View Full Version : So is PS3 GPU one generation beyond Xenon graphics chip?
Shinjisan
22-Dec-2004, 01:23
Since you're so well informed here I have a question:
http://www.digitimes.com/news/a20041220A7039.html
"Nvidia has not given up on the NV48 and is expected to roll out the GPU in the second quarter of next year, despite market rumors saying the company has cancelled the chip, according to sources at Taiwan graphics card makers.
However, Nvidia plans to have Taiwan Semiconductor Manufacturing Company (TSMC) produce the NV48, instead of IBM, as originally planned, the sources noted. The NV48 will be built using a 0.11-micron process, the sources added.
Improved yields of the GeForce 6600 and GeForce 6200 manufactured at TSMC would probably be the reason for Nvidia to shift the production of the NV48 back to TSMC, the sources said.
The NV48 will take on ATI’s next-generation high-end graphics chip, the R520, which is also expected to be launched in the second quarter of 2005. The R520, however, will be the industry’s first GPU produced using a 90-nm process at TSMC."
Xenon GPU is code named R500,if ATI R520 is out in Q2 2005 competing with NV48 and the PS3 GPU will be based on NV5X (or equivalent technology),does this put the PS3 GPU potentially one generation beyond the Xenon graphics processor?
Of course we talk about a PC generation so a 6 months difference but if NVidia and Sony decide to use 65nm for PS3 GPU the gap could be really consistent.
In the end does more powerful mean better? Will anyone see the difference is the real question.
Shinjisan
22-Dec-2004, 01:54
More powerful doesn't mean better games but better graphics for sure.
The R520 is completely unrelated to the xenon gpu.
Brimstone
22-Dec-2004, 02:31
My guess is both ATI and Nvidia GPU's destined for the Xenon and PS3 are designed for .90 nm. The amount of eDRAM will depend on if it is fabbed on .90 nm or .65 nm. The target for Sony would seem to be .65 nm, while .90 nm is more likely for MS. Supposedly TSMC will have .65 nm available to customers at the end of 2005, so the Xe R500 GPU has the possibility of being a .65nm product.
Xenon .90 nm CPU and GPU with 512mb of GDDR-4 or 1024mb of DDR-II with QBM-2 integrated into the SIS chipset to double the bandwidth of the standard DDR-II DRAM.
SIS jons the QBM alliance (http://www.sis.com/news/press/qbm.htm)
aaronspink
22-Dec-2004, 02:45
My guess is both ATI and Nvidia GPU's destined for the Xenon and PS3 are designed for .90 nm. The amount of eDRAM will depend on if it is fabbed on .90 nm or .65 nm. The target for Sony would seem to be .65 nm, while .90 nm is more likely for MS. Supposedly TSMC will have .65 nm available to customers at the end of 2005, so the Xe R500 GPU has the possibility of being a .65nm product.
Probably a better way to think of it is that they are both targetted for 65 nM but that the initial implementation will be done in 90 nM for timing reasons. There is a tradition of console companies launching the console with die sizes that are expensive to seed the market while working on shrinking the design to the next process.
Xenon .90 nm CPU and GPU with 512mb of GDDR-4 or 1024mb of DDR-II with QBM-2 integrated into the SIS chipset to double the bandwidth of the standard DDR-II DRAM.
AFAIK, QBM never existed and QBM-2 certainly doesn't exist. Xenon is most likely using 4 32b wide 256Mb or 512Mb GDDR3/4 chips @ 1600 Mtps.
Aaron Spink
speaking for myself inc.
QBM-2 certainly doesn't exist
Why do you say that, IMO would be very good for Xe 2X the bandwidth at almost the same price, they win over the PCs in cost!
Joshua Luna
22-Dec-2004, 03:16
Maybe some of the HW gurus can help me with this.
If R500 is at 90nm, the PS3 GPU would need to be at ~ 45nm to have twice as many transistors on the same die size, right?
And at 65nm it would be what? 30-50% more transistors?
Or am I way off...
Anyhow, I know that clock speed, die size, and the architecture make a big difference. If R500 taped out recently, and the PS3 will tape out at the end of 2005, Nvidia has about a 12 month area to do some work. Not sure if 12 months is enough to double the power... but all is speculation since we know nothing really about both chips. Who knows, maybe R500 will be some killer chip. I wonder what features, other than VS/PS, are being planned for these chips. I think some great new features will be more important than just raw speed (although that is important too), but that is just an opinion. Exciting time :)
Brimstone
22-Dec-2004, 03:18
My guess is both ATI and Nvidia GPU's destined for the Xenon and PS3 are designed for .90 nm. The amount of eDRAM will depend on if it is fabbed on .90 nm or .65 nm. The target for Sony would seem to be .65 nm, while .90 nm is more likely for MS. Supposedly TSMC will have .65 nm available to customers at the end of 2005, so the Xe R500 GPU has the possibility of being a .65nm product.
Probably a better way to think of it is that they are both targetted for 65 nM but that the initial implementation will be done in 90 nM for timing reasons. There is a tradition of console companies launching the console with die sizes that are expensive to seed the market while working on shrinking the design to the next process.
Xenon .90 nm CPU and GPU with 512mb of GDDR-4 or 1024mb of DDR-II with QBM-2 integrated into the SIS chipset to double the bandwidth of the standard DDR-II DRAM.
AFAIK, QBM never existed and QBM-2 certainly doesn't exist. Xenon is most likely using 4 32b wide 256Mb or 512Mb GDDR3/4 chips @ 1600 Mtps.
Aaron Spink
speaking for myself inc.
The QBM concept is real. You wouldn't have all those big name companies apart of the alliance if it didn't work.
I don't know if there's any technical reason stopping GDDR-3/4 from getting the QBM treatment.
Maybe some of the HW gurus can help me with this.
If R500 is at 90nm, the PS3 GPU would need to be at ~ 45nm to have twice as many transistors on the same die size, right?
And at 65nm it would be what? 30-50% more transistors?
....
)
(90/65)^2 ~ 2x
The area available to transistors would be a square of the process drop...
There is a tradition of console companies launching the console with die sizes that are expensive to seed the market while working on shrinking the design to the next process.
Where did you get that idea from ? It isn't a tradition, that's just what Sony did with PS2.
Looking at the shortage in the early stages, it isn't a good idea IMO. They're lucky that Sega managed to screw up Dreamcast marketing. This time around good supply is more essential than technology.
Joshua Luna
22-Dec-2004, 05:59
Maybe some of the HW gurus can help me with this.
If R500 is at 90nm, the PS3 GPU would need to be at ~ 45nm to have twice as many transistors on the same die size, right?
And at 65nm it would be what? 30-50% more transistors?
....
)
(90/65)^2 ~ 2x
The area available to transistors would be a square of the process drop...
Thanks Jaws :)
aaronspink
22-Dec-2004, 09:55
QBM-2 certainly doesn't exist
Why do you say that, IMO would be very good for Xe 2X the bandwidth at almost the same price, they win over the PCs in cost!
QBM doesn't provide any benefit in an application that isn't DIMM based. For an application with directly connected memory, you will also be targetting to run at a much higher data rate than the QBM asics can handle.
Also QBM hasn't to my knowledge EVER been shipped.
Aaron Spink
speaking for myself inc.
aaronspink
22-Dec-2004, 09:59
The QBM concept is real. You wouldn't have all those big name companies apart of the alliance if it didn't work.
I don't know if there's any technical reason stopping GDDR-3/4 from getting the QBM treatment.
What big names? Is AMD shipping support for it? Is Intel shipping support for it? Is Nvidia shipping support for it? Is ATI shipping support for it?
looking at the QBM web page, I don't see any big names.
As far as technical reasons... Signalling speeds.
Aaron Spink
speaking for myself inc.
london-boy
22-Dec-2004, 10:17
Maybe some of the HW gurus can help me with this.
If R500 is at 90nm, the PS3 GPU would need to be at ~ 45nm to have twice as many transistors on the same die size, right?
And at 65nm it would be what? 30-50% more transistors?
Or am I way off...
Hey this is elementary geometry we're talking about!! :twisted:
Edit: Oh, Jaws answered...
Guden Oden
22-Dec-2004, 10:41
And at 65nm it would be what? 30-50% more transistors?
The actual question has already been answered, but then there's the issue of actual die size.
If we assume PS3's GPU uses a deeper process than nextbox (quite an assumption already), and we assume die size is exactly the same, then PS3 GPU will have more trannies. But we could just as well assume Sony wants a similar-class GPU as what ATi is designing for MS, so they'll end up having roughly same amount of trannies. PS3 die size will therefore be smaller.
Then again, looking at PS2, we know Sony isn't afraid of large chip dies, so it could be a monster chip too, who knows. We're assuming left and right with no solid foundation to base any of this on. :)
london-boy
22-Dec-2004, 10:48
In the end it all comes down to cost, like i've said so many times before.
Sony will draw a line somewhere, when they think they have enough performance for the money they'll be spending, and anything more would be a luxury they might not be able to afford.
I don't think Sony should worry too much about being a lot more powerful than MS, they'll end up spending a lot of money for something that really doesn't affect the return of investment to them as much as GTA or FF do.
It will be valuable to them to have ze bigz numberz, but look at what happened with PS2, many people still think PS2 is more powerful than the competition.
