32nm process for ~2009

Acert93

Artist formerly known as Acert93
Legend
http://www.tomshardware.com/hardnews/20050706_161230.html

Intel has entered a joint development agreement with Corning to develop ultra low thermal expansion (ULE) glass photomask substrates required for Extreme Ultraviolet (EUV) lithography and semiconductors with transistor structures smaller than 40 nm.

EUV has been discussed to be key for next-generation production techniques for several years. After launching the first EUV lithography unit in August of last year, Intel slowly is taking the technology out of its labs and prepares it for real world use. First chips using EUV are expected to be available in 2009.

Currently, semiconductors are built using a 193 nm wavelength of light to print transistors in processes as small as 80 nm in mass production and 65 nm for next-generation chips. According to Intel, 193 nm will lead to transistor sizes as small as 50 nm. EUV lithography will take over at this point and will allow manufacturers to create smaller transistors in the future as the technology uses a wavelength of 13.5nm.

My pontificating...

90nm - 2004
65nm - 2005/2006
45nm - ?
32nm - 2009?

"First chips using EUV are expected to be available in 2009."

My guess (based on typical chip transition times with processes changes... still a guess) is that if Intel can avoid any significant snags is that mainstream/mass market 32nm chips would be seen early 2010, and maybe GPUs on the 32nm process by the end of 2010/early 2011.

Of course EUV could first be used for 45nm. This would lineup well with the fact process shrinks are taking longer and are becoming more and more costly. If 45nm chips were introduced in 2009, that would indicate that chips manufactured on the 32nm process would be in the 2011 timeframe.

I also believe ~2010 is Intel's goal to have 400mm wafer production lines active.

Finer Process + Larger wafers = A Good Thing (TM)

It is difficult to project how this could affect the consoles. My guess is that even if a process smaller than 32nm was available in late 2011, I have a hard time believing console makers would use it [with an aggressive chip size] due to yield issues on a new process and that it would mean at least another 2 years before they could shrink the chip.

It seems console makers prefer the safe route of using a proven process (like 90nm) and use an initially large chip, only to reduce its size sometime shortly after launch. Less risk and more control on pricing for a price sensative product.

On the other hand, if 45nm becomes available in 2009/2010, with 32nm 2 years off... egad! :cry: Process shrinks usually allow less heat/power consumption, higher frequencies, and more transistors (varying combinations of each based on chip design of course). If 45nm is the target process next next-gen, then I think we will begin seeing more multi-die solutions of some type, unless these smaller processes allow some significant jumps in frequency (which I doubt because as you get smaller there is bound to be more leakage).

Maybe we will have to wait until 2012 for the next round of consoles? With PS3 coming in 2006 (rough 6 years after PS2's launch), maybe we wont see PS4 until 2012?

Anyhow, I thought the above news was interesting because it gives us a sneak peak at what may be possible in 5-6 years.
 
Wow...139 Views and no replies...ummm..hi...I like rice pilaf....

So, you already stated the obvious things that benifit this when the process gets smaller and smaller...I wonder at that point what would Handhelds be capable of as far as CPU and GPU performance.
 
I say we'd see it late 2010 mabye even 2011 . The leaking seems to be a huge deal on 90nm and lower right now and it looks like everyone hit this big snag again. Intel hit it , then amd , ati is having problems with the r520 . It also looks like they are all having trouble with 65nm too .
 
Acert93 said:
http://www.tomshardware.com/hardnews/20050706_161230.html

Intel has entered a joint development agreement with Corning to develop ultra low thermal expansion (ULE) glass photomask substrates required for Extreme Ultraviolet (EUV) lithography and semiconductors with transistor structures smaller than 40 nm.

EUV has been discussed to be key for next-generation production techniques for several years. After launching the first EUV lithography unit in August of last year, Intel slowly is taking the technology out of its labs and prepares it for real world use. First chips using EUV are expected to be available in 2009.

Currently, semiconductors are built using a 193 nm wavelength of light to print transistors in processes as small as 80 nm in mass production and 65 nm for next-generation chips. According to Intel, 193 nm will lead to transistor sizes as small as 50 nm. EUV lithography will take over at this point and will allow manufacturers to create smaller transistors in the future as the technology uses a wavelength of 13.5nm.

My pontificating...

90nm - 2004
65nm - 2005/2006
45nm - ?
32nm - 2009?

"First chips using EUV are expected to be available in 2009."

My guess (based on typical chip transition times with processes changes... still a guess) is that if Intel can avoid any significant snags is that mainstream/mass market 32nm chips would be seen early 2010, and maybe GPUs on the 32nm process by the end of 2010/early 2011.

Of course EUV could first be used for 45nm. This would lineup well with the fact process shrinks are taking longer and are becoming more and more costly. If 45nm chips were introduced in 2009, that would indicate that chips manufactured on the 32nm process would be in the 2011 timeframe.

I also believe ~2010 is Intel's goal to have 400mm wafer production lines active.

Finer Process + Larger wafers = A Good Thing (TM)

It is difficult to project how this could affect the consoles. My guess is that even if a process smaller than 32nm was available in late 2011, I have a hard time believing console makers would use it [with an aggressive chip size] due to yield issues on a new process and that it would mean at least another 2 years before they could shrink the chip.

It seems console makers prefer the safe route of using a proven process (like 90nm) and use an initially large chip, only to reduce its size sometime shortly after launch. Less risk and more control on pricing for a price sensative product.

On the other hand, if 45nm becomes available in 2009/2010, with 32nm 2 years off... egad! :cry: Process shrinks usually allow less heat/power consumption, higher frequencies, and more transistors (varying combinations of each based on chip design of course). If 45nm is the target process next next-gen, then I think we will begin seeing more multi-die solutions of some type, unless these smaller processes allow some significant jumps in frequency (which I doubt because as you get smaller there is bound to be more leakage).

Maybe we will have to wait until 2012 for the next round of consoles? With PS3 coming in 2006 (rough 6 years after PS2's launch), maybe we wont see PS4 until 2012?

Anyhow, I thought the above news was interesting because it gives us a sneak peak at what may be possible in 5-6 years.

I expect PS4 to be on a process no larger than 32nm / 0.032 micron and arrive sometime around 2012. could be as soon as 2011, or as late as 2013, but 2012 seems the most likely.

I am less certain as to when the Xenon / Xbox 360's successor will arrive.
Microsoft is launching 360 exactly 4 years after Xbox. I think with 360, they want at least 5 years out of it, before they'll introduce another Xbox.
so that itself says 2010 for 360's successor, but if Sony isnt coming out until 2012, will Microsoft launch 2 years earlier ? who knows at this point
 
Back
Top