of the P4 every 4-6 months. It's at it's what? 5th redesign now since it's release? I didn't say it was easy work or anything, but much much easier than you're making it out to be.
All PowerPC chips use a RISC core design.
Post by: MikeHrusecky on March 13, 2003, 08:03:31 AM
Quote
Originally posted by: Christberg
Did I mention IBM also owns Motorola?
Where did you get that? They work together a lot, but Moto is its own company and has its own publicly traded stock. Both companies do, however, have rights to PowerPC.
Moto could theoretically pick up where IBM left off on the Gekko more or less. But then again, who's to say Nintendo needs to continue to work with PPC in the first place?
Post by: Christberg on March 13, 2003, 09:34:50 AM
Quote
Originally posted by: MikeHruseckyQuote
Originally posted by: Christberg
Did I mention IBM also owns Motorola?
Where did you get that? They work together a lot, but Moto is its own company and has its own publicly traded stock. Both companies do, however, have rights to PowerPC.
Moto could theoretically pick up where IBM left off on the Gekko more or less. But then again, who's to say Nintendo needs to continue to work with PPC in the first place?
I could be wrong about that, but I remember reading a press release to that effect ages (as in years) ago. It'd be pretty hard for me to dig up, but at the least I'm pretty damn sure IBM has a very large amount of Motorola stock at the very least.
Nobody ever said Nintendo would go with another PPC core chip this time around. It just seems likely that they'd just stick with the same partners (to me) at least because it'd shorten their R&D cost and time considerably, which is something that would be in their best interests if they're really planning on a 2005 launch. They've already said they're going with ATI again and NEC again, so the only remaining partners are MoSys and ATI.
They could very well go with NEC, IBM, AMD, Transmeta, Sun, or any other company that makes microprocessors if they want to. It's just semi educated speculation that they'll go with IBM again, partially due to market pressure to make the unit backward compatible (the other 2 systems WILL have it) and also because of the fact that they stated they wanted to launch at the same time as the other systems.
Post by: oohhboy on March 14, 2003, 02:30:53 AM
Anyway, the Cell sound like a real bitch to program for as it sounds like a Saturn with more processors, but intergrated on to one chip. It sounds like alot of power, but games beening dynamic programs and all, paralle procesing does not sound that great. I mean it is great if you are doing sequenal thngs like breaking codes, but games having variable one dependant on variable two, procces one dependant on two etc, you end up only using a fraction of the power as parts of the chip end up idle or bottlenecks by other parts.
Post by: manunited4eva22 on March 14, 2003, 09:59:28 AM
Post by: Dirk Temporo on March 16, 2003, 04:56:36 AM
Post by: Unseensoul on March 25, 2003, 08:55:49 AM
Quote
Sat February 22, 2003 2:50 AM
ATI made GCN's processor? O_o That's news to me. I always thought it was IBM...
-Grey Ninja
I thought every GameCube owner knew that ATi was in charge of the graphics... After all there is a logo of ATi on the bottom right-hand side of the GameCube. Unless someone said that ATi made the GC main CPU in previous posts, then, my bad for not catching the sarcasm.
-Unseensoul
Post by: JonTD on March 31, 2003, 07:53:10 PM
In other words, CRAY? Fujitsu? Why!?!?
Quote
)Dark-LInk(:
i think they should ally AN AMERICAN COMPANY(WEIRD) the great NVIDIA! then theyl have the best GRAPHICS for sure! what do ya think?
nVIDIA doesn't really seem to produce the type of GPU Nintendo will be looking for. Sure, Nintendo wants something competitive, but it also needs something cheap to mass-produce and cool. The Gamecube would be more like the Gamecooler if it had a big NV40 under the hood. Sure, nVIDIA could switch gears, but ATi seems to have a good handle on the type of technology Nintendo will want to employ.
Quote
Darc Requiem:
As far as the CPU goes, since Sony has allied with IBM, I think Nintendo should look into Motorola. Motorola and IBM has partnered for the PowerPC standard for Mac's and when it comes to floating point calculations Motorola built PowerPCs with there AltiVec technology actually outperform IBM built PowerPC's.
IBM, Motorola, and Apple all have rights to the PowerPC instruction set. Furthermore, they all have the ability to use AltivVec technology. Be careful: Altivec is simply Apple's name for the SIMD instruction set. All three major players in the now debunk AIM consortium (Motorola, IBM, Apple) can use it anyway they want. IBM is currently producing a POWER4 derivative for the desktop market that actually uses Altivec too.
And if you think x86 chips don't have SIMD units... well... look at 3D Now! and SSE2 people.
