|
Make it so
|
04/08/2005
|
| |
Given there is so much talk about design for manufacture in chip design circles right now, it’s hard to pin down anyone who actually likes the term. After all, shouldn’t DFM really stand for design for manufacturability and isn’t that just another way of saying design for yield?
But prepare yourself for another surprise. Whilst eda tool vendors may be investing millions in R&D, and launching and developing tools that rise to the DFM and DFY challenges, they don’t think that many need worry about buying them just yet. And when they do start marketing this software outside the power user community, their hope is the DFM elements will be invisible.
As Joe Sawicki, vice president and general manager of Mentor Graphics’ Design to Silicon division, puts it: “It is a fundamental belief of mine that if you can avoid giving the designer a DFM tool, that should be your absolutely highest priority.”
David Thon, group director for Cadence Design Systems’ Nanometer Analysis and Verification activities, makes another important point. “Ask yourself what process you’re designing for now. Is it 130nm? The research is clear that 130nm will still account for most design activity for next year and probably beyond. If you are at 130nm, then this isn’t an issue. If you are looking at 90nm, that’s yielding well, so it is not so much of a problem there either.”
|
| |
Author
Paul Dempsey
|
| |
| |
|
| Email this article |
|
|