TSMC May Initiate Production of Low-Power 28nm Chips Shortly - Estimates

[Unknown Soldier]

The issues with Taiwan Semiconductor Manufacturing Company's 40nm and 32nm fabrication technologies have caused a massive gloom around the industry in the recent quarters. But according to some optimistic analysts, TSMC plans to finally initiate production of low-power chips using 28nm process technology.

A report from China Economic News Service claims that TSMC plans to start production of Altera's field-programmable gate arrays (FPGAs) in Q4 2010 using a 28nm process technology, which will mostly likely be 28LP, which uses previous-generation materials, such as poly gate and silicon oxide nitrate. Interestingly, but even if optimist market watchers are to be believed, production of 28LP chips seems to be behind the schedule.

According to claims by a senior executive of TSMC back in February, the company intended to start producing chips using 28LP in Q3 2010, not sometimes in the fourth quarter.

“The first node we are going to release for the 28nm will be called the 28LP. This is our poly gate and silicon oxide nitrate version. We will establish production at the end of June this year, about four months from now, and this is for the low power application," said Shang-Yi Chiang, senior vice president of R&D at TSMC at TSMC Japan Executive Forum in Yokohama.

Even though technically TSMC is ahead of all companies making various types of processors or logic in terms of 28nm node, its 28LP process not only uses old materials, but even with this one TSMC is late by one quarter from its own schedule.

TSMC did not comment on the news-story.

Source: http://www.xbitlabs.com/news/other/...f_Low_Power_28nm_Chips_Shortly_Estimates.html

Good news I suppose.

 

[Alexko]

Sure, but it says little about 28nm HP, which is what GPUs will use.

 

[Squilliam]

I wonder how far behind 28nmHP will be. Perhaps the reason why we didn't see the 67xx on the roapmaps was because it was intended as an early 2011 pipecleaner part?

 

[Arun]

xbit is wrong, Altera uses 28HP: http://danielnenni.com/2010/09/26/tsmc-yields-28nm/
This is about the expected timeframe I think, but it's easy to be (very!) confused when everyone uses a different nomenclature.

 

[3dilettante]

28HP is the HKMG gate-last process, right?

 

[Arun]

Yeah, HKMG direct successor to 40G. The naming is a bit confusing on the question of triple gate oxide though; on 28nm, that's theoretically 28LPT or 28LPG instead of 28LP if SiON, but I don't think anybody bothers pointing out the difference. And with HKMG, there's the 28HPM process that certainly uses triple gate oxide, but my guess is 28HP can also use it (which is what Altera probably uses). Yay, confusion ahoy! Feels like the good ole days of TSMC temporarily calling a 40nm process as 45GS so as not to tip their hand (and before anyone tells me it's the other way around, what do you call 45LP? a 50nm process?)