Now, moving along, Motorola has been nothing but asstacular in the CPU field for years. It's a company that has, for a very long time, been near to cashing out of the CPU industry entirely. Look at the state of Apple hardware and you'll see exactly why Motorola isn't a good bet for high-powered CPUs. Altivec is the only thing the G4 has going for it. IBM has had much more success at scaling up its own G3 processors, as well as producing PowerPC chips with more bandwidth. A G4s _only_ advantage is when Altivec comes into play, and I imagine the only reason Apple has stuck with the G4 is because they put their eggs into the Altivec basket (for a variety of reasons). Otherwise, they would have jumped ship for higher clockspeed G3s from IBM.
Motorola has still yet to make the jump to a CPU capable of handling the increased bandwidth available with DDR memory. They have struggled with the move to a .13 die process when the rest of the market is gearing up for .09. Do we want them producing a chip for a game console? Motorola's market is small and embedded (as in, embedded chips) where power consumption and heat dissipation are far more serious issues than performance.
Not that Nintendo isn't also very, very concerned about heat and power.
Quote
AMD chips are faster than Intel chips at the same clock speed but they aren't close to the speed of IBM and Motorola's Power PC chips.
Clock for clock? Clock for clock goes out the damn window when Intel is pimping twice as many clocks as its nearest competitor (hypothetical, not literally). What you are talking about is IPC (Instructions Per Cycle), and IBM and Motorola, despite popular belief, are not that far ahead of AMD. Intel isn't in the game, quite simply, because the P4 isn't designed to play the IPC game.. it's designed to play the scaling game. It's all about trade-offs, and being efficient as possible is not always the best way to go.
I mean, I bet even Nintendo could make a manufacture a really efficient 100 MHz processor... doesn't mean its usable in a console. A 2.0 GHz Athlon is going to beat a 1.4 GHz G4 no matter what way you look at it. Of course, in the console industry, really fast CPUs really aren't all that, so something with a high IPC, low heat requirements, low energy needs, and reasonable clock is probably what Nintendo is looking for.
Which will probably keep them in the PowerPC arena, though Intel is doing some interesting things. Look at their Centrino line of mobile chips for a good example.
Quote
- CUBE:
ATi and NEC but not IBM anymore they are traitors it's better to have individual companies work on individual components and focus all their strengths to that like the ATI Flipper and IBM Gekko
Traitors? IBM is a large company... a billion dollar business unto itself. So large, in fact, that it actually competes WITH ITSELF in certain markets. Saying they are traitors is a bit much. They can do well for both Sony and Nintendo without hurting either of them in any way just to advance the other.
Quote
)Dark-LInk(:
they should stick with IBM but pay them MORE millions to make the "GEKKO 2" 10x faster!!(and of course abit stronger then the PS3 "CELL")
ALSO they should get NVIDIA and pay them ALOT OF CASH to make a GRAPHICS chip at least 3x better then the GEFORCE4 ti 4200!then N wil have a very strong console!
And they would also be bankrupt. Being frugal can pay just as well as throwing money hand over fist at your next generation console. Look at some of the advantages of the Gamecube over the XBox. A lot of that is because Nintendo knows quite well that processing power and money does not mean market dominance. Remember? They learned that lesson the hard way.
Nintendo also couldn't afford to get in a money war with Microsoft. They would lose. Microsoft has over $30 billion in the bank. Nintendo is lucky to have 1/30 of that. Microsoft can afford to pay big, spend big, and lose big when Nintendo can't. It's all about the long term, and Microsoft's plan is to entrench themselves now at high cost so they can rake in even higher amounts later.
It's been their strategy all along and has been successful in every market they've ever tried to enter. (MSN, for example, and the browser wars. Not to mention productivity software and the desktop market.)
Quote
rodtod:
what about Silicon Graphics? just a thought...I mean they did revolutionize the digital industry.
Silicon Graphics doesn't have the resources or the will to take on such a project these days. They're struggling to stay alive on Intel hardware and have pretty much given up chip production. They're no longer a contender.
Quote
manunited4eva22:
The cell is taking up a lot of the team that has worked on the powerpc architecture I am sure, so you would be working with an already limited core of developers. IBM will probablly not be the next nintendo CPU
The team that has worked on the PowerPC architecture is still busily working on the PowerPC architecture. IBM is doing some really interesting things, such as ramping up a POWER4 derivative for the desktop market and producing the new POWER5 core that will release processors across several market segments (server, desktop, etc.) For IBM, PowerPC is just getting competitive, and it would make no sense for them to take away from their existing teams to throw developers on Cell.