Sony and everyone else better spend their money somewhere they know they'll make a profit out of. GTA. Pokemon. Halo. Those are the things that make the most money for them, people buy consoles based on what big game they can play on them, not because one GPU runs 500MHz faster than the other one.
Brimstone
22-Dec-2004, 10:54
The QBM concept is real. You wouldn't have all those big name companies apart of the alliance if it didn't work.
I don't know if there's any technical reason stopping GDDR-3/4 from getting the QBM treatment.
What big names? Is AMD shipping support for it? Is Intel shipping support for it? Is Nvidia shipping support for it? Is ATI shipping support for it?
looking at the QBM web page, I don't see any big names.
As far as technical reasons... Signalling speeds.
Aaron Spink
speaking for myself inc.
VIA and SIS are a among those listed as supporters and they're certainly chipset middleweights.
Kentron isn't seeking royalties on the patents they have, so I assume companies can customize the technologys as they please. However, assuming what a company would allow is dangerous though.
in the end it doesnt matter cuz sony is going to dominate next gen.
aaronspink
22-Dec-2004, 11:57
VIA and SIS are a among those listed as supporters and they're certainly chipset middleweights.
Kentron isn't seeking royalties on the patents they have, so I assume companies can customize the technologys as they please. However, assuming what a company would allow is dangerous though.
Neither VIA nor SIS have support for QBM in their chipsets. This is a technology that has so far been all hype with no products on the market.
But this is all irrelevent for the area of consoles and graphics chips since they use directly connected on board memory and would not benefit from any intermediate chip. If they wanted faster memory, they would order faster memory.
Aaron Spink
speaking for myself inc.
Probably a better way to think of it is that they are both targetted for 65 nM but that the initial implementation will be done in 90 nM for timing reasons.
Why is that a better way to think about it? Specifically, why do you assume everyone is shipping (anything) at 90nm? Your all inclusive comment is wrong, more to the point, why not posit a company launching at 65nm is "targeting" 45nm?
NocturnDragon
22-Dec-2004, 15:06
in the end it doesnt matter cuz sony is going to dominate next gen.
It's just me or the signal/noise ration is lowering everyday more?
in the end it doesnt matter cuz sony is going to dominate next gen.
It's just me or the signal/noise ration is lowering everyday more?
Don't mind, they are subliminal voices :lol:
in the end it doesnt matter cuz sony is going to dominate next gen.
Oh ok, lets call up Microsoft and Nintendo to tell them to quit before its too late. Honestly, this forum makes me sad about the ignornace and stupidity of people.
Vince,
Why is that a better way to think about it? Specifically, why do you assume everyone is shipping (anything) at 90nm? Your all inclusive comment is wrong, more to the point, why not posit a company launching at 65nm is "targeting" 45nm?
out of curiosity, what process did sony use for PSP chips? was it any lower than people on the net expected?
out of curiosity, what process did sony use for PSP chips? was it any lower than people on the net expected?
ASC9, a Sony developed 90nm Low-K embedded DRAM logic process. It was in line with what was stated by Sony and stated by us.
Out of curiousity, why did you post? The only way for your argument to be analogous is if it was expected that PS3 launch on a 90nm process. Which is a fringe position at this time and against everything from SCE comments to their investment (http://www.sony.net/SonyInfo/News/Press/200402/04-0202E/) to sheer inability to process enough wafers. The consensus is a 65nm launch with quick transition to 45nm for early 2006 (atleast concerning the OTSS/SCE fabbed GPU) as stated here (http://www.sony.net/SonyInfo/News/Press/200402/04-0212E/).
Not sure about that being 'stated'. All it states there is 'positive results are expected in 2005'. It doesn't define 'positive' (as 'ready for production in 2006' for example).
The first results of the collaboration are expected within two years, in 2005, the companies said. They plan to beat the rest of the semiconductor industry to the 45-nm process node and are starting 65-nm designs for application to sample chips "soon".
The joint development agreement calls for completion of the project by late 2005, with the ultimate goal of being first to market with 45-nm know-how. The project has a budget of ¥20 billion (about $190 million) and approximately 150 engineers from the two companies are expected to work on the project at Toshiba's Advanced Microelectronics Center in Yokohama, Japan and Oita Operations in Kyushu.
I, personally, do anticipate them sitting on this untill 2007 so they can launch PlayStation4 with it.
Not sure about that being 'stated'. All it states there is 'positive results are expected in 2005'. It doesn't define 'positive' (as 'ready for production in 2006' for example).
If 45nm is positive, what's 65nm like? Super-duper-positive? 8)
Of the 120 billion yen, approximately 53 billion yen will be used to reinforce the fabrication line located in the clean room on the first floor in SCEI's Fab2, for the manufacturing of 65 nanometer generation high-performance LSIs such as "Cell", using SOI process technology.
You can make a different assumption like the PS3's processors may launch in 90nm then will move to 65nm, but considering the demand of the PSP and the PSTwo next year, 90nm lines will be at full capacity for some time.
Out of curiousity, why did you post?
What kind of question was that? I was curious is people on the net were expecting PSP to be targetting 65nm and it ended up being 90nm for now with a die shrink coming in the future.
What kind of question was that? I was curious is people on the net were expecting PSP to be targetting 65nm and it ended up being 90nm for now with a die shrink coming in the future.
I think you're quite mistaken. The IC was allways intended to be a 90nm SoC. There currently is, from what I hear, some debate over if they should transition quicker to 65nm or revise the 90nm design. There being several external aspects to balance outside of the sheer process gains in terms of profitability and power consumption, such as fab capacity.
in the end it doesnt matter cuz sony is going to dominate next gen.
Oh ok, lets call up Microsoft and Nintendo to tell them to quit before its too late. Honestly, this forum makes me sad about the ignornace and stupidity of people.
since when did stating the obvious become ignorant?
in the end it doesnt matter cuz sony is going to dominate next gen.
It's just me or the signal/noise ration is lowering everyday more?
sorry i dotn speak foblish
aaronspink
22-Dec-2004, 21:47
Probably a better way to think of it is that they are both targetted for 65 nM but that the initial implementation will be done in 90 nM for timing reasons.
Why is that a better way to think about it? Specifically, why do you assume everyone is shipping (anything) at 90nm? Your all inclusive comment is wrong, more to the point, why not posit a company launching at 65nm is "targeting" 45nm?
Because they can't afford to lose that much money waiting until '08 to ship at reasonable die sizes. And those that don't ship initially at 90nm will do developer seeding using 90nm parts or be late to market.
Aaron Spink
speaking for myself inc.
aaronspink
22-Dec-2004, 21:55
If 45nm is positive, what's 65nm like? Super-duper-positive? 8)
well once you use the sony hype processor to industry standard process decoders your sentence reads as follows...
"if 65nm is positive, what's 90nm like? Super-duper-positive? 8)"
Sony has a habit of embelishing their technology. Lets state a positive. I am positive that no one will beat Intel to a process node. In general the companies pushing the technology are doing good to hit within 6 months of the Intel transition. This is generally the case because Intel does a large portion of the tools development with the process equipment vendors.
Aaron Spink
speaking for myself inc.
Sony has a habit of embelishing their technology. Lets state a positive. I am positive that no one will beat Intel to a process node.
OTTAWA, CANADA, FEBRUARY 04th, 2004 – Chipworks Inc. (“Chipworks”), the standard setting supplier of reverse engineering services, commented today that they consider that the Sony PSX device is manufactured on one of the most advanced processes in production.
“Chipworks obtained a sample of the Sony PSX chip and had started to take it apart based on the reports that it was a 90 nm part, but we have come to somewhat different conclusions than those reported last week.” stated Chipworks senior technology analyst Dick James. “For one thing, we found transistors with a physical gate length of 45 – 50 nm, which immediately led us to believe that we did have a 90 nm part.”
While the Sony/Toshiba process may not meet the letter of the International Technology Roadmap for Semiconductors (ITRS) 2003 it is comparable with the leading edge Intel process in terms of gate dimension. The highly anticipated Intel 90 nm device announced last fall at the Intel Developer Forum is reported to have a gate length of 45 nm. This is a little larger than the gate length listed in the ITRS, which benchmarks the polysilicon gate length at 37 nm.
In addition the Sony device uses an advanced two stack low-k dielectric structure. The combination of this dielectric process and the smallest transistor seen so far by Chipworks makes this one of the most advanced processes in volume production today.
<center>http://www.chipworks.com/news/newsimages/SonyPSX.gif</center>
When did the EE+GS@90 powered PlayStation2's go on sale, refresh my memory Aaron.
Because they can't afford to lose that much money waiting until '08 to ship at reasonable die sizes. And those that don't ship initially at 90nm will do developer seeding using 90nm parts or be late to market.
:roll: I'm sorry... I had a recollection from 1999. I can't say much more than this, you just continue to defy all logic -- there are 90nm first generation Cell chips in circulation today. 65nm samples are being evaluated today with an investment in 65nm production of well over $3Billion and a Sony contract on initial production from this node. Ohh, and 45nm has a ~2006 introduction -- stop being obtuse.
It depends .
If we look at recent past history you get
nv3x new nvidia tech
r3x0 new ati tech
The new ati tech as clearly the winner
nv4x new nvidia tech
r42x modified ati tech
THey are pretty much even in performance with the nv4x having the edge in features though a small diffrence
The r520 is not a new tech either. Its an r3x0 with sm3.0
So it may very well be that both the ps3 gpu and xbox system would be similar to each other .
