Wii broadway update

The details of the Nintendo chip, code-named Broadway, are being kept under wraps.

"The first chips are in our possession," said Genyo Takeda, senior managing director for Nintendo's Integrated Research and Development Division. "Today's milestone marks the final stage of our drive to reach both core and nontraditional gamers with an inviting, inclusive and remarkable gaming experience."

IBM only says the chip is 20 percent more power-efficient than the first game chip it began making for Nintendo in 1999, which was called Gekko.


hmmm


edit: ah here we go, this is more like it

http://today.reuters.com/news/artic...IDST_0_TECH-IBM-NINTENDO.XML&rpc=66&type=qcna


IBM said the chips use technology that will substantially
boost the processing power in the Wii while using 20 percent
less energy and throwing off less heat.
 
Last edited by a moderator:
IBM said the chips use technology that will substantially boost the processing power in the Wii while using 20 percent less energy and throwing off less heat.
Yes, the technology dubbed "90nm process with higher operating frequency" that boosts processing power by nearly 100% over existing GameCube CPU, while keeping the power usage low. :cool:
 
Yes, the technology dubbed "90nm process with higher operating frequency" that boosts processing power by nearly 100% over existing GameCube CPU, while keeping the power usage low. :cool:

Wow! Never heard of that teachnology. Do you know if it will be available in enough time for the Wii launch or could we be looking at a delay for such cutting edge technology? :p
 
Yes, the technology dubbed "90nm process with higher operating frequency" that boosts processing power by nearly 100% over existing GameCube CPU, while keeping the power usage low. :cool:

Well, they had bolt on a power-save mode as well, as Wii supposed to be always powered-on , according to Nintendo.
 
Yes, the technology dubbed "90nm process with higher operating frequency" that boosts processing power by nearly 100% over existing GameCube CPU, while keeping the power usage low. :cool:

Since when does a 50% higher operating frequency boost processing power by nearly 100%? :LOL:
 
Quantifying "power efficiency" for something as complex as a general purpose CPU is pretty mean. Bad IBM! Sit! No cookies!
 
Teasy said:
Since when does a 50% higher operating frequency boost processing power by nearly 100%?
Actually I thought I remembered multiplier was more then 1.5 - hence I was just "rounding up" like any good marketting person would :p
 
Well, they had bolt on a power-save mode as well, as Wii supposed to be always powered-on , according to Nintendo.

I believe gecko was a sub 5-watt processor, even at 700+ mhz, a similar processor idling would use even less. (but 5 watts always on would add up to more money over a year than a more powerful system only played some of the time)

Oh, Wii's processor also supports SOI technology. Considering how late it's being produced though, I think it's really disappointing that such a simple processor didn't make 65nm SOI.
 
http://forum.teamxbox.com/showthread.php?t=468997

There is a video from IBM fabs doing the CPU.

Plus a nice theory (well if the chip is indeed 25mm^ at 90nm then it should have quite a few more tech wonder the wood, unless they go for 1Mg of L2 but then the CPU is like Gekko and what is the point of so much cache(?)).

Gecko has a 256KB L2 Cache
Broadway has 512KB L2Cahce
GX has 1mB of L2 Chache

The other minor note is that is that L1 cache of Broadway is equal to the L2 Cache of gecko. Hopefully Matt at Ign reveals some the other gecko tidbits, because at least one is a really big clue to nintendo's future and this cpu when you connect the dots.
 
I'm not speculating on the size I'm saying as fact that's what it is.

Sorry, I think you're probably confused. Maybe you're thinking of the D-Cache size in Kbits. There's simply no way IBM would implement such a large L1 cache (larger than any other PPC out there) -- particularly given the dubious performance gains that would come from such a large and likely slow L1.
 
My bad on the number size Though increase by multplier 4 or 2x in most areas is in effect. From what I know as a concrete the L2 512K, while the L1 is 256K 1st report or 128K second report.
 
Well if it draws 20% less power, while "substantially boosting processing power", it'll obviously be more then 20% more "power efficient" than Gekko. Besides, they'll probably have the most brainiac cpu this gen, which, if your propa...aeh marketing department isn't all gung ho about silly spec wars, is a good thing, imho.
 
Well if it draws 20% less power, while "substantially boosting processing power", it'll obviously be more then 20% more "power efficient" than Gekko. Besides, they'll probably have the most brainiac cpu this gen, which, if your propa...aeh marketing department isn't all gung ho about silly spec wars, is a good thing, imho.

Gecko wasn't that brainy. Anyhow, at 700-800mhz, I doubt broadway is going to performance competitive in anything with xenon or cell. From what's been done on xbox 360 so far, I'd estimate each core to be about the performance of a 2ghz to 2.4ghz pentium 4 (or at least that's the currently realized performance, it could do better as games become better optimized) though I'm also assuming current games are only single threaded. Most people would probably consider that estimation on the low side though, but it's way beyond what a sub 1ghz g3 should be capable of.
 
Gecko wasn't that brainy. Anyhow, at 700-800mhz, I doubt broadway is going to performance competitive in anything with xenon or cell. From what's been done on xbox 360 so far, I'd estimate each core to be about the performance of a 2ghz to 2.4ghz pentium 4 (or at least that's the currently realized performance, it could do better as games become better optimized) though I'm also assuming current games are only single threaded. Most people would probably consider that estimation on the low side though, but it's way beyond what a sub 1ghz g3 should be capable of.

With much higher code density and instruction decoding out of the critical path, much more sophisticated branch prediction & cache architecture & elaborate prefetching, a robust OOOE implementation and Integer ALUs at 4 and 4.8 GHz respectively, a single Xenon core would surpass it only on very,very biased workloads; but given a fair and representive set of real world problems and equally elaborate implementations on both architectures, the P4 would pull far ahead on average.
 
Back
Top