Making a cheap PowerPC chip for Nintendo, however, would take very little in R&D but bring in a great deal of cash. I wouldn't be at all surprised to see a scaled-down POWER4 or POWER5-based chip in Nintendo's next console. It worked quite well with the G3 after all. Hopefully power and heat issues will not be a problem by that point in time. People who tout the POWER4 often don't realize or just plain forget that it was never intended for the desktop to start with. It has extremely high heat output and power requirements while having a relatively low clock frequency--it is intended for the server space where these issues don't matter. It crunches numbers.
Things are changing though. Also, Nintendo won't need to spend the likes of what Sony is spending. They want their hardware to do very, very different things, and Sony is willing to invest heavily for a specialized chip that does everything it needs it to do. Nintendo no longer has that "everything and the kitchen sink" mentality.
Quote
MikeHrusecky:
Moto could theoretically pick up where IBM left off on the Gekko more or less.
I don't think Motorola would, to be honest. They just don't have the fab in place to produce the type of high performance, high bandwidth chip (relatively speaking to their current offerings) that Nintendo would desire in the large quantities Nintendo would demand. IBM does. They just built a new fab geared towards the .09 process--something to watch.
Quote
Christberg:
I could be wrong about that, but I remember reading a press release to that effect ages (as in years) ago. It'd be pretty hard for me to dig up, but at the least I'm pretty damn sure IBM has a very large amount of Motorola stock at the very least.
If they did they sold it. IBM most definitely does not own Motorola nor does it have much say in how the company is run. Otherwise, Motorola wouldn't be the dog it is now. As a company, the late-90s tech boom just wasn't there for Moto. They bled cash like a cow at a slaughterhouse. They're only now digging themselves out of the mess they were in.
Relations between IBM and Motorola cooled considerably after the AIM consortium fell apart. It seems like relations between Motorola and Apple are heading the same way, as rumor-has-it that Apple is looking to IBM for its high-end chips instead of Motorola's much anticipated but ever-late "G5". It's amazing that the PowerPC instruction set has kept from splintering between the two companies this long, but it's reassuring to see that both IBM and Motorola take efforts to keep PowerPC consolidated in their processors.
Post by: oohhboy on March 31, 2003, 11:42:25 PM
Quote
Nintendo also couldn't afford to get in a money war with Microsoft. They would lose. Microsoft has over $30 billion in the bank. Nintendo is lucky to have 1/30 of that.
Nintendo has closer to 10 bil. Not 1/30 aka, 1 bil. you can't find out the real value as such are corparate secerts. But that is just plain wrong.
Post by: manunited4eva22 on April 01, 2003, 02:17:13 AM
Post by: oohhboy on April 01, 2003, 03:40:09 AM
Post by: JonTD on April 01, 2003, 05:51:58 AM
Quote
oohhboy:
Nintendo has closer to 10 bil. Not 1/30 aka, 1 bil.
If you want to be technical, Nintendo's current assets are valued at around 1,026,478 million yen. That's about $8.6 billion in US dollars. This is according to their end-of-year financial report for 2002. However, that $8.6 billion is not the same as Nintendo's cash reserves, nor is it the same as their cash-on-hand.
Nintendo does not have $10,000,000,000 sitting in the bank. Somewhere between $5 and $10, yes. (I imagine it's still closer to five, but I can't find hard numbers.)
But anyway, it does not matter. I was being sarcastic to make a point. If you want to split hairs, Microsoft has over $40 billion in cash reserves and is looking to increase it by at least another $9 billion by the end of this financial year alone.
Do you think Nintendo can compete on the money front with Microsoft? And Microsoft isn't dumb... they're looking at the growing billion dollar gaming industry just like Sony did, and at how more money comes into Sony from the gaming market than any other market they're involved in. Microsoft wants a piece of the pie, and they WILL get it.
Quote
you can't find out the real value as such are corparate secerts.
No, they're not. Nintendo is a publicly traded company. If they're keeping their current assets a public secret then I believe SEC (and its equivilant in Japan) would be very interested to know why.
Quote
manunited4eva22:
The point was Nintendo doesn't have the resources to spend 5 billion on a single chip, it wouldn't be worth it them...
Exactly.
Quote
oohhboy:
They did spend a billion for gekko. But I believe that was over five years? I also remember it was also for production, a factory and R&D. Regardless, they are willing to spend atleast a billion on a chip. 5 billion for a chip would not be worth it for Microsoft either.
They signed a billion dollar agreement with IBM over Gekko. They did not spend a billion dollars for Gekko R&D. I believe that deal was for design AND manufacture. Remember, Sony-IBM-Toshiba are going to drop $400 million on Cell development _alone_, and that cost will probably rise.