When did the EE+GS@90 powered PlayStation2's go on sale, refresh my memory Aaron.
I don't know but Intel were producing 90nm processors in September 2003.
Those reports about wether it 90nm or not seem to indicate EE+GS@90 being Jan 2004.
I don't know but Intel were producing 90nm processors in September 2003.
I know Deano, I wanted Aaron to answer since it seems implicit that he knows...
Production started in 2003:
Sony Launching 90nm Embedded DRAM
Sony's subsidiary that makes the PlayStation 2, Sony Computer Entertainment Inc. said today that it would be in production on 90nm embedded DRAM process in fiscal 2003...
Sony claims it will be ahead of anyone else in the industry on 90nm embedded DRAM volume production. Production of the chip will start from this spring at Oita TS Semiconductor (OTSS), a joint venture between Sony Computer Entertainment and Toshiba Corp., and in the fall at Sony's own fab in Isahaya City, Nagasaki Prefecture, Japan.
CX-News, which is Sony's publication, stated they had begun shipping PS2's with the 90nm processor out of Nagasaki in October of 2003.
<center>http://www.chipworks.com/mail/2004_02/sony_90nm.jpg http://www.chipworks.com/mail/2004_02/Intel_90nm.jpg
Sony CXD9797GB -- (via Chipworks) -- Intel "Prescott"</center>
so if intel shipped them in sept wouldn't it have been after intel which means it wasn't the market leader ?
Correct me if I'm wrong, but Intel was still sampling it's 90nm Prescott and Dothan cores in September of 2003 and suffered problems which delayed production from Q2 2003 to Q1 2004 (http://www.cooltechzone.com/reviews/cpu/itwbu64p_030.php). SCE was sampling according to the one report (posted above) that summer and you could buy a 90nm PS2 or PSX that Fall or Christmas (Dec 13th) respectively.
So, today, on February 2, 2004 Intel Corporation announced and started selling new processors formerly known as Prescott. These processors based on a new 90nm core will continue the Pentium 4 family at least for another year little by little ousting the previous 130nm Northwood core.
If 45nm is positive, what's 65nm like? Super-duper-positive? 8)
well once you use the sony hype processor to industry standard process decoders your sentence reads as follows...
"if 65nm is positive, what's 90nm like? Super-duper-positive? 8)"
Sony has a habit of embelishing their technology. Lets state a positive. I am positive that no one will beat Intel to a process node. In general the companies pushing the technology are doing good to hit within 6 months of the Intel transition. This is generally the case because Intel does a large portion of the tools development with the process equipment vendors.
Aaron Spink
speaking for myself inc.
Intel is free from hype? Where's my 4GHz? :cry:
Well I hate to point out something to those who don't bother to use google by themselves :lol:
Intel released 90nm Pentium 4E (Prescott) at 2004/02/01.
Sony released 90nm EE+GS (SoC with eDRAM) in the PSX at 2003/12/13.
I don't know jack about how well the process technology at Intel is, but I'm sure the design of Prescott and its too high TDP prevented Intel from moving in 90nm process any quicker as it was too hot for a commercial product. IIRC they planned to release 90nm Prescott in late 2003 but the fact is they couldn't and it's the official history too. Enough of execuse already :P
Correct me if I'm wrong, but Intel was still sampling it's 90nm Prescott and Dothan cores in September of 2003 and suffered problems which delayed production from Q2 2003 to Q1 2004 (http://www.cooltechzone.com/reviews/cpu/itwbu64p_030.php). SCE was sampling according to the one report (posted above) that summer and you could buy a 90nm PS2 or PSX that Fall or Christmas (Dec 13th) respectively.
So, today, on February 2, 2004 Intel Corporation announced and started selling new processors formerly known as Prescott. These processors based on a new 90nm core will continue the Pentium 4 family at least for another year little by little ousting the previous 130nm Northwood core.
vince my fathers centrino is 90nm tech and was puchased last nov .
Intel has a habit of using the new manufacturing process for mobile chips first where it is needed the most then moving the desktop chips to it
Megadrive1988
22-Dec-2004, 23:38
I think both Xenon and PS3 GPUs will be similar in many ways. both should be capable of peak vertex performance in the low billions and in-game real-world performance of hundreds of millions of vertices/polygons with lighting, shaders, textures, rendering features on. it might be Xenon is is the low-to-mid 100s of millions and PS3 is in the mid-to-upper 100s of millions. there will be some difference but not too much. kinda like the difference between Gamecube and Xbox (geometry wise) and about the same in terms of shader capability. still, i'm just specu-guessing 8)
vince my fathers centrino is 90nm tech and was puchased last nov .
Intel has a habit of using the new manufacturing process for mobile chips first where it is needed the most then moving the desktop chips to it
Pentium M @ 90nm (Dothan) was released at 2004/05/10.
vince my fathers centrino is 90nm tech and was puchased last nov .
Intel has a habit of using the new manufacturing process for mobile chips first where it is needed the most then moving the desktop chips to it
Pentium M @ 90nm (Dothan) was released at 2004/05/10.
Hmm time to look in the laptop again !
In other news tsmc has been installing 65nm machines since november
TSMC believes immersion lithography may be called upon for 65nm production and are the preferred method for 45nm production. TSMC said it began installing its first 65nm immersion lithography system in early November.
http://www.reed-electronics.com/electronicnews/article/CA489760.html?ref=nbth
vince my fathers centrino is 90nm tech and was puchased last nov
Interesting, what core is it? It was my understanding that Banias was 130nm and that the 90nm Dothan was pushed back to 2004:
Most recently scheduled for an early Q1 launch, the 90nm chip will now ship during Q2, probably in May
EDIT: One, you're too damn fast. :P
vince i can be wrong , one says i'm wrong , I have to check again , I could have read the core wrong .
I wouldn't have responded to you if I saw his post, which is why I said he's too damn fast. I do step away from the laptop from time to time ;)
Concerning 65nm, TSMC's roadmap. (http://www.tsmc.com/english/technology/t0112.htm) Makes you wonder about the supposed eDRAM since TSMC is fabbing the GPU (http://forum.teamxbox.com/showpost.php?p=3456709&postcount=1), something (perhaps the roadmap) doesn't seem right.
EDIT: It's listed as a "Fall 2004" Roadmap and has the following:
Logic
[list:2ced363522] High-Speed
[list:2ced363522] CLN90GT (1.2/2.5V) Sampling 2004, Production 2005
CLN65HS - Sampling 2H2006, Production 2006
General Purpose
CLN90G (1/1.8/3.3V) Available Now
CLN65G - Sampling 2H2006 (4Q), Production to follow[/list:u:2ced363522]
Embedded High Density
1T-Q - CL013G Available Now
1T-P - CL013LP Available Now
1T-MiM - CLN90GT (1.2/2.5V) Sample 1H2006, Production 2006[/list:u:2ced363522]
whats it say , i refuse to have the pdf viewer installed on my system (used to crash my system for so long )
anyway thats the news from today according to that website haha
aaronspink
23-Dec-2004, 00:31
When did the EE+GS@90 powered PlayStation2's go on sale, refresh my memory Aaron.
We can get into quote wars if you like. I stand by my statement.
:roll: I'm sorry... I had a recollection from 1999. I can't say much more than this, you just continue to defy all logic -- there are 90nm first generation Cell chips in circulation today.
I never said there weren't first generation 90nm chips in circulation today.
65nm samples are being evaluated today with an investment in 65nm production of well over $3Billion and a Sony contract on initial production from this node. Ohh, and 45nm has a ~2006 introduction -- stop being obtuse.
well over $3B eh? And the only way they'll get 45nm production in 2006 is if they lie their asses off. 2008. Mark my words. They may do a press demonstration of something that call a 45nm PS3 before that but they will not be in volume production.
Aaron Spink
speaking for myself inc.
Intel is producing when it supplies them for OEM production test, AFAIK that occured in September 2003 for 90nm.
Test was a good few months before. Intel and AMD have to supply OEMs with lots of chips so they can test them, the time you can actually buy them is significantly delayed.
Intel are currently in test with 65nm processors. Which means if you have a very good reason its possible to get one (last time I was involved with getting a test Intel processor (it was a Katmai 300Mhz) it was hand delivered and I was told it cost about $50,000 to produce). Sampling (relatively low volume production) will occur very early 2005, you'll be able to buy one late next year.
We can get into quote wars if you like. I stand by my statement.
What quote war? The fact is anyone could buy a PS2 or PSX fabricated on a Low-K 90nm process in Fall of 2003, you couldn't do the same from Intel or anyone else in the consumer sector for that matter. You can't argue this fact.
well over $3B eh? And the only way they'll get 45nm production in 2006 is if they lie their asses off. 2008. Mark my words. They may do a press demonstration of something that call a 45nm PS3 before that but they will not be in volume production.
Yeah. And I'll assume you're not talking for AMD/IBM for that matter as AMD has stated they anticipate 2007 for 45nm and their development agreement covering 32nm runs untill 2008. So, I'll mark your words all right; I look forward to 2008 and my PS4.
