PS4 to use Cell, NOT PS3?

[Panajev2001a]

The point behind Cell is that a Software Cell/Apulet can be run by any APU of any Cell based device ( with the exception of the "APUs requested field" in the Apulet header which requires a certain number of free APUs to be executed ) if using a MIPS64 core instead of a PPC core in a PU ( both of them would not be full size cores... a PU should not be a huge processor ) does not compromise this aspect ( easy inter-operability between Cell based devices without the need of complex Virtual Machines and complex Abstraction Layers... ) then it would be fine...

In the case that changing the PU core involves the come bakc of huge and complex abstarction layers for compatibility then you would not use a different PU core compared to the standard one used in the Cell Processors built by IBM and Toshiba...

The patent does not really specify the core used in the PU...

 

[Paul]

http://www.scei.co.jp/corporate/release/pdf/030421ae.pdf

"With this investment, SCEI will manufacture the new microprocessor for the broadband era, code-named “Cellâ€￾, as well as other system LSIs, to be used for the next generation computer entertainment system."

 

[jvd]

Man posting old quotes is getting boring esp since we have gone through them all time and time again and ever article points both ways. Lets just wait and see what happens . For the love of god they haven't even started building the fabs yet .

 

[Paul]

bbot asked for the quote jvd. And it's not "vauge" or it can't go both ways.. It specificly states that Cell, as well as other LSI's will be used in the next computer entertainment system.

 

[jvd]

bbot asked for the quote jvd. And it's not "vauge" or it can't go both ways.. It specificly states that Cell, as well as other LSI's will be used in the next computer entertainment system.

I said every article . Read what i posted. There is also no reason to bring up something that is being discused in another thread. Why don't u just send him to that thread instead of having two splinter arguments going on. Also we both know there are quotes going against the cell chip being in the ps3. There are also quotes saying the construction on the fabs used to make this chips will start next fiscal year. I've also stated that it can be the other Lsi's that are in the ps3 not the cell chip.

 

[Paul]

You said ever article. Also, why send him places where we know he won't read it?

"I've also stated that it can be the other Lsi's that are in the ps3 not the cell chip."

....I take it you didn't read the quote or just don't want to read it.

Cell AS WELL AS other system LSI's TO BE USED in the next generation comp entertainment system.

If your having trouble reading it than it's your problem, now where is that post where that guy had the grammar discusion about this to you.

 

[jvd]

You said ever article. Also, why send him places where we know he won't read it?

"I've also stated that it can be the other Lsi's that are in the ps3 not the cell chip."

....I take it you didn't read the quote or just don't want to read it.

Cell AS WELL AS other system LSI's TO BE USED in the next generation comp entertainment system.

If your having trouble reading it than it's your problem, now where is that post where that guy had the grammar discusion about this to you.

Oh ya i yield that point . Okay . Go to that thread and see where we are at though . Before the rumors of it pushed back we still don't know 100% sure it will be used although we all feel it will. Production still hasn't started yet on the plants to build the cell chip. There will be a pstwo before the ps3 comes out. Jvd is awsome .

 

[Paul]

"Before the rumors of it pushed back we still don't know 100% sure it will be used although we all feel it will."

So If I make a up a rumour that Sony will use a CPU based on bio-tech that will be 10tflops we shouldn't be 100% sure of Cell being in PS3? Despite ALL the articles, and despite SCEI confirming it? Don't give me that BS, "oh other LSI's mean another CPU"

Because that's NOT what it says, your reading it wrong, or only reading what you want to hear. Also, show me the yearly reports which show that a second CPU has funding, oh wait it's not in any of them!

Sony is spending a few billion dollars on a CPU that will not go inside of their best selling product? Despite them saying it will numerous times, and SCEI confirming it on Monday.

"Production still hasn't started yet on the plants to build the cell chip."

Point being? gamecube 2's chip isn't in production either yet, neither is MS's for the xbox2.

Cell is schedueled to be ready 2005, production starts late 2004 or early 2005, and it should be ready for a late 2005 japan launch.

 

[jvd]

"Before the rumors of it pushed back we still don't know 100% sure it will be used although we all feel it will."

So If I make a up a rumour that Sony will use a CPU based on bio-tech that will be 10tflops we shouldn't be 100% sure of Cell being in PS3? Despite ALL the articles, and despite SCEI confirming it? Don't give me that BS, "oh other LSI's mean another CPU"

Because that's NOT what it says, your reading it wrong, or only reading what you want to hear. Also, show me the yearly reports which show that a second CPU has funding, oh wait it's not in any of them!

Sony is spending a few billion dollars on a CPU that will not go inside of their best selling product? Despite them saying it will numerous times, and SCEI confirming it on Monday.

"Production still hasn't started yet on the plants to build the cell chip."