Of course Microsoft isn't going to drop $5 billion on one chip, but do you really think a company who put $2 billion on XBox marketing doesn't have the capability to heavily push funds into chip development just like Sony does? So far they've played it safe with Intel, but... the debate isn't over WHAT they will do as much as what they COULD do.
Anyway, the writing is written all over the walls. Microsoft is willing to spend everything necessary for whatever necessary in order to make themselves a fixture on the gaming scene.
Post by: oohhboy on April 01, 2003, 07:09:22 PM
As far as I can see, there is no point for Nintendo to change from IBM as thier CPU maker as IBM has made a chip which has statisfied all the conditions set out. Low cost, most powerful for the dollar, runs relativly cool, easy to program for. I also don't see why just because Sony has IBM to make a chip for them that another divsion could just do the same and make one for Nintendo.
Post by: JonTD on April 01, 2003, 07:31:59 PM
Quote
Originally posted by: oohhboy
A believe me, total value of a company weather it is in cash reserves or not are a corparate secert as the books are always juggled one way or the other. It just depends on the extent. How do you thing Enron happened? I am not saying that Microsoft or nintendo are screwing over the SEC or equivalent totally, it just does and will happen.
Not every company has skeletons to hide. Innocent until proven guilty.
Quote
As far as I can see, there is no point for Nintendo to change from IBM as thier CPU maker as IBM has made a chip which has statisfied all the conditions set out. Low cost, most powerful for the dollar, runs relativly cool, easy to program for. I also don't see why just because Sony has IBM to make a chip for them that another divsion could just do the same and make one for Nintendo.
Did I ever suggest anything to the contrary? Though most powerful for the dollar is not a true statement (x86 chips are cheaper and faster than PowerPC), and "easy to program for" is really a baseless claim. Easy to program for compared to...? What makes it any easier than an x86 chip, for example? Or a PowerPC derivative from Motorola?
Not that I'm saying it will happen. All hypothetical. I'm interested in your reasoning.
And please don't think I'm making any conjecture about WHAT Nintendo WILL do. I'm just thinking along the lines of what Nintendo's options are. Though I do have a few opinions about that, I really didn't think this thread was appropriate.
Post by: oohhboy on April 01, 2003, 11:34:54 PM
Quote
Did I ever suggest anything to the contrary? Though most powerful for the dollar is not a true statement (x86 chips are cheaper and faster than PowerPC), and "easy to program for" is really a baseless claim. Easy to program for compared to...? What makes it any easier than an x86 chip, for example? Or a PowerPC derivative from Motorola?
Quote
Being frugal can pay just as well as throwing money hand over fist at your next generation console. Look at some of the advantages of the Gamecube over the XBox. A lot of that is because Nintendo knows quite well that processing power and money does not mean market dominance. Remember? They learned that lesson the hard way.
As they say, everything is relative. The current CPU almost matches what Xbox currently offers, even at a lower clock speed. Nintendo got thier most bang for the buck. You could get more bang if you spent more, but you would get less bang per buck. Diminishing returns.
Easy to pogram for, well you know that is relative too. Relative to the current generation would be a good example. Definatly better than the PS2. Developers are willing to say this. Equal to the Xbox? Not quite due to the fact that x86 series is found in 95% of all computers in the world, therefore my guess is that they had alot of pracite with it. Relative to the last generation? Say the N64? hell yes. I can't say for the PSX. Of course all the conditions set out were relative to Nintendo's expectations, not Sony's, not Microsoft's. So the entire argument from both sides are rendered moot.
Hell, the point about motorola, you already answer that(Both have the same core instruction set, inferior fab workshops/tech) and from all the options presented so far I made the statment, not to you, not to anybody in particular. It was just a statment I made that I beleive to be correct from what I know and that I was trying to bring the thread back on topic, which I hope it has.
Post by: JonTD on April 02, 2003, 05:46:34 AM
Quote
Originally posted by: oohhboy
As they say, everything is relative. The current CPU almost matches what Xbox currently offers, even at a lower clock speed. Nintendo got thier most bang for the buck. You could get more bang if you spent more, but you would get less bang per buck. Diminishing returns.
Well, the problem is, do we really know how much Microsoft paid for the CPU inside of the XBox? Do we really know that Gekko "almost" matches? I would think it would be pretty hard to benchmark processors across gaming platforms because it's impossible to have a closed system. Everything from the GPU to the bus to the memory to the CPU are all finely woven together to work in tandum. You don't want the CPU to be starved by the bus, you don't want the memory to go unused by the CPU, you don't want a bottleneck waiting for the GPU, etc. etc.