Mulciber
23-Dec-2004, 02:36
What quote war? The fact is anyone could buy a PS2 or PSX fabricated on a Low-K 90nm process in Fall of 2003, you couldn't do the same from Intel or anyone else in the consumer sector for that matter. You can't argue this fact.
you're going to compair availability of a process when one companys target clockspeed is >3ghz and the others is 200mhz :?: :?:
What quote war? The fact is anyone could buy a PS2 or PSX fabricated on a Low-K 90nm process in Fall of 2003, you couldn't do the same from Intel or anyone else in the consumer sector for that matter. You can't argue this fact.
you're going to compair availability of a process when one companys target clockspeed is >3ghz and the others is 200mhz :?: :?:
Isn't the argument is regarding 90nm in a commercial device available for purchase (thus mass production)?
It wasn’t regarding GHz, otherwise they'll be debating which had the samples first.
Can we safely assume a Cell chip at 4.6 GHz @ 90nm exist, given its going thru ICCC? If so what does Intel have that they can reveal to the world? I am sure they have something beyond their current top crop, but how ready is it?
Mulciber
23-Dec-2004, 05:20
What quote war? The fact is anyone could buy a PS2 or PSX fabricated on a Low-K 90nm process in Fall of 2003, you couldn't do the same from Intel or anyone else in the consumer sector for that matter. You can't argue this fact.
you're going to compair availability of a process when one companys target clockspeed is >3ghz and the others is 200mhz :?: :?:
Isn't the argument is regarding 90nm in a commercial device available for purchase (thus mass production)?
It wasn’t regarding GHz, otherwise they'll be debating which had the samples first.
no, that was not "the" arguement. that was the arguement that Vince put forth. considering the nature of what is being argued, clockspeed and transister count both play a crucial role in determining how quickly a chip can be put into a commercialy viable product.
Isn't the argument is regarding 90nm in a commercial device available for purchase (thus mass production)?
It wasn’t regarding GHz, otherwise they'll be debating which had the samples first.
no, that was not "the" arguement. that was the arguement that Vince put forth. considering the nature of what is being argued, clockspeed and transister count both play a crucial role in determining how quickly a chip can be put into a commercialy viable product.
So you want to factor in process size (@90 nm) and speed (xHz)?
Firstly, when comparing the EE+GS to the P4 is already "apples vs oranges".
Secondly, haven't AMD (and Sony and others) proved it’s not just the speed of the chip, but the performance? At lower Hz some chips still blow those a lot fast out of the water under certain conditions.
So what was it we're debating again? :lol:
Mulciber
23-Dec-2004, 05:45
Isn't the argument is regarding 90nm in a commercial device available for purchase (thus mass production)?
It wasn’t regarding GHz, otherwise they'll be debating which had the samples first.
no, that was not "the" arguement. that was the arguement that Vince put forth. considering the nature of what is being argued, clockspeed and transister count both play a crucial role in determining how quickly a chip can be put into a commercialy viable product.
So you want to factor in process size (@90 nm) and speed (xHz)?
Firstly, when comparing the EE+GS to the P4 is already "apples vs oranges".
um, did you miss the point of my question to Vince entirely, then?
that is exactly what i was pointing out
Secondly, haven't AMD (and Sony and others) proved it’s not just the speed of the chip, but the performance? At lower Hz some chips still blow those a lot fast out of the water under certain conditions.
I dont care to look up the flops and gips counts of the p4 or the ee, but my guess is the 3.8ghz p4 is still faster than a 200mhz ee.
edit: though i'd like to add that (to forstal any rebuttles of "well the ee clockrate could go as high as...") that actually has no bearing on the point I was attempting to make
Brimstone
23-Dec-2004, 06:08
Can we safely assume a Cell chip at 4.6 GHz @ 90nm exist, given its going thru ICCC? If so what does Intel have that they can reveal to the world? I am sure they have something beyond their current top crop, but how ready is it?
The SRAM running a 4.6 GHz might be related to "CELL" needing to have low latency because of all the cores and wire delay issues. I believe Deadmeat has guessed at 1.2 Ghz per SPU. To me Deadmeats estimates are more plausible, but maybe he is wrong. It won't be long before the world finds out or someone leaks some real info.
:roll: I'm sorry... I had a recollection from 1999. I can't say much more than this, you just continue to defy all logic -- there are 90nm first generation Cell chips in circulation today. 65nm samples are being evaluated today with an investment in 65nm production of well over $3Billion and a Sony contract on initial production from this node. Ohh, and 45nm has a ~2006 introduction -- stop being obtuse.
Nice guess. You have no idea. I love how all of your "evidence" comes from links in the press room on sony.com.
ISSCC != circulation and test chip != sample
About the only thing we know is that at 90nm the Cell test chip is huge, and Sony knows they better get to 65nm PDQ.
Megadrive1988
23-Dec-2004, 07:12
I believe PS3 will launch on 65nm process. then be moved down to 45nm in 2007. then down to 32nm or 35nm or whatever the next step is from 45nm, sometime in 2008-2009 timeframe. I'm merely guessing that PS4 will be on 25nm or smaller, if that is possible, sometime early in the next decade (2011-2013)
no, that was not "the" arguement. that was the arguement that Vince put forth. considering the nature of what is being argued, clockspeed and transister count both play a crucial role in determining how quickly a chip can be put into a commercialy viable product.
So you want to factor in process size (@90 nm) and speed (xHz)?
Firstly, when comparing the EE+GS to the P4 is already "apples vs oranges".
um, did you miss the point of my question to Vince entirely, then?
that is exactly what i was pointing out
Re-read your question, and I can't answer for Vince, but personally I just see it as another challenge for Intel when dealing with such high speed as it was a different challenge for Sony to integrate the EE+GS. Hence its still "apples vs oranges".
Secondly, haven't AMD (and Sony and others) proved it’s not just the speed of the chip, but the performance? At lower Hz some chips still blow those a lot fast out of the water under certain conditions.
I dont care to look up the flops and gips counts of the p4 or the ee, but my guess is the 3.8ghz p4 is still faster than a 200mhz ee.
Quick google and we get:
EE (core + V0 + V1) = ~ 6.2 GFlops (single precision)
P4 @ 3.4GHz on 90nm = ~ 13.6 GFlops (single precision)
So what's your point? EE+GS @ 90nm was bound to the initial target set back in late 1990s during its design, whereas at the time when Intel transit to 90nm it was pushing the limits with Hz?
Edit: We don't know if Sony could have made the EE alone earlier or faster if the goals were different as the same apply to Intel P4 @90nm. All we can take out of it is which product was first to market at 90nm without factoring in other variables which could have delayed them.
london-boy
23-Dec-2004, 10:05
They're very different chips.
Although it's kinda funny to see the EE @ 300MHz pushing 6.2 GFLOPS and a Pentium at ~3Ghz pushing 13GLOPS. 10 times the clockspeed for double the FLOPS performance...
But they're very different chips, so the comparison has no point.
Heh. Who brought Intel which is not in a console in the next gen? It's aaronspink screaming "Intel is da king in da process technology dammit!" and it's greatly irrelevant here :roll:
Also, if you give Intel 1up for making 3Ghz chip in the 90nm process, you should give SCEI 1up for mixing eDRAM in a chip in the 90nm process. Needless to say, the latter is the more difficult challenge in terms of the process technology.
BTW, 3Ghz and 90nm have no relation to each other - Intel expected shrinking it to 90nm would resolve the heat problem in Prescott, but it was a mistake and they had to redesign their release plans, not the process technology. Pentium4E in 90nm underperformed Pentium4 in 130nm IIRC.
I think both Xenon and PS3 GPUs will be similar in many ways. both should be capable of peak vertex performance in the low billions and in-game real-world performance of hundreds of millions of vertices/polygons with lighting
The GPUs could be a lot different if the Vertex Programs are handle solely by the CPU on the PS3, and if the XeVPU is indeed built using unified architecture (Judging by what we know the XeCPU would also be used for the vertex, though). But that doesn't change the point you made, which is that the two machine, would be more or less in the same ballpark.
BTW, and that's slightly off-topic, but the more i hear about next-gen games, from various statements to actual screenshots, and the less i believe we'll see games with more than 1M polygons per frame (@60Hz, of course), especially the first batch of games.
london-boy
23-Dec-2004, 14:09
BTW, and that's slightly off-topic, but the more i hear about next-gen games, from various statements to actual screenshots, and the less i believe we'll see games with more than 1M polygons per frame (@60Hz, of course), especially the first batch of games.
Getting pessimistic huh?
I'm kinda like that, but more on the side of "the first next gen games will probably have more than 1M polys per frame with 23 shader programs for each pixel, but it will still look like tacky Z-movie kind of crap. *cough*DOOM3*cough...."
That is, until the real games come out. I can only imagine what Temco, Sony's first party teams, Polyphony Digital, Konami will throw at us...
Dave Baumann
23-Dec-2004, 14:09
Heh. Who brought Intel which is not in a console in the next gen? It's aaronspink screaming "Intel is da king in da process technology dammit!" and it's greatly irrelevant here :roll:
Personally I don't ignore those with unique industry insights, whatever they are saying...
Panajev2001a
23-Dec-2004, 14:27
Heh. Who brought Intel which is not in a console in the next gen? It's aaronspink screaming "Intel is da king in da process technology dammit!" and it's greatly irrelevant here :roll:
Personally I don't ignore those with unique industry insights, whatever they are saying...
I don't ignore them either, but I still debate them if I am not convinced... how else am I going to learn ;) ?
I don't ignore them either, but I still debate them if I am not convinced... how else am I going to learn ;) ?