Point being? gamecube 2's chip isn't in production either yet, neither is MS's for the xbox2.

Cell is schedueled to be ready 2005, production starts late 2004 or early 2005, and it should be ready for a late 2005 japan launch.

How about the quotes of the pres of sony saying cell wont be used. As i have said a thousand times nothing is offical and untill they offical come out with specs for the ps3 anything is up in the air. Why don't u go read the other thread again and see where we have gotten. Your right gamecube chips are not inproduction. of course nintendo isn't trying to push the envolope and will go with less risky tech. Ms ... well I'm pretty sure the cpu is in production now since they will most likely use a p4 or a hammer based chip. Of course ms and nintendo aren't make any anouncements and shouldn't be dragged into this . when they do then we can arguee about thier tech. The fabs to be used for the cell chips wont even start being built untill the start of fiscal year 2004 and that is if there is no delays .

 

[Panajev2001a]

There are also quotes saying the construction on the fabs used to make this chips will start next fiscal year. I've also stated that it can be the other Lsi's that are in the ps3 not the cell chip.

There are fabs that Sony and Toshiba have ( and IBM will have 65 nm technology as well in 2005 ) that can be upgraded and the fab that will be started next fiscal year should be the Nagasaki one and not the new one in Oita.

Also in fiscal 2004, even before the fab is finished, 25,000+ wafers a month should be a good output for the NEW Oita fab using 65 nm technology and getting ready for 45 nm.

Also Sony and Toshiba affirmed in several occasions that they feel positive about a March-April 2004 start for 65 nm manufacturing...

Toshiba presented 65nm-node devices at IEDM this week (Dec. 9-11) and intends to establish volume production technology by 2004 spring.

http://www.eetimes.com/semi/news/OEG20021213S0016

TOKYO — Toshiba Corp. said it is leveraging its deep-trench-capacitor DRAM structure, which it has championed as a process driver since the 0.25-micron generation, to migrate its system-on-chip devices to the 90-nanometer and 65-nm process nodes ahead of its competitors.

Process development partner Sony Corp., meanwhile, is formulating plans for production at the two next-generation nodes and said it will migrate to the trench-capacitor structure from a stacked-capacitor architecture developed with Fujitsu.

Toshiba has been shipping engineering samples of 90-nm embedded-DRAM devices since November and expects to begin volume production in March with a startup capacity of 1 million units per month. Small-scale production of devices at the 65-nm node is expected a year later, in March 2004.

http://www.eetimes.com/story/OEG20030117S0029


Again, I do not expect PlayStation 3 to launch in Japan before March-April ( or even September ) 2005 and by early-mid December 2005 in the U.S.

Sony and Toshiba have quite a lot of fabs already and the new fabs they are building will be able to produce chips even if not at full speed before PlayStation 3 is launched... plus you also have IBM which could help if the situation calls for it.

 

[Paul]

How about the RECENT quotes of SCEI, and countless other articles saying it will be? The designchain article DIDNT even quote kutagari himself.

At a joint press conference held in Tokyo today, Sony Corp. and Toshiba Corp. announced plans to build advanced semiconductor plants for the fabrication of faster and more complex chips as they look ahead to the next-generation PlayStation console and other broadband network products. Among the announcements include confirmation that the Cell processor will be the main engine behind the PlayStation 3.

So your telling me Sony lied again at the press conference?

"The fabs to be used for the cell chips wont even start being built untill the start of fiscal year 2004 and that is if there is no delays."

Oh really?

"The Sony line will be constructed at the company's Nagasaki plant, while the Toshiba line will be at that company's Oita facility, both in Japan. Construction is planned to begin this June."

http://www.gamepro.com/index.html?/gamepro/famitsu/games/news/29018.shtml

So the Toshiba fab begins Cell production early in the game to make Cell chips for the immediate ps3 launch, than later when the Sony fab's are ready they take over.

 

[jvd]

There are also quotes saying the construction on the fabs used to make this chips will start next fiscal year. I've also stated that it can be the other Lsi's that are in the ps3 not the cell chip.

There are fabs that Sony and Toshiba have ( and IBM will have 65 nm technology as well in 2005 ) that can be upgraded and the fab that will be started next fiscal year should be the Nagasaki one and not the new one in Oita.

Also in fiscal 2004, even before the fab is finished, 25,000+ wafers a month should be a good output for the NEW Oita fab using 65 nm technology and getting ready for 45 nm.

Also Sony and Toshiba affirmed in several occasions that they feel positive about a March-April 2004 start for 65 nm manufacturing...

Toshiba presented 65nm-node devices at IEDM this week (Dec. 9-11) and intends to establish volume production technology by 2004 spring.

http://www.eetimes.com/semi/news/OEG20021213S0016

TOKYO — Toshiba Corp. said it is leveraging its deep-trench-capacitor DRAM structure, which it has championed as a process driver since the 0.25-micron generation, to migrate its system-on-chip devices to the 90-nanometer and 65-nm process nodes ahead of its competitors.