Is there hard numbers available comparing the Gekko to the P3?
Really, the cloest you can get is a G3 and a P3 on the computing side of things, and a large clock-per-clock performance delta between a G3 and a P3 is just not there. And x86 is generally cheaper, not less expensive, than PowerPC. Though with their current deal Nintendo really didn't see the cost per CPU of a chip... if there ever was a bidding war, AMD or Intel could/would definitely be JUST as cost effective as anything IBM could conjure up.
Quote
Easy to pogram for, well you know that is relative too. Relative to the current generation would be a good example. Definatly better than the PS2. Developers are willing to say this. Equal to the Xbox? Not quite due to the fact that x86 series is found in 95% of all computers in the world, therefore my guess is that they had alot of pracite with it. Relative to the last generation? Say the N64? hell yes. I can't say for the PSX.
Ah. See, I think we've gotten our signals crossed. I really am talking about CPUs, not systems. That's slightly different. I took what you said as anything PowerPC is going to be easier to develop for than say... x86 or MIPS or Alpha (now debunk) or whatever simply because it's PowerPC. Of course, it's very easy to design a PowerPC system that's hard to develop for, and I think you understand that.
Quote
Hell, the point about motorola, you already answer that(Both have the same core instruction set, inferior fab workshops/tech) and from all the options presented so far I made the statment, not to you, not to anybody in particular. It was just a statment I made that I beleive to be correct from what I know and that I was trying to bring the thread back on topic, which I hope it has.
Well, once again, I think we may be talking about slightly different things. A chip from Motorola wouldn't be any hard to develop for, but if the tech wasn't there (a crippled bus and subpar memory system) would be hard to develop for because developers would be forced to tap dance and jazz hands their way around a CPU starved by the long wait for data to come in from other areas (such as memory) down a saturated bus.
However, IF (and that's a big, fat, capitalized IF, as I think you know) Motorola could/would develop a PowerPC chip with a nice, wide bus... well... I doubt it would be any more difficult to develop for.
A complex CPU in and of itself isn't really a development hurdle. It's just that it can be. For example, one really strong reason that the Gamecube is so easy to develop for is Nintendo's partnership with Metrowerks. However, I agree. Nintendo does nothing but benefit by keeping their platform simple while offering some really elegant developing tools to developers. It also does not hurt that Metrowerks has long been creating compilers for PowerPC.
I really think the best thing for Nintendo to do would be to return to ATi and IBM for the GPU and CPU. Even though the Gamecube isn't the most successful platform on the market, it is, without a doubt, exactly what Nintendo needed. They need a repeat... staying out of the set-top-box market and really courting third party developers hard. Listen to them, find out what they want (broadband and modem built in, anyone?), make the platform cheap and easy. If Nintendo can get the Gamecube2 out at the same time as Sony and Microsoft... well... I think they will only gain marketshare. (Especially if they can make development tools even better.)
Though a long, hard, calculated look at the changing demographic of gamers might be in order.
Post by: oohhboy on April 03, 2003, 01:44:32 AM
Quote
As they say, everything is relative. The current CPU almost matches what Xbox currently offers, even at a lower clock speed. Nintendo got thier most bang for the buck. You could get more bang if you spent more, but you would get less bang per buck. Diminishing returns.
I am sorry, this pargraph should have been merged with the nest one as it all relative. Nintendo did get what they wanted for the price they offered. Nobody has hard numbers, but with the two CPU's there is a large difference. RAM. While Xbox uses off the shelf DDR, Nintedo opted for t1-SRAM. In the current set up, the Cube's CPU is going to be far more effecient than that of the Xbox. There is also game based evidence that says there are no reason what Xbox can do that GC can't. The Xbox was suppose to blow away everything else out of the water, yet it has failed to do so. We have yet to see a game that truely sets apart (In terms of graphical quality) the two apart. We have yet to see the insane ploy-count promised. How ever may I try, I cannot explain it as well as an old forum member before Ezboard. His name eludes(Unless you are him, bt he used numbers, not reasoning), but he had the numbers to back up the claim that the GC's CPU almost did match that of Xbox's. Note I am running on the assumption that you know what t1-SRAM is.
Quote
Though with their current deal Nintendo really didn't see the cost per CPU of a chip... if there ever was a bidding war, AMD or Intel could/would definitely be JUST as cost effective as anything IBM could conjure up.
But would it satistfy all the other conditions Nintendo demanded?