You shouldn't question the omniscient deities Pana, fucks with your karma (as you can see).
Getting pessimistic huh?
Yep, and it's kind of a good thing, since I could only be pleased if things turns better than I expected. Nonetheless, i heard some numbers (polygon numbers) for a few games, and that's low... I'm mean really.
As you said thoses polygons will features tons of shaders, but as you also pointed out, fragment programs could easily be irrelevant, and uselless if they're not coupled with a good amount of polygons... In my opinion.
Riddlewire
23-Dec-2004, 14:58
That is, until the real games come out. I can only imagine what Temco, Sony's first party teams, Polyphony Digital, Konami will throw at us...
Well, Tecmo has supposedly been working on Code Cronus for 2 years already. It will be about 3 by the time Xenon is released. It's very likely that we'll see something very impressive at (or very near) launch. I suspect the game will be very much in the modern Tecmo tradition (Fighting or Hybrid-on-rails), so it ought to be pushing the polys pretty hard, including every 'Special Effect' they can think of. It will most likely be the best looking game on the planet for a few months.
Fafalada
23-Dec-2004, 15:10
I dont care to look up the flops and gips counts of the p4 or the ee, but my guess is the 3.8ghz p4 is still faster than a 200mhz ee.
That would be 300mhz, and care to tell me how that is even relevant? P4's clock wasn't frozen for the last 5years, or else you would be comparing EE to a 1.6Ghz P4...
Mind you, EE already had overclocking headroom back when it was at 180nm, and IIRC GSCube parts were in fact clocked at 500mhz (at 180nm as well).
90nm part could clearly clock higher yet, if there was any need for it.
Dave Baumann
23-Dec-2004, 15:31
I don't ignore them either, but I still debate them if I am not convinced... how else am I going to learn ;) ?
Of course, but dismissing things as "irrelevent" is not condusive to a learning process.
That would be 300mhz, and care to tell me how that is even relevant?
I'm astounded that people are still missing his point. Sony may have gotten their 90nm part out first, but it may well have also been a far easier part to develop and manufacture.
london-boy
23-Dec-2004, 15:42
That would be 300mhz, and care to tell me how that is even relevant?
I'm astounded that people are still missing his point. Sony may have gotten their 90nm part out first, but it may well have also been a far easier part to develop and manufacture.
We all got his point the second time he repeated himself. 83 times ago. :twisted:
I don't ignore them either, but I still debate them if I am not convinced... how else am I going to learn ;) ?
Of course, but dismissing things as "irrelevent" is not condusive to a learning process.
Oh, is that the case? :shock: IIRC (I won't bother to look up sorry) in a few pages before aaronspink basically wrote that all Sony's PR are full of hype and irrelevant to a serious discussion. 65nm - 45nm is a talk of future and can be a target of speculation but he extended it to 90nm and brought Intel in then wrote off 90nm facts as 'Sony hype', which is a very stereotypical and illogic jump IMO. Hope he learns something as I do in using Google :P
I do believe the question was concerning process technology and not the implimentation of it. Untill Aaron jumped in, namely:
Sony has a habit of embelishing their technology. Lets state a positive. I am positive that no one will beat Intel to a process node.
Which is unambiguiously false. Concerning "embelishing," this comment is something expected from PC-Engine or Deadmeat. Chipworks cleared this up well when they stated:
The combination of this dielectric process and the smallest transistor seen so far by Chipworks makes this one of the most advanced processes in volume production today.
<center>http://www.chipworks.com/mail/2004_02/sony_90nm.jpg http://www.chipworks.com/mail/2004_02/Intel_90nm.jpg
Sony CXD9797GB -- (via Chipworks) -- Intel "Prescott"</center>
And then concerning the TTM aspect, Sony had a 90nm fabricated product on sale almost 6 months before Intel came to market in Febuary of 2004. I'm still waiting for Aaron, whom Dave thinks we should all listen to and never question, to tell me what 90nm part from Intel I could buy in the fall of 2003. Anytime you're ready, just post away.
Concerning implimentation which IS irrelevent to the base technology discussion, I think Faf's answered this aspect quite well. The Emotion Engine is a static design, the subsequent work on the design was to maximize yeilds and profitability, not to maximize preformance.
Dave Baumann
23-Dec-2004, 16:02
I'm still waiting for Aaron, whom Dave thinks we should all listen to and never question, to tell me what 90nm part from Intel I could buy in the fall of 2003.
:roll:
I'm astounded that people are still missing his point. Sony may have gotten their 90nm part out first, but it may well have also been a far easier part to develop and manufacture.
People got his point, that's not the problem, what people didn't understand is the relevance of that point with regards to the subject, the 90nm processes from Sony and Intel.
Both EE+GS and the P4 Prescott from an architectural point of view are already in the Apple VS Orange land, pointing out another difference between the chips is not susceptible to change the discussion much, at this point.
Intel certainly think they were first (not that, that is any real evidence of course :-)
"Intel Pentium 4 processor on 90nm technology has the distinction of being the world’s first high-volume processor on the new technology"
http://www.intel.com/technology/itj/2004/volume08issue01/foreword.htm
So it looks like Sony claims they were first and Intel claim they were first. Bit of a stalement then so lets have some fun...
I conclude that both Sony and Intel chip engineers suck whereas many others are good.
Why?
I see no pretty chip art from either Sony or Intel, miserable gits :-)
http://www.chipworks.com/gallery/gallery_home.asp
Panajev2001a
23-Dec-2004, 16:48
I don't ignore them either, but I still debate them if I am not convinced... how else am I going to learn ;) ?
Of course, but dismissing things as "irrelevent" is not condusive to a learning process.
I do not dismiss them as irrelevant, I would not debate with them if I did.
Dave Baumann
23-Dec-2004, 16:51
...but the poster I replied to did...
that's getting boring..
Even industry insiders do mistakes, and even industry insiders can have a bad attitude sometimes..just like everyone.
Dave Baumann
23-Dec-2004, 17:08
Regardless of the exact minutia of the timings for individual availability of 90nm parts, the fundamental point is that they are likely to be projecting similar timescales for future process availability, and that’s the important facet of the discussion.
Regardless of the exact minutia of the timings for individual availability of 90nm parts, the fundamental point is that they are likely to be projecting similar timescales for future process availability, and that’s the important facet of the discussion.
True and the whole point of this thread is that Sony are very likely targetting their chipsets one process generation ahead of MS.
Intel certainly think they were first (not that, that is any real evidence of course :-)
"Intel Pentium 4 processor on 90nm technology has the distinction of being the world’s first high-volume processor on the new technology"
http://www.intel.com/technology/itj/2004/volume08issue01/foreword.htm
So it looks like Sony claims they were first and Intel claim they were first. Bit of a stalement then so lets have some fun...
No, Intel didn't claim they were first in 90nm.
"Intel Pentium 4 processor on 90nm technology has the distinction of being the world’s first high-volume processor on the new technology"
They only claimed "high-volume processor" was the first. Looking even this page (http://www.intel.com/research/silicon/nanometer.htm) still holds an erroneous description ("Intel began shipping products built on 90 nm technology in the fourth quarter of 2003"), it seems they strongly hoped 90nm to be a nice hype of them, but it didn't work because of delay.
Actually, Sony only follows Toshiba in process technology.
http://www.sony.net/SonyInfo/News/Press/200212/02-1203E/
TOKYO, December 3, 2002 -- Toshiba Corporation and Sony Corporation today announced the world's first 65-nanometer (nm) CMOS process technology for embedded DRAM system LSIs -- a major breakthrough in process technology for highly advanced, compact, single-chip system LSIs that will be only one-fourth the size of current devices while offering higher levels of performance and functionality.
...
Current system LSI devices on the market are produced with 130 nanometer process technologies. Toshiba, the recognized industry leader in advanced process technology, is the only company with mass production technology for 90nm process embedded DRAM system LSI, a technology it is currently deploying and that will meet ever increasing demand for more and more compact devices.
http://www.reed-electronics.com/electronicnews/article/CA270489.html
Toshiba Takes 90nm Process Lead
Online staff -- Electronic News, 1/13/2003
Toshiba Corp. today disclosed details of its 90nm process that could place it ahead in the race among chip companies to be the first to reach volume production at such an advanced technology node.
Toshiba has begun sampling its 90nm SOC process and aims to begin mass production in Q2, with high-volume production beginning in Q3. That could place Toshiba ahead of the likes of Intel, IBM and TSMC, all of which are expected to reach production at 90nm in the second half of 2003.
Since this thread is becoming another branch of Open PC Architecture vs. Closed Game-console/CE Architecture, I'd like to quote this interesting article about "digital consumer". Who benefits from iPod selling well? One of them is Toshiba, the manufacturer of the mini-HDD in iPod.
http://www.electronicsweekly.com/Article5700.htm
by David Manners Wednesday 28 April 2004
Japan's semiconductor firms build on their strengths
The Japanese are back. After the 'lost decade' of the 1990s when the Japanese semiconductor industry lost market share, under-invested and retrenched, the companies are now cock-a-hoop.
Japan's domestic market is booming, the local companies have cash to invest in new capacity, they have cutting edge process technology delivered by industry consortia and the electronics industry's macro trends are moving in their favour.
Last year, Japanese domestic semiconductor consumption overtook US domestic semiconductor consumption for the first time in over a decade, reflecting the change in shift from a PC dominated semiconductor market to a digital consumer led semiconductor market.
