Err, why would they do that when they can just remove it altogether and use software emulation?(which is their plan anyway...)
Uhm cause they are still producing PStwo's? And also, until that emulation is ready, it makes sense to move EE+GS to 45?
Cell 65nm production started
- Thread starter Ragnarok
- Start date
Uhm cause they are still producing PStwo's? And also, until that emulation is ready, it makes sense to move EE+GS to 45?
They have hinted at "major changes" to the firmware in time for the European launch.
Of course this could be many things, but in my opinion is likely to be the inclusion of the emulation software and hopefully the up scaling of all PS2 games to 720p as they had initially promised.
Of course this could be many things, but in my opinion is likely to be the inclusion of the emulation software and hopefully the up scaling of all PS2 games to 720p as they had initially promised.
When did they promise that?
Agreed, makes sense.
Fu3lFr3nzy
Regular
I thought of the first answer as soon as I posted my previous statement(not editing ftl
) and the second one: ah, it all makes sense now Still, if software emulation doesnt come out this year, then "wtf" is all I can say about the situation
Who cares about PS2 scaling, they have a list of things to fix or add before something petty like that. They need a 1080i fix for 720P games, background DLing in the PS store, more PSN demos and arcade games, etc.
I also don't remember Sony ever saying they were going to add features to BC games, it works great now, move on to important things.
Fu3lFr3nzy
Regular
But.... Upscaling and Free AA on PS2 games is teh fun
But yeah, I agree on resolving the PS3 problems first before the others
So if things go well for Sony, they get 65nm Cells, and PS2 emulation so they can drop the EE+GS, saving them a lot more on the hardware just around the European launch...where they charge us the most for the hardware when it's at it's cheapest!
It'd be nice to know when RSX @ 65nm will occur. That'd probably be the ideal early system to get (if price is no concern) as you'll have the coolest, quietest system for a good while.
Although, are smaller processes more or less reliable? This is something that's been concerning me of late, having more high-tech electronic widgets die on me than elder low-tech electronic widgets ever seemed to. Smaller circuitry means less ability to endure wear and tear. Does the lower power consumption mean less wear and tear and 65nm chips will last longer, or do smaller and smaller nodes mean shorter and shorter lifespans for processors?
Panajev2001a
Veteran
It depends on how the shrink is performed. The reason why it is taking them so long (relatively to competitors which already ship 65 nm chips) is also because CELL using quite a bit of custom designed logic and tailored to their specific 90 nm process dynamic logic. All things that have to be redone from scratch for the 65 nm die-shrink. A lot of work has to be done to reduce power consumption and also to better extract the heat dissipated (it is much easier to dissipate heat and cool down the chip when the chip has a large surface), but I do not think there is any reason to believe that life of the chip should be any worry for the current chips or the future chips nowadays (I'd hope so, they are not newbies
).Cheaper (after a while) and cooler CELL CPU, no EE+GS, and no 32 MB of Direct RDRAM should help quite a bit to cost-reduce PLAYSTATION 3 and with a cheaper and cooler 65 nm RSX too they should be able to save quite a few bucks as far as cooling system is concerned.
No, well, that session is about SRAM and not Cell itself contrary to the reports by some media.
I think we're talking about the same thing, I just wrongly highlighted the processor speed when that was not the central focus of the presentation (though they mentioned it in their blurb on the ISSCC site). I guess rather what I meant was, are their any presentation details containing any of the specifics from the presentation?
I was hoping we'd get some insights into improved thermal and/or power profiles since that seemed to be a thrust with this.No more details on the site I pulled the pic. Apparently the 65nm Cell pic was briefly shown for 20 seconds to quench those who anticipated too much from the presentationI think we're talking about the same thing, I just wrongly highlighted the processor speed when that was not the central focus of the presentation (though they mentioned it in their blurb on the ISSCC site). I guess rather what I meant was, are their any presentation details containing any of the specifics from the presentation?
Hiroshige Goto put his article with pics including the shmoo charts for the 65nm Cell.
http://pc.watch.impress.co.jp/docs/2007/0216/kaigai338.htm
Basically not much new info from the ISSCC presentation itself, so Goto supplemented it with other topics and speculations.
Toshiba was at Rambus Developer Forum in Nov. 2006 and suggested they'd manufacture 65nm Cell for consumer appliances. It has 4 SPEs and they plan to remove PPE too. FlexIO may be replaced with a more general interface, but XDR RAM is still necessary for bandwidth.
Also Goto speculates the SOI eDRAM in the other IBM presentation may be used for larger LS in SPE in the future.
http://pc.watch.impress.co.jp/docs/2007/0216/kaigai338.htm
Basically not much new info from the ISSCC presentation itself, so Goto supplemented it with other topics and speculations.
Toshiba was at Rambus Developer Forum in Nov. 2006 and suggested they'd manufacture 65nm Cell for consumer appliances. It has 4 SPEs and they plan to remove PPE too. FlexIO may be replaced with a more general interface, but XDR RAM is still necessary for bandwidth.
Also Goto speculates the SOI eDRAM in the other IBM presentation may be used for larger LS in SPE in the future.
PPE is not an essential part of Cell. They can use MIPS or whatever they want.