Quote
Everything from the GPU to the bus to the memory to the CPU are all finely woven together to work in tandum. You don't want the CPU to be starved by the bus, you don't want the memory to go unused by the CPU, you don't want a bottleneck waiting for the GPU
Quote
A lot of that is because Nintendo knows quite well that processing power and money does not mean market dominance. Remember? They learned that lesson the hard way.
Funny thing to note is that the PS2 has the largest Busbandwidth/fillrate out of all of them by a long shot, but you may know that the lag time from the RAM pretty much negated that advantage along with the lack of any native graphical effects forcing developers to do almost everything in software. There are many other problems, but the PS2 makes for an intersting example. I believe you would agree this is not the way to design a system. You would probaly agree that the Xbox uses brute strenght to overcome it's bottlenecks etc, oppose to the GC's design where effeneicy is the main word. This sets up the argument, which is better, being effecient, or just plain strong.
Quote
I really think the best thing for Nintendo to do would be to return to ATi and IBM for the GPU and CPU. Even though the Gamecube isn't the most successful platform on the market, it is, without a doubt, exactly what Nintendo needed. They need a repeat... staying out of the set-top-box market and really courting third party developers hard. Listen to them, find out what they want (broadband and modem built in, anyone?), make the platform cheap and easy. If Nintendo can get the Gamecube2 out at the same time as Sony and Microsoft... well... I think they will only gain marketshare. (Especially if they can make development tools even better.)
Well said.
I see that we have almost come to an understanding and so far it has been a nice debate.
Post by: JonTD on April 03, 2003, 06:51:26 AM
Quote
Originally posted by: oohhboy
Nobody has hard numbers, but with the two CPU's there is a large difference. RAM. While Xbox uses off the shelf DDR, Nintedo opted for t1-SRAM.
Ah, yes, but that is a limitation of the system itself, not, inherently, the CPU. There's no reason an x86 CPU shouldn't or couldn't use high-speed RAM. Once again I think it's just me thinking in terms of the CPU and you thinking more about each system as a whole. It was definitely a huge boon to Nintendo that the Gamecube is so well-balanced. Hence why it's unfair to say the CPU inside of the Gamecube in and of itself is better than that of the XBox.
Quote
Note I am running on the assumption that you know what t1-SRAM is.
Of course.
Quote
But would it satistfy all the other conditions Nintendo demanded?
Such as...? I don't really see any hurdle that PowerPC overcomes that x86 doesn't. Remember, IBM had to redesign (slightly) the G3 so it could be coupled with the wide bus of the Gamecube. No reason AMD or Intel could not do the same with their CPUs. The Pentium 4, for example, has already been designed from the start to love a wide bus.
POWER4/5 are the same.
Quote
Funny thing to note is that the PS2 has the largest Busbandwidth/fillrate out of all of them by a long shot, but you may know that the lag time from the RAM pretty much negated that advantage along with the lack of any native graphical effects forcing developers to do almost everything in software. There are many other problems, but the PS2 makes for an intersting example.
Ugh. I know. I seriously think Sony wanted to just look good on paper and did not care (or didn't think) how things would work out in real world situations. I'm not really sure, though, that they completely saw both the XBox and Gamecube coming... nor thought they would turn out to be what they are. (Two fairly powerful systems that are easy to develop for, and, compared to the PS2, far more balanced.)
Quote
I believe you would agree this is not the way to design a system. You would probaly agree that the Xbox uses brute strenght to overcome it's bottlenecks etc, oppose to the GC's design where effeneicy is the main word. This sets up the argument, which is better, being effecient, or just plain strong.
I'd argue for a well-balanced, efficient system any day. The XBox could have done better. I really think Microsoft came to the XBox with a PC-centric mindset. Hopefully (or, I guess, unfortunately if you are Nintendo and Sony) Microsoft will rectify this with the XBox2. Consoles are different, and Microsoft had a bit of a learning curve to follow. If anything, the Gamecube has proven that a well-designed, balanced system is more cost-effective than brute strength while garnering similar real world results.
Post by: manunited4eva22 on April 03, 2003, 11:06:32 AM
Another thing to consider would be the MagRAM, but I haven't seen a lot on it for awhile now. Either of you guys have an idea on it?
Post by: oohhboy on April 04, 2003, 03:32:28 AM
Quote
Ah, yes, but that is a limitation of the system itself, not, inherently, the CPU. There's no reason an x86 CPU shouldn't or couldn't use high-speed RAM. Once again I think it's just me thinking in terms of the CPU and you thinking more about each system as a whole. It was definitely a huge boon to Nintendo that the Gamecube is so well-balanced. Hence why it's unfair to say the CPU inside of the Gamecube in and of itself is better than that of the XBox.