The PC industry was an US-dominated industry where Intel and Microsoft set the standards and took most of the profit with a host of companies supplying components and software on a 'horizontal' industry model.
The digital consumer industry is a Japan-dominated industry where the industry is organised on a 'vertical' model of companies, which each have total control of the product from chip design to chip manufacturing to end product assembly.
For the Japanese semiconductor industry, this return to favour of the integrated, vertical industry model is a harbinger of good times.
"The Japanese semiconductor industry is coming back," says Toshihiko Ono, president of Fujitsu's electronics devices business group, "and the reason is that with the 0.13µm and 90nm and future technology generations the fabless business model no longer works, in my opinion."
"For the last ten years, everything in the US and Europe followed the fabless/foundry, horizontal business model," says Sartoru Ito, president of Renesas Technology, the world's third largest semiconductor company, "now the industry is coming back to the vertically integrated business model."
"The market is shifting from a focus on the PC - which is an American dominated area - to a focus on digital consumer (DC) products where Japan is dominant," says Dr Tsugio Makimoto, chief technology officer of the networking group at Sony Semiconductor.
The effect of the PC to DC transition is that the vertically integrated semiconductor model will gain strength at the expense of the horizontal fabless/foundry model.
"As structures get smaller, mismatches between design and manufacturing magnify fluctuations in the outcome of the production process," explains Makimoto, "so design has to understand the capabilities of manufacturing which are constantly changing. So if manufacturing is not good enough, the design people can build in safety factors to take account of the limitations of the process. That's the benefit of the synergetic effect of having control over everything from design to manufacturing. This is why 0.13µm was delayed on the foundry side because of the miss-matching of the design and the manufacturing side."
IBM agrees. In February, Dr Tom Reeves, v-p and general manager of IBM's Asic division, said the semiconductor industry generally is getting a 50 to 60 per cent first-time-right rate for 0.13µm silicon, but the foundry industry is only getting a five to ten per cent first-time-right rate.
So, as structures get smaller at 0.13µm and below, the difficulties on matching design to manufacturing are increased, so favouring the vertical industry model and consequently favouring Japan. "The strength of Japan is in the vertical model because of our culture - valuing teamwork and paying attention to very small details," says Makimoto.
With the long reign of the PC as the major market for semiconductors coming to an end, the major semiconductor market for the next decade, or even two, will be DC products: digital TV, digital video recorders, DVD, camera-phones, games machines, digital cameras, camcorders, car navigation systems etc, which were first developed and marketed in Japan. Adding to the DC market is that products like mobile phone handsets, PDAs and notebook/handheld computers are becoming DC products.
The 'digital consumer' tag is a bit of a misnomer. "There's no difference between 'analogue' consumer products and 'digital' consumer' products," says Dr Susumu Kohyama, corporate senior v-p at Toshiba Semiconductor. "The quality of the consumer's experience comes mostly from the quality of the picture and the quality of the sound and those are totally analogue."
Whatever you like to call it, the market effect of DC is significant. Last year, for the first time, semiconductors sold for use in consumer, communications and automotive products were worth more than semiconductors used in computers, according to World Semiconductor Trade Statistics (WSTS). That has huge implications for Japan.
"There's been no major market in Japan for the last ten years," says Ito of Renesas, "now digital consumer is a major market segment and Japan is clearly leading that, and, from the semiconductor point of view, we have major customers here and we have the technologies to support them, such as high density packaging and low power."
"The digital consumer type of business is very different from the PC business," adds Ito, "the PC business is more like box-making with a clear standard from Intel and Microsoft. If you have a limited expertise in one area you can be part of it. It's characterised by fragmented expertise around a standard. Digital consumer has no standard. How you make the product attractive to consumers decides the competitiveness of the product."
Dr Hajime Sasaki, chairman of NEC, says: "The American players are coming from the PC side, the Japanese players are coming from the TV side. Going from a PC approach to a TV approach is more difficult than from TV to PC because of the structure of the market."
While the US industry will have to adjust to that structure; the Japanese industry already has it. A big advantage of the DC consumer products for the semiconductor industry is that they have much greater value of chip content than the old 'analogue' consumer products even if much of the value-added in DC products comes from analogue components.
DC products can use up to ten times more semiconductors than the old ones. Semiconductors in film-based cameras were worth an average $4, in digital cameras it's around $40. Semiconductors in analogue TVs are worth $20, in digital TVs it's $110.
In products like car navigation systems the semiconductor content, at around $100, represents 17 per cent of the product's end value. In mobile phones a chip content of $40 represents 20 per cent of end product cost.
So the new, chip-intensive, DC products will help drive semiconductor industry output and, with the current semiconductor upswing being driven by consumer products, the Japanese semiconductor industry has the advantage of having most of the world's consumer electronics giants on its doorstep.
Sony, for instance, uses $8bn worth of semiconductors a year. Others, like Matsushita, Toshiba, NEC, Hitachi, Sanyo, Fujitsu, Mitsubishi and a host of other Japanese producers of electronic consumer goods keep Japan at the top of the consumer electronics industry.
Collectively Japanese companies make over 80 per cent of the world's camcorders, over 70 per cent of the world's games machines and camera phones, and over 60 per cent of the world's digital cameras and car navigation units.
Supplying these companies with semiconductors could set the Japanese semiconductor industry on a 20 year-long roll.
"In the past, the technology and market-place in Japan, south east Asia, Europe and the USA were all different - consumer in Asia, communications in Europe and computer in the US - and Toshiba operated in all those technologies and markets. But now those technologies and markets are converging and so Toshiba can offer all those technologies and all that knowledge," says Toshiba's Kohyama.
Another factor driving the Japanese upswing is the effect of various industry/academia/government consortia over the past few years. "In the case of technology development, government investment helps to overcome Brick Walls," says NEC's Sasaki. "In the 1990s the semiconductor industry had a Brick Wall at 100nm. We had similar consortia in the 1970s when we had a Brick Wall at 1µm. Once it was overcome, the technology could be refined without basic changes."
In the 1990s, the US and the EU started all sorts of industry support activities, says Ito of Renesas: "Makimoto did a report to the Japanese government which resulted in the SNCC (Semiconductors in the New Century Committee). Makimoto was the central guy in that initiative."
The SNCC gave rise to a number of industry consortia in which the Japanese government matched industry financial contributions 50/50. One of them was the ASPLA (Advanced SOC Platform Technologies Activity) which helped companies make the transition into 300mm/90nm manufacturing.
Others were ASCA (Advanced Semiconductors through Collaborative Achievement), HALC (High Performance and Agilent Clean-Room Association) and MIRAI (Millennium Research for Advanced Information Technology).
"The consortia have played an important role in strengthening the basic technology," says NEC's Sasaki, "but at the same time we've seen a big improvement in the market with three key products - flat panel LCD/PDP TV, the DVD-Recorder, and the digital still camera."
A sparkling 2003 saw the Japanese domestic semiconductor market grow 40 per cent, and allowed the Japanese semiconductor companies to replenish their financial coffers giving them the opportunity to invest in new technologies and new manufacturing capabilities.
"Sony spent $1.2bn in 2003 - the fourth highest capital spend in the industry. Semiconductor investment is the highest priority in the company," says the firm's Makimoto.
NEC, with $1bn from its sale of shares and another $1bn gained from operations has a $2bn pot to invest in a new 300mm 90/65nm wafer fab.
Toshiba's semiconductor division generated $1bn of positive cash-flow last year and announced the ground-breaking on a new 300mm 90/65nm wafer fab in April.
Fujitsu, without the huge cash balances of Toshiba and NEC, has nonetheless come up with a strategy to build a 300mm 90/65nm fab. It has persuaded partners to contribute towards the cost.
The reason why those companies are ready to put up their money is because they get access to Fujitsu's 90nm technology - which claims to be the best in the world for low power. That's because it has solved the problem bedeviling everyone else's 90nm process - current leakage.
"We have achieved high speed with low leakage," says Fujitsu's Ono. "Data from an American customer shows that, in our 90nm process, the speed has been increased by 30 per cent and the current has been reduced by one fifth, compared to other foundries." 'Other foundries' includes, says Ono, IBM, TSMC and UMC.
Japan does not see the moves of the US computer companies into DC products as a threat.
"It's very cheap stuff - just assemble and ship," says Toshiba's Kohyama, "we can't yet know the impact on products like low-end flat panel TV, but it may accelerate the commoditisation of the product. I wouldn't want one in my living-room - it would spoil the atmosphere."
Ito of Renesas reckons: "I think the US will have some success in the consumer business but not to the extent that they succeeded in the PC world. They will be viable competitors, but not dominant."
Distribution may also be a problem for the US companies. "The US companies most likely won't have direct channels to the customer through stores - they will sell via the Internet," says Kohyama, "if customers require any more detailed explanation of products they can't get it."
Can the Japanese semiconductor industry recapture its glory days of the late 1980s when they had six companies in the world semiconductor top ten and over 50 per cent world market share? No one is predicting that. Yet. But the mood bubbling away in Tokyo is bullish.
BTW, and that's slightly off-topic, but the more i hear about next-gen games, from various statements to actual screenshots, and the less i believe we'll see games with more than 1M polygons per frame (@60Hz, of course), especially the first batch of games.
