A week ago, in the series of articles at PC Watch, Hiroshige Goto had this article titled "The Rushed Xbox 360 And The Lost Xbox 1.5 Plan."
http://pc.watch.impress.co.jp/docs/2005/1227/kaigai231.htm
Basically it's a technical postmortem of the Xbox 360 launch, but he reveals some new info he got from insiders, and here's my summary.
1. The lost Xbox 1.5 plan
Microsoft had actually planned to release an updated version of Xbox, called Xbox 1.5, shortly after the original Xbox 1, and before the real next-generation Xbox. This Xbox 1.5 has the full hardware backward compatibility with Xbox 1, and has a better CPU/GPU. It can run both Xbox 1 softwares and better looking Xbox 1.5 softwares. Basically it breaks the 5-year cycle of game console and runs away from the competition by the absolute power. Goto heard of this Xbox 1.5 from various sources, and one of them was people in the CPU industry not at all related to the games industry. Goto suggests it's likely that MS was seeking a CPU supplier.
Goto also writes back then he was confused by the info that what's thought to be Xbox 1.5 would be the next-gen Xbox 2 rather than a mere upgraded version, so now he seem to have another evidence that what's called Xbox 1.5 and the actual Xbox 360 are 2 different things. But he doesn't know the reason why Xbox 1.5 was ditched and speculates because of this fluctuation Microsoft couldn't get enough design time for Xbox 360.
IBM disclosed at FPF 2005 that the Xbox 360 CPU development including silicon manufacturing only took 24 months, which is unusually short for CPU developement and almost rivals with a GPU development schedule. According to Goto it means Microsoft had a basic design database for the Xbox 360 CPU before going to IBM.
2. Too optimistic heat estimation
Goto thinks while it seems the PS3 case was designed after Cell was completed, the design of the Xbox 360 case had to be done before silicon validation. At E3, he heard from a Microsoft insider that they hesitated to make the PSU external, as seen in the back panel of the mockup unit at E3. The ugly PSU will stay for a while as the symbol of the rushed design until the chip set is shrinked and the PSU becomes internal.
3. The 2-core plan for the Xbox 360 CPU
He heard from game developers that devkits based on the actual Xbox 360 design had various problems such as units under the target clockspeed, and other problems not related to the clockspeed. Considering the transistor size and the frequency, the biggest heat source is the 3.2Ghz 3-core CPU. According to what Goto heard from a Microsoft insider, there was actually a plan about a 2-core version and they tried to lower the clockspeed too. But the Microsoft insider told him that eventually there had been a reason that it has to be both 3-core and 3.2Ghz. Goto doesn't know the exact reason, but he speculates it's to run games and other services simultaneously.
-------------------------------------------------------------------------------------------------------
Anyway, I've never heard of this Xbox 1.5 plan, so posted this here to ask if someone knows about this plan or how do you think about this.
http://pc.watch.impress.co.jp/docs/2005/1227/kaigai231.htm
Basically it's a technical postmortem of the Xbox 360 launch, but he reveals some new info he got from insiders, and here's my summary.
1. The lost Xbox 1.5 plan
Microsoft had actually planned to release an updated version of Xbox, called Xbox 1.5, shortly after the original Xbox 1, and before the real next-generation Xbox. This Xbox 1.5 has the full hardware backward compatibility with Xbox 1, and has a better CPU/GPU. It can run both Xbox 1 softwares and better looking Xbox 1.5 softwares. Basically it breaks the 5-year cycle of game console and runs away from the competition by the absolute power. Goto heard of this Xbox 1.5 from various sources, and one of them was people in the CPU industry not at all related to the games industry. Goto suggests it's likely that MS was seeking a CPU supplier.
Goto also writes back then he was confused by the info that what's thought to be Xbox 1.5 would be the next-gen Xbox 2 rather than a mere upgraded version, so now he seem to have another evidence that what's called Xbox 1.5 and the actual Xbox 360 are 2 different things. But he doesn't know the reason why Xbox 1.5 was ditched and speculates because of this fluctuation Microsoft couldn't get enough design time for Xbox 360.
IBM disclosed at FPF 2005 that the Xbox 360 CPU development including silicon manufacturing only took 24 months, which is unusually short for CPU developement and almost rivals with a GPU development schedule. According to Goto it means Microsoft had a basic design database for the Xbox 360 CPU before going to IBM.
2. Too optimistic heat estimation
Goto thinks while it seems the PS3 case was designed after Cell was completed, the design of the Xbox 360 case had to be done before silicon validation. At E3, he heard from a Microsoft insider that they hesitated to make the PSU external, as seen in the back panel of the mockup unit at E3. The ugly PSU will stay for a while as the symbol of the rushed design until the chip set is shrinked and the PSU becomes internal.
3. The 2-core plan for the Xbox 360 CPU
He heard from game developers that devkits based on the actual Xbox 360 design had various problems such as units under the target clockspeed, and other problems not related to the clockspeed. Considering the transistor size and the frequency, the biggest heat source is the 3.2Ghz 3-core CPU. According to what Goto heard from a Microsoft insider, there was actually a plan about a 2-core version and they tried to lower the clockspeed too. But the Microsoft insider told him that eventually there had been a reason that it has to be both 3-core and 3.2Ghz. Goto doesn't know the exact reason, but he speculates it's to run games and other services simultaneously.
-------------------------------------------------------------------------------------------------------
Anyway, I've never heard of this Xbox 1.5 plan, so posted this here to ask if someone knows about this plan or how do you think about this.
That seems like it. Especially the paragraph at the end of the windowsitpro article echoes what Goto wrote in the article...