"If" both CPU's did use the same RAM then it would be true that the Xbox would have had been more powerful and it was unfair of me to compare other wise. my aplogies.
Quote
Such as...? I don't really see any hurdle that PowerPC overcomes that x86 doesn't. Remember, IBM had to redesign (slightly) the G3 so it could be coupled with the wide bus of the Gamecube. No reason AMD or Intel could not do the same with their CPUs. The Pentium 4, for example, has already been designed from the start to love a wide bus.
What prevented them from picking AMD or intel over IBM? This-
Quote
The IBM processor leverages IBM's experience with complex system designs and incorporates enhancements specifically required by Nintendo. These include functions specifically designed to accelerate games processing and more efficient data bandwidth management between the processor and the game system's primary graphics chip.
Quote
"We are pleased to continue our long-standing affiliation with IBM," said Satoru Iwata, president of Nintendo. "With the ability to push to even higher performance and integration levels
I believe it was all about efficenecy all along. Maybe AMD or Intel could not deliver on that note or it was political. Artical There are more articals, but how relavent they are is up to you. This may provide some more relavent inormation. Another artical Note it mentions the the CPU runs at 405 MHz(now 485MHz). Old artical, but the general facts are still there.
Post by: JonTD on April 04, 2003, 05:50:53 AM
Quote
Originally posted by: oohhboyI believe it was all about efficenecy all along. Maybe AMD or Intel could not deliver on that note or it was political.
Well, it's really a bit of both. You have to keep in mind that the current state of PowerPC and x86 is a little different than what it was a few years ago. At the time, IBM was really doing some things no one else had quite managed yet... they were the first to move to copper, for instance. The IPC of PowerPC was stomping all over Intel, AMD, and Cyrix as well.
The problem is, however, that PowerPC really hasn't lived up to its promise as a desktop contender. Competition between AMD and Intel fueled fierce competition that pushed their chips ahead of Moore's Law. Later revisions of the Athlon and Pentium III (especially) closed the IPC gap between x86 and PowerPC. Today the last stronghold of PowerPC (heat dissipation) is also being addressed in the x86 world... top of the line PowerPC chips are hotter than ever.
So while the x86 manufacturers have stampeded ahead, PowerPC has stagnated. Motorola, IBM, and Apple had a falling out. Motorola's G4 sat at 500 MHz for over a year as the company hit hard times. Their market shifted away from the desktop to the embedded industry as Apple's marketshare slipped further and further behind. IBM waffled on the PowerPC, trying to figure out what exactly it was... an x86 distributor or a PowerPC manufacturer, finally deciding it was both.
Anyway, to make a long story short, the demand necessary to really push the PowerPC ahead in the type of applications that Nintendo would want were not there for a long time. Only now is this starting to look like its changing, and only because IBM has once again decided to start pushing the PowerPC architecture. I'm really glossing over a lot of events to give a quick summary, so don't think x86 is just stomping all over PowerPC or anything. Basically, the moral of the story is that, since AIM fell apart, IBM has focused on the really high-end (servers) and left the mid-range desktop market (and the low-end desktop market is where Nintendo will be shopping around) mostly to Motorola. Keeping up with Motorola isn't exactly difficult...
Two interesting reads: 1 GHz G3 and PowerPC 970. Here's a roadmap, though I'm not particularly fond of IBM's roadmaps myself.
SOI, RapidIO, SIMD on-chip, etc. are all pretty interesting to follow. SIMD, especially, might have some interesting benefits for Nintendo. It's good to remember that the Gamecube2 is two years away... PowerPC970/POWER5 will be coming down the pipe then and the G3 will be scaled even higher. IBM know how is a cut above the rest.
However, as far as IBM's press releases on the Nintendo deal... seems just like a lot of PR to me. PowerPC isn't the end-all-be-all, and I just don't really think it's applicable anymore to call the PowerPC more "efficient." As I said, it's all about trade-offs. I think what will keep Nintendo with IBM is IBM, not just PowerPC.
Post by: JonTD on April 04, 2003, 09:21:38 AM
Quote
Originally posted by: manunited4eva22
I wouldn't mind seeing Nintendo throw a bone out to the DRAM guys. DDR-II is QDR as it is and it takes up a tad less space than the T1-SRAM (both have 1 transistor, though I believe I read that the 1T-SRAM is about 1.2x larger.
I haven't kept up... though the data throughput should/could be better for DDR-II, the big question is latency. How does it stack up to 1T-SRAM? I haven't seen hard numbers (has anyone?) benchmarking 1T-SRAM, but it's obviously very fast... though it is NOT true SRAM, it acts somewhat like it. As far as density goes, I'd imagine if Nintendo stuck with MoSys they would incorporate 1T-SRAM-Q. It's quad density (don't confuse that with quad data rate) and incorporates a lot of other improvements to 1T-SRAM that have come along the way. You can read a brief white paper on it here.