UE3 use more than 1.200.000 polygons, and TS said that it is runing great in the SDK ( 9800XT :?: ) and the shader instruction set is from 50 to 200
The R520 is completely unrelated to the xenon gpu.
I dont know if I would take it that far. It appears that this is true due to the 3D architectural split at R400/R420. However, 3D is not the only aspect of the GPU and I wouldnt be suprised if other areas of the chip made there way from the R400 line to the R420 line and vice versa. This would make them not so unrelated.
The R520 is completely unrelated to the xenon gpu.
I dont know if I would take it that far. It appears that this is true due to the 3D architectural split at R400/R420. However, 3D is not the only aspect of the GPU and I wouldnt be suprised if other areas of the chip made there way from the R400 line to the R420 line and vice versa. This would make them not so unrelated.
I think you are stretching a bit to show that they are related. Thats like saying an airplane and car arn't unrelated different because they both use tires (ignore sea planes pleaze).
Wow, That Electronics Weekly piece sounds familiar in tone...
Anyways, thanks for the article One. It was a delight to read. :P
Brimstone
23-Dec-2004, 23:24
"The market is shifting from a focus on the PC - which is an American dominated area - to a focus on digital consumer (DC) products where Japan is dominant," says Dr Tsugio Makimoto, chief technology officer of the networking group at Sony Semiconductor.
The effect of the PC to DC transition is that the vertically integrated semiconductor model will gain strength at the expense of the horizontal fabless/foundry model.
"As structures get smaller, mismatches between design and manufacturing magnify fluctuations in the outcome of the production process," explains Makimoto, "so design has to understand the capabilities of manufacturing which are constantly changing. So if manufacturing is not good enough, the design people can build in safety factors to take account of the limitations of the process. That's the benefit of the synergetic effect of having control over everything from design to manufacturing. This is why 0.13µm was delayed on the foundry side because of the miss-matching of the design and the manufacturing side."
Synergy, the new corporate culture buzzword for Sony? It could be just a coincidence "synergetic" was used, afterall it's a common word, but in light of the small cores in "CELL" being named Synergistic Processing Units I'm not so sure. Anyway I just found the use of that word amusing.
Inane_Dork
24-Dec-2004, 00:02
Every PR group uses "synergy" all the time. It's probably not worth your time to wonder about.
Brimstone
24-Dec-2004, 00:51
Every PR group uses "synergy" all the time. It's probably not worth your time to wonder about.
I've already returned to being a pro-active forum member. Thanks for the insight number two. I'll be seeing you.
overclocked
24-Dec-2004, 09:15
On the topic i think that it´s safe too assume that the "next gen gpu" from nVidia is built with 90nm in regards to cost and transistor budget.
That is for the pc-part ofcourse.
Now what i wonder is it crazy too belive that $ony has said to nVidia that you can add a couple of quads cause we are going to fab this at 65nm so we can "afford" a bigger die/relative to the pc-part that´s likely going to be built using 90nm at another foundry?
btw Merry Christmas to you all at B3D :D
Every PR group uses "synergy" all the time.
But still not as often as 'compelling'.
cthellis42
24-Dec-2004, 11:44
I see no pretty chip art from either Sony or Intel, miserable gits :-)
<laughs> Much love for Deano to be able to introduce much-needed levity and a fun site to boot. ;) One could even say it's "compellingly synergistic!"
...or not.
pahcman
24-Dec-2004, 12:17
Chipworks vs Silicon Insights
Chipworks: PSX Chip Is Too 90nm
Online Staff -- Electronic News, 2/4/2004
Technical services company Chipworks Inc. said today that Sony's Playstation X processor was manufactured with 45nm gates, leading it to believe that it is indeed made with a 90nm process.
Another chip market research company, Silicon Insights, claimed that Sony's EmotionEngine and Graphics Synthesizer chip, found in its latest Playstation game console and which Sony says is built on a 90nm process, was actually built with a 0.13-micron, or 130nm process technology. Silicon Insights said it removed the chip from a PSX model DESR-5000, and determined it was built on 130nm technology with a die size of 90mm square.
But Chipworks announced on its site today that after reverse engineering the EmotionEngine chip, that the device has gate lengths more consistent with a 90nm process. Chipworks even has a cross section SEM image on its site of what it says are the gates of the Sony chip in question.
"Chipworks obtained a sample of the Sony PSX chip and had started to take it apart based on the reports that it was a 90nm part, but we have come to somewhat different conclusions than those reported last week," Dick James, Chipworks' senior technology analyst, said in a statement on the company's Web site. "For one thing, we found transistors with a physical gate length of 45nm to 50 nm, which immediately led us to believe that we did have a 90 nm part."
Chipworks conceded that the Sony chip, fabricated in partnership with Toshiba, while perhaps not meeting the exact definition of a 90nm process as spelled out in the International Technology Roadmap for Semiconductors, is nevertheless comparable with the leading edge Intel Corp. process in terms of gate dimension. Intel 90nm chips, as announced at last autumn's Intel Developer Forum, reportedly have gate lengths of 45nm, slightly larger that the 37nm polysilicon gate length considered the benchmark by the ITRS for the 90nm node.
The EmotionEngine chip also utilizes an advanced two stack low-k dielectric structure, according to Chipworks. This, combined with the 45nm gates, makes the chips one of the most advanced in production today, the company maintains.
Chipworks observed that the ATI RADEON 9600XT graphics processor, built by foundry Taiwan Semiconductor Manufacturing Co., like the Sony chip uses Applied Materials' Black Diamond low-k intermetal dielectrics, but with 85nm gates. This indicates it is clearly a 0.13-micron device, according to Chipworks.
Semi Insights stands by 'not 90-nm' description of PSX chip
By Peter Clarke
Silicon Strategies
02/05/2004, 12:11 PM ET
LONDON -- Despite seemingly contradictory photographic evidence put forward by a rival engineering company, Semiconductor Insights has said it stands by its claim that the manufacturing process used to fabricate the processor for Sony's Playstation X entertainment system, is not a 90-nm process.
On Wednesday Chipworks Inc, a rival Canadian engineering consultancy, had claimed it had found transistors on the processor with physical gate lengths considerably smaller than those Semiconductor Insights said it had found, and that this was indicative of a leading-edge process. At the same time Sony continued to insist its Emotion Engine and Graphics Synthesizer, or EE+GS, processor was being built on a 90-nm manufacturing process, as it had always claimed.
The previous week Kanata, Ontario-based Semiconductor Insights had said the smallest physical gate lengths it had found, after sampling multiple sites on the chip, were of 70-nm, and that this was slightly larger than the 65-nm the International Technology Roadmap for Semiconductors (ITRS) equates to a 130-nm manufacturing process.
This, together with a metal-one pitch that Semiconductor Insights measured as being closer to the ITRS 2003 definition for a 130-nm process than a 90-nm process, and similarly relaxed embedded DRAM measurements, had persuaded the firm to go public with its claim (see story).
At the time Edward Keyes Semiconductor Insights chief technology officer said: "It's clear that it's a 130-nm chip, not a 90-nm chip, as defined by the ITRS,"
But on Wednesday (February 4, 2003) Ottawa-based Chipworks published a photograph on its website identified as showing a cross-section of deep-submicron transistors within a CXD9797GB, the part number stamped on the package of the EE+GS processor. On the photograph two transistors are marked as having physical gate lengths of 46.7-nanometers and 47.5-nanometers, respectively.
Although Chipworks acknowledged these did not meet the ITRS 2003 definition for a 90-nm process, the engineering firm observed that the 90-nm manufacturing process from Intel Corp. is reported to have a gate length of 45-nm, barely different from the smallest gate lengths found in the Sony PSX processor.
When asked by Silicon Strategies on Wednesday (February 4, 2004) if Semiconductor Insights could have missed some more aggressively scaled transistors in its original measurements, Edward Keyes, chief technology officer of Semiconductor Insights said that Sony has never claimed to have such small geometry 47-nm transistor gate lengths in its 90-nm process.
Sony's published gate length claims for the CMOS4 process and the ASC9 embedded DRAM process match Semi Insights' findings of 70-nm, Keyes said. Semi Insights has confirmed those measurements and pointed that is not a 90-nm process by the ITRS tables, Keyes added.
Keyes acknowledged it is not possible to measure every transistor but he also said it is not necessary, as transistors don't come in an infinite number of sizes.
"Engineers don't have the freedom to alter the gate length," said Keyes.
Silicon Strategies asked if it was possible that different logical gates, such as different versions of NAND, with different fan-outs or with different drive strengths, selectable from a library, might result in transistors which when finally etched and diffused in silicon have different physical gate lengths.
"If half of one percent of the gates are at some extreme point does that make the chip belong to the next process node? I would say no," Keyes said in answer.
Reasons why measurements may differ
Keyes said he was reluctant to comment on the Chipworks scanning electron microscope photograph as he did not know how the microscope had been set-up to take the photograph. "There are a couple of reasons you might get different measurements though."
Keyes said that typically gate polysilicon is etched back in preparing the sample. This produces sharp edges that will emit electrons strongly. This can make the surrounding buffer oxides appear thicker than they really are, and make the gate appear shorter.
The second possibility is the geometrical effect of taking an end-of-gate slice, rather than going through the middle of the transistor, when cross-sectioning the chip.
Rather like slicing though the edge, rather than the middle of an orange this has the effect of making the polysilicon gate (the flesh of the orange) look narrower and the oxides (the pith) appear wider.
