Broadway specs

[Shifty Geezer]

These things aid efficiency. They allow you to either run a processor cooler with less power, or clock it higher, or mixes of the two. They don't fetch you better performance unless you use them. The performance is roughly determined by the number and type of functional units in the CPU, and the clock speed. The rest mostly just determines yields and costs.

 

[cubeboy101]

These things aid efficiency. They allow you to either run a processor cooler with less power, or clock it higher, or mixes of the two. They don't fetch you better performance unless you use them. The performance is roughly determined by the number and type of functional units in the CPU, and the clock speed. The rest mostly just determines yields and costs.

im sure you wil find your wrong, if i got a broadway made of aluminium connectors and no other upgrades, like silicon on insulator then benched it directly with boadway with copperwire silicon on insulator the copper and silicon upgraded chip would out perform the aluminium cpu by a good amount...
efficancy is power isnt it. copperwire beats aluminium clock for clock even ibms official press release states using silicon on insulator allows us to lower power consumtion and heat and at the same time offer a SIGNIVICANT upgrade in power, your seriously sujesting only clockspeeds matters...

its proven silicon on insulator gives a cpu core a 70% increase in data flow and a average per chip performance increase of 35%

id personally rate broadway at about 3 times the xbox celeron it has massive advantages over a old cisc aluminium chip...

broadways advantages
risc
out of order true cpu prossesing
coperwire
stained silicon
silicon on insulator
gamecentric optermized

dont underestemate this cpu

 

[StefanS]

You're making no sense, here.

Manufactoring processes is not processing power. SOI, etc. don't automatically gurantuee better processor performance.

 

[swaaye]

dont underestemate this cpu

heh. Don't overestimate this CPU. It's just a 90 nm Gecko (slightly customized G3) from what we've been able to figure out. Nintendo had their already R&D'ed CPU shrunk down to reduce power, but mainly to make it super cheap no doubt, and didn't even double the CPU clock.

 

[Teasy]

heh. Don't overestimate this CPU. It's just a 90 nm Gecko (slightly customized G3) from what we've been able to figure out. Nintendo had their already R&D'ed CPU shrunk down to reduce power, but mainly to make it super cheap no doubt, and didn't even double the CPU clock.

We haven't been able to figure out anything that specific yet so I wouldn't make those assumptions.. But it is pretty obvious that at best the CPU is around twice as fast as Gekko (maybe a bit more, maybe a bit less).

Though if Hollywood has a vertex shader the difference between Gekko and Broadway could be much more pronounced in game (3 to 4 times faster). Here's hoping...

 

[OtakingGX]

im sure you wil find your wrong, if i got a broadway made of aluminium connectors and no other upgrades, like silicon on insulator then benched it directly with boadway with copperwire silicon on insulator the copper and silicon upgraded chip would out perform the aluminium cpu by a good amount...
efficancy is power isnt it. copperwire beats aluminium clock for clock even ibms official press release states using silicon on insulator allows us to lower power consumtion and heat and at the same time offer a SIGNIVICANT upgrade in power, your seriously sujesting only clockspeeds matters...

its proven silicon on insulator gives a cpu core a 70% increase in data flow and a average per chip performance increase of 35%

id personally rate broadway at about 3 times the xbox celeron it has massive advantages over a old cisc aluminium chip...

broadways advantages
risc
out of order true cpu prossesing
coperwire
stained silicon
silicon on insulator
gamecentric optermized

dont underestemate this cpu

What? Are you just making this crap up? If you were to take two processors with the same clock, same architecture, and they only differed in aluminum and copper interconnects you'd find they would have the exact same performance. Case in point, AMD's Athlon Thunderbird core was produced in both copper and aluminum interconnect variants. Each chip performed identically at the same speed, while it was claimed that the copper chips overclocked better. You'll find the numbers here to show the differences are statistically insignificant.

I don't doubt that the 750CL (or whatever Broadway truly is) is much better suited for gaming than the castrated PIII in the Xbox, but not for the reasons listed. Copper, strained (and possibly stained) silicon, and SOI all helped IBM achieve a design on a process node at a speed that conventional techniques probably would not have allowed, and within a tiny power budget. Would a chip with such a short pipeline have been possible at 90nm without such technologies? Maybe an electrical engineer can help us out here.

 

[crystall]

Would a chip with such a short pipeline have been possible at 90nm without such technologies? Maybe an electrical engineer can help us out here.

It could have been possible at the same performance level IMHO, though power would have probably been significantly higher. What's excellent about it is its very low power consumption and probably - from Nintendo POV - its price.

 

[MDX]

Does anyone know why CANADA is stampled on the Broadway?
http://www.informit.com/articles/article.asp?p=677908&rl=1

IBM is still a US company correct?
Broadway was manufactured in the US as well, correct (Washington)

Im wondering if this is due any influence by AMD/ATI? If so, did they design the CPU?

 

[swaaye]

IBM is definitely multinational.

 

[MDX]

Does anyone know why CANADA is stampled on the Broadway?
http://www.informit.com/articles/article.asp?p=677908&rl=1

IBM is still a US company correct?
Broadway was manufactured in the US as well, correct (Washington)

Im wondering if this is due any influence by AMD/ATI? If so, did they design the CPU?