Process development partner Sony Corp., meanwhile, is formulating plans for production at the two next-generation nodes and said it will migrate to the trench-capacitor structure from a stacked-capacitor architecture developed with Fujitsu.

Toshiba has been shipping engineering samples of 90-nm embedded-DRAM devices since November and expects to begin volume production in March with a startup capacity of 1 million units per month. Small-scale production of devices at the 65-nm node is expected a year later, in March 2004.

http://www.eetimes.com/story/OEG20030117S0029


Again, I do not expect PlayStation 3 to launch in Japan before March-April ( or even September ) 2005 and by early-mid December 2005 in the U.S.

Sony and Toshiba have quite a lot of fabs already and the new fabs they are building will be able to produce chips even if not at full speed before PlayStation 3 is launched... plus you also have IBM which could help if the situation calls for it.

The current fabs will have to be dismantled and the new .65nm instealed unless they are upgradable which we don't know . Once the construction of the actually physical building is done yes they can bring in the machines and get it up but i'm sure it will take a few months two put the buildings up(esp in the middle of the winter) . I don't see the ps3 launching till 2006. Ibm is saying they will put the cell out 2007ish. So i doubt they will help sony . Alsa we will have to wait and see what happens . For me it doesn't matter . If the ps3 doesn't have the games i want i wont buy it (Well untill it becomes dirt cheap) Same goes for the other 2 systems .

 

[iapetus]

One thing that a lot of people seem to be assuming is that a large number of articles mentioning a rumour provides some strength to it. This isn't the case. Given the incestuous nature of online reporting, it's likely that they all come from the same original source, and the other sites have jumped on the news and written it up in their own words.

The only multiple sources that are worthwhile are those which actually show different initial source information.

 

[Panajev2001a]

Generally you build multi-billion dollars fab with the idea of upgrading them and not dismantling/destroying/rebuilding from head to toe: also notice that a major portion of PlayStation 2 manufacturing is being moved to China and that means more and more fab space..."unutilized"

The new Oita fab is expected to start yelding 25,000+ 300mm wafers a month by April 2004 and I bet even the new Nagasaki fab will be able to start relatively "low volume" production before PlayStation 3 is released ( in Japan I expect any date between March and September 2005 and early-mid December 2005 in the North American market )...

I think 600,000-800,000 consoles for the Japanese launch and 800,000+ consoles for the North American launch should be possible...

That would mean that you would need up to 1,600,000 consoles produced by December 2005 with a part of that produced between the Japanese launch and the North American launch.

25,000 Wafers a month for 1 fab ( Oita )... April 2004 to April 2005 is 12 months...

That would mean 300,000 wafers...

Let's say they can do up to 6 Broadband Engines or Visualizers per wafer and let's say that the amount of working chips per wafer ends up being 2-3...

That would bring you in the 600,000-800,000 ball-park we were mentioning before...

Of course the situation is not going to be as nice as the one I presented, but also remember that I was only talking about the new Oita fab and I was not adding early production from the new Nagasaki plant and other chips coming from upgraded fabs in Oita and Nagasaki...

 

[jvd]

Generally you build multi-billion dollars fab with the idea of upgrading them and not dismantling/destroying/rebuilding from head to toe: also notice that a major portion of PlayStation 2 manufacturing is being moved to China and that means more and more fab space..."unutilized"

The new Oita fab is expected to start yelding 25,000+ 300mm wafers a month by April 2004 and I bet even the new Nagasaki fab will be able to start relatively "low volume" production before PlayStation 3 is released ( in Japan I expect any date between March and September 2005 and early-mid December 2005 in the North American market )...

I think 600,000-800,000 consoles for the Japanese launch and 800,000+ consoles for the North American launch should be possible...

That would mean that you would need up to 1,600,000 consoles produced by December 2005 with a part of that produced between the Japanese launch and the North American launch.

25,000 Wafers a month for 1 fab ( Oita )... April 2004 to April 2005 is 12 months...

That would mean 300,000 wafers...

Let's say they can do up to 6 Broadband Engines or Visualizers per wafer and let's say that the amount of working chips per wafer ends up being 2-3...

That would bring you in the 600,000-800,000 ball-park we were mentioning before...

Of course the situation is not going to be as nice as the one I presented, but also remember that I was only talking about the new Oita fab and I was not adding early production from the new Nagasaki plant and other chips coming from upgraded fabs in Oita and Nagasaki...

Remember not all fabs can be updated to .65nm . Amd's fabs were made to go from .18-.13 SOI . They said they would need new machines for .9 .