So the size of DDR-II may not be such a boon. It's pretty annoying that MoSys doesn't talk about how high 1T-SRAM clocks, though. Or maybe they have and I missed it. It's a low key player in the memory world.
Quote
Another thing to consider would be the MagRAM, but I haven't seen a lot on it for awhile now. Either of you guys have an idea on it?
It's interesting, but it's really up in the air if it will be ready in time. It won't hit production until 2004, so it's possible... production samples should be available soon. With Nintendo's relationship with IBM... MRAM could be a real possibility. It's just that the word on MRAM has been quiet for a while. Information here and here. Does anyone have anything more recent? The big silence on it could either make it a sleeper surprise or just mean that production is much more difficult than anticipated.
Post by: oohhboy on April 05, 2003, 02:48:58 AM
Quotevolatile
Another thing to consider would be the MagRAM, but I haven't seen a lot on it for awhile now. Either of you guys have an idea on it?
Based on what JonTD has provided (Since I don't generally read white papers for fun), MRAM has one flaw. Capacity. 32MB is not that much and in the next generation memory requirements would be much greater than what MRAM can offer without having muliple units increaing cost. Also volitability and power comsuption would be a non-issue in a console. Based on previous generations along the Nintendo line of machines, memory requirements would jump by 10 to 16 times of what is found today or about 400 to 640 MB, whether that is affordable in 2 years time is unknown. Even if the amount of RAM contained only increased by 3,4 or 5 times of today, MRAM would have trouble filling it. Of course, 400 to 640 MB of RAM is some what over kill. some where around 256MB would probaly work out with todays current displays as we come to the limit of what the human eye can see on them.
MRAM though would work wounders for the next Gameboy. But FeRAM would prove the logical choice for cost, size, capacity and latency(this being a non-issue).
Post by: rodtod on April 08, 2003, 08:11:59 PM
Post by: JonTD on April 08, 2003, 09:24:04 PM
Quote
Originally posted by: rodtod (emphasis mine)
oohhboy, one of memory's main uses is as a temporary hard drive, so to speak. If Nintendo isn't going to use a hard drive in their next console, then they should plan to include a whole bunch of meg's of memory.
No it's not. If you'd like to clarify...?
If anything, it's the other way around.
Post by: manunited4eva22 on April 09, 2003, 11:37:01 AM
Post by: JonTD on April 10, 2003, 01:58:48 AM
Quote
Originally posted by: manunited4eva22
If you are referring to virtual memory, it is worthless for real time use.
No, it is not. And it wouldn't be virtual memory, per se. Accessing data from the hard drive is much, much faster than reading data from optical media. There are lots of things developers can utilize a hard drive for in their games. I wouldn't be at all surprised if game developers are already utilizing some of the things I'm thinking about in my head on the XBox. (Matter-of-fact, I'd be surprised if they didn't.)
At any rate, I (I'm assuming your comment is directed at me) am NOT talking about virtual memory. I'm just stating that a hard drive is not memory, memory is not a hard drive, and they're not the same in any sense of the word. (Except, of course, for virtual memory systems.) Games use memory to hold data before it is processed by the CPU or GPU. Once that data is used and no longer necessary, you want to flush it from memory as soon as possible to free up space.
A hard drive's use stems from the fact that it does not require constant power to maintain data. You cut the power and its still there. When you run a program it is loaded from the hard drive into memory. When parts of memory go unused, it is cached out to the hard drive. When a program is closed (and to close it, you would have to reload the program back into memory from the hard drive if it were cached out) some memory is then freed up for other processes. Some of the data remains in memory for a while... in case you reload the program... depending on which operating system you use. (You can test this on MacOS X. Load a program after a cold shut down and reboot. Count the bounces and then start the application again (just hope it doesn't load in one to start for this test
). Chances are you will notice the much slower run time. You can do the same with Windows... cold start, load something from quick launch, remove it from quick launch, restart, load the application again.)I can't see how on current consoles this behavior can be replicated in any sense with any form of RAM. Which is why I disagree with the statement, "one of memory's main uses is as a temporary hard drive, so to speak." However, rodtod did say "so to speak" which is why I'm asking for a more thorough explanation... I don't currently follow his reasoning, but I'm willing to listen to what he is saying in order to clear up any misunderstandings.
Post by: rodtod on May 05, 2003, 04:14:30 PM
crap, I meant memory cards btw.