In answer to my own question, some member on another forum said that the first
run of CPUs came from

"Canada IBM manufacturing plant... Bromont, Quebec to be exact..."

Now the rest are being done stateside.

 

[Flux]

What? Are you just making this crap up? If you were to take two processors with the same clock, same architecture, and they only differed in aluminum and copper interconnects you'd find they would have the exact same performance. Case in point, AMD's Athlon Thunderbird core was produced in both copper and aluminum interconnect variants. Each chip performed identically at the same speed, while it was claimed that the copper chips overclocked better. You'll find the numbers here to show the differences are statistically insignificant.

I don't doubt that the 750CL (or whatever Broadway truly is) is much better suited for gaming than the castrated PIII in the Xbox, but not for the reasons listed. Copper, strained (and possibly stained) silicon, and SOI all helped IBM achieve a design on a process node at a speed that conventional techniques probably would not have allowed, and within a tiny power budget. Would a chip with such a short pipeline have been possible at 90nm without such technologies? Maybe an electrical engineer can help us out here.

so we still don't know the hardware specifications for broadway other than the rumored specifications?

 

[MDX]

After doing benchmark tests on both the Wii and PS3's CPU it was found that the PS3's Cell processor which is over 3GHz was only about equal to a 1.5 GHz CPU. While the Wii's Broadway chip was about 8oo MHz. If benchmarking is to be believed, then the Cell processor is less than twice as fast as the Broadway. More tests should be made but this is a interesting find no doubt.
video:
http://video.google.com/videoplay?docid=-5136252806353451439&q=felix+domke

source:
http://revoeyes.blogspot.com/2007/01/ps3-cell-processor-less-than-2-times-as.html

 

[INKster]

After doing benchmark tests on both the Wii and PS3's CPU it was found that the PS3's Cell processor which is over 3GHz was only about equal to a 1.5 GHz CPU. While the Wii's Broadway chip was about 8oo MHz. If benchmarking is to be believed, then the Cell processor is less than twice as fast as the Broadway. More tests should be made but this is a interesting find no doubt.
video:
http://video.google.com/videoplay?docid=-5136252806353451439&q=felix+domke

source:
http://revoeyes.blogspot.com/2007/01/ps3-cell-processor-less-than-2-times-as.html

Any real hacker would know instantly that benchmarking -especially on a console with wildly different CPU designs and power planes- based on absolute clockspeed numbers is completely wrong.
I sure would like to know how he "hacked" the PS3, but somehow i doubt he will ever tell us... because it probably didn't happen.
Just some guy trying to make it look like the Nintendo Wii is even remotely comparable to either the PS3 or the X360 in terms of hardware processing power.

 

[3dilettante]

It might be somewhat in the ballpark if the guy just looked at one thread on the PPE.

If you extrapolate from this probably fake or horribly flawed data, that makes Cell worth about 8 double-speed Broadways with software that was developed for what it's running on.

 

[JohnyBeGood]

He says that this "benchmark" measures general purpose processing capabilities and not vector processing.

 

[3dilettante]

He says that this "benchmark" measures general purpose processing capabilities and not vector processing.

Make that maybe 5 double-speed Broadways, then.

I can't follow the links right now, did he mention how threaded things were, or what excactly the benchmark did?

 

[JohnyBeGood]

Make that maybe 5 double-speed Broadways, then.

I can't follow the links right now, did he mention how threaded things were, or what excactly the benchmark did?

His Wii vs PS3 comparison is rather simplistic and humorous , favoring Wii. He doesn't make clear how the cpus performance was measured or ,in case cell's case,if the results are only about PPE's performance. He goes up to saying that PS3:
1) is more difficult to carry with one hand (compared to Wii) , due to it's weight , thus reducing it's portabillity,
2) it is more power consuming ( 380 W compared to Wii's 25 W ) and
3) 2,4 times more expensive.
He ends up comparing the FUN factor with PS3 gaining an unipressive 24 mark whereas Wii gets a mighty 212 , which means that Wii is 8,83 times more fun to play (don't ask what benchmark he used for that).
The whole thinq looks like a joke ,imo.

 

[3dilettante]

His Wii vs PS3 comparison is rather simplistic and humorous , favoring Wii. He doesn't make clear how the cpus performance was measured or ,in case cell's case,if the results are only about PPE's performance. He goes up to saying that PS3:
1) is more difficult to carry with one hand (compared to Wii) , due to it's weight , thus reducing it's portabillity,
2) it is more power consuming ( 380 W compared to Wii's 25 W ) and
3) 2,4 times more expensive.
He ends up comparing the FUN factor with PS3 gaining an unipressive 24 mark whereas Wii gets a mighty 212 , which means that Wii is 8,83 times more fun to play (don't ask what benchmark he used for that).
The whole thinq looks like a joke ,imo.

I see he left out the name normalization factor: where anyone with console named "Wii" must quarter the size of his e-penis.

This would reduce Wii to little more than double the PS3's value.

 

[Teasy]

Make that maybe 5 double-speed Broadways, then.

Nah I'd say 3.75 double speed Broadways

 

[TurnDragoZeroV2G]

What kind of Double-Speed Broadway rating would the Core 2 Duo at 1.8GHz get?

I bet it's higher than Xenon's DSB rating!