When asked how it was possible to know whether a sliced transistor was a good representation of the gate length or an end-of-gate aberration, Keyes said: "It comes down to the law of averages. If 99 percent are the transistors are constant and there's a few 30 percent smaller you discount it."
Keyes acknowledged that he would expect end-of-gate slices to vary and not to show identical reduced measurements.
Keyes added that Semiconductor Insights assessment had not just rested on the lack of sub-70-nm gate length transistors but also on the metal-one pitch which had measured at 260-nm. Keyes said the ITRS definition gives 210-nm for a 90-nm manufacturing process and 295-nm for a 130-nm process.
Chipworks findings
Dick James, senior technology analyst for Chipworks, said that although the two transistors Chipworks had photographed were measured to be below 50-nm in gate length he could not identify where in the EE+GS processor the sub-50-nm transistors had been found. This, he said, was because, until an example chip is delayered, the characteristic transistor layout remains hidden under upper metal layers.
James also said he believed that the transistors Chipworks had found were not end-of-transistor anomalies, because several transistors in a line had produced similar measurements in each transistor. If an end-of-transistor anomaly was being measured you would expect the measurements to vary, he said, unless the cut happened to hit the arc of each transistor-end in exactly the same place.
James also said that Chipworks was familiar with edge-effects in samples prepared for scanning electron microscopy by etching. "There is a bit of distortion depending on beam energy. But when you're looking at a gap like this I've got more confidence. Our SEMs are calibrated to within plus or minus 5 percent," said James.
"We did see wider polysilicon lines. But we see p-channel transistors at 48 to 49-nm and n-channel transistors at 46 to 47-nm. The embedded DRAM is more relaxed but you'd expect that."
After studying Chipworks' photograph Semiconductor Insights' Keyes wrote an email to Silicon Strategies saying: "It's apparent that the line-end effect that we talked about is not occurring in this cross-section and hence cannot be the explanation for their smaller measurement. We do still stand by our original statement that this is not 90-nm technology according to the roadmap and we will be investigating further."
Who right who wrong? Who side are you on? Hmmm....
But deno say it best, no pretty chips everywhere. :lol:
pahcman
24-Dec-2004, 12:40
One cool thing, Semi Insights gave PS2 GS an award for impressive SoC remember? Some Sony fan always quick to point out this award to highlight Sony exellence in enigineering.
So is Semi Insights reliable or not?
Business Editors & Technology Writers
OTTAWA--(BUSINESS WIRE)--April 11, 2001
Semiconductor Insights is pleased to announce that it has awarded Sony Computer Entertainment (SCE) with an INSIGHT Award for the Most Innovative Computing Semiconductor Product of the Year for the Sony Graphics Synthesizer(TM) from the Sony PlayStation(R)2 game console.
"System-on-a-Chip (SoC) design has been a buzzword in the semiconductor industry for the past 5 years. Sony has developed a chip that exemplifies the philosophy of SoC as it not only integrates many functions on-chip but also integrates different technologies," said Derek Nuhn, Chief Operating Officer of Semiconductor Insights. "The extreme performance of the resulting chip is phenomenal."
"This is the first year Semiconductor Insights has given out the INSIGHT Awards and we plan to make it an annual tradition," Nuhn said. "Semiconductor Insights performs competitive technical intelligence on devices in every technology area each year putting us in the unique position of analyzing the true technical merits of leading edge semiconductor products."
The Sony Graphics Synthesizer is used in the PlayStation2 game console and can handle 20 times the simultaneous data of some PC graphics accelerators and can render 75 million polygons per second. The device is implemented in a sophisticated DRAM process with high performance logic in a single die. Some game reviewers describe this chip as "wicked fast."
"PC technology has generally led the advancement of process geometries in semiconductors," a senior analyst indicated. "Sony's Graphics Synthesizer has a logic core with embedded DRAM. The embedded DRAM uses sophisticated sense amplifiers that contain an SRAM latch and two read data buses. This combination of DRAM and SRAM is intended to improve the speed of the video memory for better drawing performance."
Semiconductor Insights' analysts also discovered that although Sony's Graphics Synthesizer was likely implemented in a 0.18 micron process, transistor sizes were found to be as small as 0.1 microns. Sony uses a process approach that results in near-perfect alignment of the transistors and capacitor contacts.
The Sony Graphics Synthesizer was chosen amongst other innovative products nominated by Semiconductor Insights' analysts, including such devices as the Intel Pentium 4 Processor, the AMD Athlon Processor and the nVIDIA Geforce2 Graphics Processor. These runner-up devices also represent true innovations:
Intel Pentium 4 Processor -- This device is developed on Intel's 0.18 micron process and has transistor gates as small as 0.1 microns. Intel uses a "notched" gate to achieve the smaller transistors that also results in other benefits in the process. The version investigated by Semiconductor Insights runs at 1.5 GHz -- over twice as fast as the Pentium III examined in early 2000. Intel is using low-k dielectric in their higher speed processors to reduce capacitance between metal lines and allow for faster operation. Most other manufacturers currently use copper for high speed operation.
AMD Athlon Processor -- AMD is implementing their 1GHz and higher processors in both aluminum and copper probably to verify and test their copper process. Semiconductor Insights has analyzed the copper version of the Athlon. Like the Intel Pentium 4 Processor, AMD has transistor gates as small as 0.1 micron.
NVIDIA Geforce2 Graphics Processor -- nVIDIA is a fabless company that uses TSMC as its foundry for manufacturing integrated circuits. Developed on a 0.18 micron process, the Geforce2 has more functionality that previous generations, including transform and lighting functions, leading to faster process speeds, dynamic per pixel shading and videoprocessing that, in conjunction with a receiver, allows for computers to view HDTV signals.
Semiconductor Insights congratulates Sony Computer Entertainment and their design team for demonstrating innovation and implementing a design that advances the development of semiconductor technology.
:D
Spidermate
24-Dec-2004, 13:47
On the topic i think that it´s safe too assume that the "next gen gpu" from nVidia is built with 90nm in regards to cost and transistor budget.
That is for the pc-part ofcourse.
Now what i wonder is it crazy too belive that $ony has said to nVidia that you can add a couple of quads cause we are going to fab this at 65nm so we can "afford" a bigger die/relative to the pc-part that´s likely going to be built using 90nm at another foundry?
btw Merry Christmas to you all at B3D :D
That may not be far from the truth. They have approached problems many times like this in the pass.
And no, I don't believe these GPUs will be anything alike as what most people may think. I have my own reasons for this, but I'll leave it at that.
london-boy
24-Dec-2004, 15:18
One cool thing, Semi Insights gave PS2 GS an award for impressive SoC remember? Some Sony fan always quick to point out this award to highlight Sony exellence in enigineering.
So is Semi Insights reliable or not?
:D
Why would you bring that up? It even mentions the 75million polygons figure many people cry at...
But yeah, the PS2 won some awards. It was after all a very different and powerful machine compared what was available at the time.
I don't think a P3-733Mhz and a modified NV25 could deserve awards, although i think the NV2A did win something.
L_i_n_k
25-Dec-2004, 12:20
Just wondering that.
-What information Vertical (Inhouse Design and Process model) Can use in its designs, Witch cannot be aplied to horisontal designs to Overcome its process limitations ,when only chip designs are considered?.
-If we see gadgets (mobilephoes/multimedia carrycadgets,Homestation boxes/media centers/"consoles") to replace pc with their vertical model in the future. How is this progression related to manufacturing process race between Vertical and horisontal model. Is`nt this broblem to be more in the practical/Physical design area not in the information deployment clitches, that prevent horizontal designers from taking the full use of given process in their designs. Could this phenonem be just a gimmic in present time becouse there is not good enought tools to share information between process/design or the Information sharing progress has in time started to lack behind the vertical model (Ip/money issues/?)?.
-By the way of topic!!! what did Aaron mean by "I have stop been bitter a long time a go". And something about , i have understanded the meaningless of understanding". I Cant remember latter becouse i did not found it (deleted?) I understand bitter part well. Did not understand the understanding part, what did it mean?.(Fear fades a way only throught understanding). I dont want to spoil these threads but im just too damn curious :D
Sorry for my rare posting, you guys just allways get to the end of discussion before i even have time to read the thread to its closedness. The intire reading process is mostly about learning to me. you guys seem to know a damn lot , Good Xmass to all ,it has been fun to be your apprentice, Less pride greating bitternes :wink:
L_i_n_k
25-Dec-2004, 12:34
As structures get smaller, mismatches between design and manufacturing magnify fluctuations in the outcome of the production process," explains Makimoto, "so design has to understand the capabilities of manufacturing which are constantly changing. So if manufacturing is not good enough, the design people can build in safety factors to take account of the limitations of the process. That's the benefit of the synergetic effect of having control over everything from design to manufacturing. This is why 0.13µm was delayed on the foundry side because of the miss-matching of the design and the manufacturing side."
That was targeted mainly for this problem in horizontal model. Is`nt fluctuations mainly a quantum mechanical broblem at its core.Meaning that when horizontal model designers understand these fluctuations (current passes even super resistive material in scales below 10nm) to degree of vertical mode (Fluctuation information sharing between desing/process) the broblem with horizontal model could be solved?.
vBulletin® v3.8.4, Copyright ©2000-2010, Jelsoft Enterprises Ltd.