|
Don’t mention the ‘K’ word! |
|
|
|
Dec 13, 2006 at 05:12 PM |
IBM and AMD came out with a press release concerning some of process R&D efforts at the 45nm node yesterday. Normally this would prompt a story from us on this but I am hesitant to give it an airing.
The key reason for this is the fact that it is full of subtle marketing spin without any facts and figures. That might be okay for microprocessor release that happens every couple of weeks, but to claim a process breakthrough or claim a performance improvement etc... from a process point of view, you really need to back it up.
Of course that is supposed to be going on this week at the IEDM conference, which this press release is associated with, but for once, I am not at IEDM and don't have access to the papers, and neither does anyone else, unless you are attending the conference.
Some of the data being made public at the conference within the press release seems to me an obvious thing to do, but it's simply not there.
Why is this important?
Well, the saga over low-k dielectrics has been a long one and IBM has been at the centre of it since the start. Now we have a press release claiming an ‘ultra-low-k porous material,' yet no indication is given of the Young's Modulus numbers, let alone the effective k-value.
A story yesterday by Mark LePedas over at EETimes noted that the k-value was said by IBM researchers to be 2.7. What that figure represented wasn't clear, but what is clear to many of us is that a k-value of 2.7 has never been called ‘ultra-low.'
Nitpicking?
Not at this level.
|
Blogs 
