View Full Version : TSMS beats IBM in low-k process
Taiwan Semiconductor Manufacturing Company (TSMC) may begin volume production on a low-k dielectric process on its 0.13-micron process in the third quarter of this year, gaining an important technological lead over rivals, most notably, IBM Microelectronics, sources said.
Will we see loci and nv40 with this new technology?
http://www.digitimes.com/NewsShow/Article.asp?datePublish=2003/06/23&pages=02&seq=3
silence
06-23-03, 07:51 AM
LOL....is that like one year old news??
nv30 was supposed to use that process and we all know how that ended:angel:
I'm almost certain the NV40 isn't headed to TSMC, and i'm fairly sure they weren't planning on using LowK, unlike the NV30. I was also under the impression that the bigger, more complex chips, don't benefit so much from the LowK process. Could be wrong there though.
Hanners
06-23-03, 08:37 AM
Originally posted by jpeter
Will we see loci and nv40 with this new technology?
I don't think either chip was designed with low-k in mind after all the problems fabs have had getting the process working, so I doubt either of them will use it.
R.Carter
06-23-03, 01:23 PM
Actually, it's my understanding that Nvidia is splitting its orders with 20% going to IBM and 80% to TSMS. So IBM is more of a backup foundry than for them anything else.
Who knows if this will continue to be the case with new products.
They're not going to mix and match foundaries for 1 chip (NV40, for example). Too many headaches.
silence
06-23-03, 03:02 PM
Originally posted by R.Carter
Actually, it's my understanding that Nvidia is splitting its orders with 20% going to IBM and 80% to TSMS. So IBM is more of a backup foundry than for them anything else.
Who knows if this will continue to be the case with new products.
yap...on whole range...but i bet top card will be IBM only ;)
Actually, while for now it'll be 20/80, nVidia expects it to become 50/50 in a few years.
Uttar
extreme_dB
06-23-03, 09:11 PM
That is too funny. Nvidia has been cursed lately! They switch to IBM for the NV40 to avoid TSMC's problems, then TSMC clears them up while IBM starts having trouble! Unbelievable.
Their management must be fuming.
Well, IBM also got SOI that TSMC doesn't have. So really, it balances out, IMO.
No idea if SOI will be sufficently matured for the NV40 though...
Uttar
Ninja Prime
06-24-03, 05:36 AM
I thought IBM uses SOI exclusively on all .13 micron and under processes? Haven't they had SOI in most of their chips for almost a year now? I know that the G5/PowerPC970 is on a .13 micron process with SOI, and it's already in production, with systems using it to be in retail september 1st.
GlowStick
06-24-03, 09:03 PM
I belive when they say a 20/80 split they mean that from a Total of every chip Nvidia orders, but it would pre pritty insnae to have 2 fabs make the same chip, unless somehow a chip was in such demand that TMSC coudlnt supply enough, but TMSC can pump out so many chips thats pritty much impossible...
vBulletin® v3.7.1, Copyright ©2000-2012, Jelsoft Enterprises Ltd.