Your numbers are nice . But don't forget they still need hardware to get the bugs out . SO you will prob need to shave off at least 3 months from your caculations. If its really buggy hardware then more.

 

[Panajev2001a]

Even assuming that they would have to dismant parts of some of the old fabs this is going to happen while the new fabs are being built and it would be completed IMHO before the new fabs are capable of producing 65 nm chips...

The new Oita fab is scheduled to start that kind of relatively "low volume production" by April 2004 and that include erecting the new buildings and setting up the equipment.

I think this would take more time than updating older fabs or dismanting old equipment and setting up the new one...

When they were manufacturing PlayStation 2 and they prepared for its Japanese launch they basically had 2 fabs and the second one, manufacturing the GS, started having bad yelds problems...

Now they are going to have 2 fabs + 2 being built ( which can start early production befor being fully completed ) and a relatively "low volume production" from 4 fabs should allow them to manufacture 600,000 consoles for the Japanese launch and some more to complement the number of consoles produced between the Japanese PlayStation 3's launch and the North American one.

Also consider that by September 2005 you should be able to see increased yelds as the new fabs come closer and closer to completion and the "upgraded" ( even if this involves some dismantling ) fabs are increasing their output capacity...

 

[jvd]

Even assuming that they would have to dismant parts of some of the old fabs this is going to happen while the new fabs are being built and it would be completed IMHO before the new fabs are capable of producing 65 nm chips...

The new Oita fab is scheduled to start that kind of relatively "low volume production" by April 2004 and that include erecting the new buildings and setting up the equipment.

I think this would take more time than updating older fabs or dismanting old equipment and setting up the new one...

I agree that it be quicker to upgrade the old plants than build the new ones. But of course we don't know what the older plants are going to be building. Also as i have said they need to get alpha hardware done. They need to test it and fix any problems in the silicon. This could take a short time (like 3 months ) or a longer time . We should have a better view of everything in a few short months along with rumors of the other systems to see how everything stacks up .

 

[Panajev2001a]

I understand your concerns jvd...

I think there is a reason top executives from both Sony and Toshiba are staying close to Cell design and manufacturing and I am sure IBM is going to take the necessary precautions to "limit" the number of possible problex that will appear in alpha HW... they cannot eliminate them all, but a good design planned in advance usually reduces this last minute flaws and/or allows for "easier" last minute debugging of early silicon.

Remember the Japanese launch date I gave ranged from March till September 2005... this is a ~5 months window which should take into account bugs fixing in alpha silicon.

 

[jvd]

I understand your concerns jvd...

I think there is a reason top executives from both Sony and Toshiba are staying close to Cell design and manufacturing and I am sure IBM is going to take the necessary precautions to "limit" the number of possible problex that will appear in alpha HW... they cannot eliminate them all, but a good design planned in advance usually reduces this last minute flaws and/or allows for "easier" last minute debugging of early silicon.

Remember the Japanese launch date I gave ranged from March till September 2005... this is a ~5 months window which should take into account bigs fixing in alpha silicon...

Yes i know. There is so much that can go wrong and that is why i really think they will go with a 2006 launch date. A year late to the market isn't to big of a problem. More than that is . So if ms is launching x mass 2005 and xmass or sept 2006 launch wouldn't be bad for sony. If they fight to launch to early they may end up with buggy hardware out in the market which you just can't do in the console world.

 

[Panajev2001a]

That is also true, they can afford to launch some months after Xbox 2 if this comes with some exclusive 3rd party games, good 1st party titles and better APIs ( more refined thanks to the extra development ) and this would also help them reduce costs on other technologies whcih Sony would want to include in PlayStation 3 like Blu-Ray...

The PlayStation brand is quite powerful and PlayStation 2 is helping it grow even more: they know how to hype the PlayStation 3 and get the word out and the mouth to mouth advertisement going and lots of free advertisement from the press eager to cover new details on Sony's next-generation PlayStation console.

Also, the introduction of a PSTwo (major re-design of the PlayStation 2 ) and a revised/redesigned PSOne would still get quite a bit of sales and publishers would still be inclined to keep development for those two platforms going...

There is no risk involved for them as the PlayStation 3 is supposed to be backward-compatible with PlayStation 2 and PSX ( in order to run PlayStation 2 games it would need an emulated PSX core anyways, unless they embed it too, and then you can re-use that efficiently and accurately emulated PSX core for Software PSX emulation and run the PSX GPU code on the PlayStation 3's GPU like on PlayStation 2 the graphics code is run on the GS IIRC ) so new PlayStation 2 titles would benefit from the increased experience in PlayStation 2 coding across programmers and people would not be scared of buying those games as their money would be invested in software that would not be thrown away when the user changes from PlayStation 2 to PlayStation 3.

By that time PlayStation 2 userbase will be even bigger and making PlayStation 2 games will be cheaper too...