BRIGHT SIDE OF NEWS About | Advertise | Contact BSN USER Login
| Register
SUBSCRIBE Newsletter | RSS Feeds
Saturday, May 18, 2013
Email this to a friend.
Your friend's e-mail:
Your Name:
Your e-mail:
Message subject:

by: Anonymous on 10/5/2009
Is that 100nm guy joking?
by: Anonymous on 10/4/2009
It's pathetic. AMD being the money whores they are always trying to gyp us on computing power. I bought a CPU a few years ago (don't remember exactly with) with over a hundred nanometers. CPUs today have 65 if you're lucky, and usually 45 now, and it's moving to even less and less, while they claim the CPUs are somehow more powerful with LESS NANOMETERS.

It's pathetic. I refuse to buy another processor. I'll enjoy my 100 nanometers while you guys get ripped off generation after generation.
by: Kakkoii on 10/1/2009
Ah well I was speaking purely AMD's GPU's. Not their CPU's.
need go to to dentist by: Anonymous on 10/1/2009
i need to go to the dentists today.... i hate the dentists! :-(
GPUs arn't SOI by: Anonymous on 9/30/2009
Kakkoli -

GPUs use Bulk silicon, not Silicon On Insulator. SOI is for AMD's (and IBM's) CPUs. 32nm SOI will not be used, according to AMD's road maps, until 2011 and Bulldozer, so the article is correct in that regard. Your guess is as good as mine why, if yields seem good, they don't push an AM3/1207 32nm quad or hexa core on the mainstream so they could at least eat a little less money on the CPU front, and perhaps become competitive with Lynnfield/Bloomfield/Westy cores. It would seem like a good idea considering Intel's not going to do a 32nm 'Field" quad core, and Gulftown will cost an arm and a leg. This gaping hole until Sandy Bridge could lend AMD to be competitive if they chose to do a straight shrink, but I don't see them capitalizing on it, at least not from any road maps I've seen. In my mind, that's an epic fail on their part.

As for 32nm bulk silicon...Does anyone know WTF is going on with that? They said SOI got 'delayed' because AMD didn't need to use it until their next-gen arch, which makes sense since almost noone uses SOI. IBM's bulk 32nm OTOH, that GF and Chartered both are going to be using, seems up-in-the-air. Last I heard they both were going to take orders Q409 (late this quarter), which does lend itself to ATi transitioning over for a Evergreen refresh, but news on that front has gone silent, with only talk of 32nm SOI being late (which it always was supposed to be ~6 months after bulk). If it too is delayed, ATi might stay at TSMC so it doesn't lose it's process adoption lead versus nvidia.

We know ATi WANTS to use ATIC's fabs for 32nm bulk, as IBM's process tech at any given node is superior to TSMCs, but the question remains if and when it will launch compared to TSMC, which AFAIK, is still on track for early next year. Both plan for a 28nm optical shrink (half-node) H210.

Irregardless of what happens with the refresh, you can bet your ass AMD's next-gen (r9xx) will use ATIC fabs. Be it through 32nm, or 28nm (which they can take a 32nm design and adapt easily to 28nm when the process is available.)
by: Kakkoii on 9/30/2009
"AMD won't be ready with their 32nm SOI microarchitecture before 2011"

Huh?

I highly doubt ATI is going to wait 15 months to do another die shrink for it's parts. They have already been on 40nm for a while, so they will without a doubt move to 32nm in 2010. And in all likelihood so will NVIDIA, albeit perhaps a bit later in the year.
© 2009 - 2011 Bright Side Of News*, All rights reserved.
© 2009 - 2013 Bright Side Of News*, All rights reserved.