( ESNUG 459 Item 11 ) ------------------------------------------- [12/14/06]
Subject: Gary Smith on lawsuits, Wall St., Mentor, Cadence, DFM, and 2006
> EDA is the Rodney Dangerfield of high tech. "We don't get no respect!"
>
> - Vic Kulkarni, CEO of Sequence
> from http://www.deepchip.com/gadfly/gad102506.html
From: Gary Smith <gary=user domain=garysmitheda spot mom>
Hi, John,
It's getting close to the end of the year and I thought I'd summarize 2006
in EDA terms.
For 2006, except for the return of revenue growth, the EDA industry hasn't
had a stellar year this year. As Vic Kulkarni once wrote, EDA has become
the Rodney Dangerfield of high tech.
For the first time in 12 years, I'm having conversations with semiconductor
houses about developing all of their own IC CAD tools in-house. "It works
for the FPGA vendors; why shouldn't it work for us?", is their thinking.
Vic is right on target with one point. Until we, as an industry, start
respecting one another (and stop all marketing-driven lawsuits) we aren't
going to get much respect from our customers.
We also desperately need to get a road map so we can stop delivering our
tools 2 to 4 years behind the curve. This way runaway development of
in-house tool development should stop; if the semi houses know a certain
tool is definitely coming from the commercial EDA vendors, they won't
bother wasting their own R&D dollars on duplicating it.
Fortunately, we did see some improvement in 2006. Only 36% of the engineers
I surveyed reported that they used in-house tools; down 2% from 2005.
This was a year of confusion. There's why there were no big mergers nor
IPOs this year. Wall Street didn't know what to make of EDA. Synopsys,
until its new reorg, was wandering around in the dark. Cadence was still
trying to find its way as a company. Magma was under (and is still under)
the dark cloud of the Synopsys-Magma lawsuit; which stiffled anything that
Magma could do. Mentor was the only company moving ahead. They bought
Summit, and pretty much drove the direction for DFM this year. But one
company with its act together (Mentor) isn't enough to woo Wall Street.
Heck, even my own Gartner got out of the EDA business this year.
On the DFM side, in my opinion, ClearShape introduced the first true DFM
tool. It's the first tool that's targeted to the design engineer rather
than yet another backend guy. That's what DFM stand for. All the other DFM
tools I know of are all post-GDSII tools. That's why in my landscape I have
a DFY and a RET category, but no DFM category -- there have been no true DFM
tools to populate a DFM category until ClearShape came along. It'll be
interesting to see if ClearShape can actually sell the tool; nobody knows if
design engineers are willing to take on the responsibility for yield. Most
just want to throw the yield problem over the wall to the backend fab guys.
For ESL, it's just now starting to dawn on the EDA vendors that most of the
ESL revenue is not going to come from the silicon side. A next generation
tool overwhelms an older generation tool. For the revenue stream, IC CAD
(like Calma) dominated, then PCB (like Daisy) dominated, then RTL tools
(like Design Compiler) dominated, and it'll soon be that ESL will be
dominating. And for this revenue it won't just be the silicon jockeys
drooling over the ESL tools; it will be the FPGA and embedded processor
designers who will be the major revenue stream, as they'll be purchasing
large quantities of the software-oriented (the silicon has already been
designed), lower-priced ESL tools. The EDA vendors are just now waking up
to this fact.
One big business upside this year is it looks like Mike Fister's new Cadence
has finally decided they aren't a microprocessor vendor after all. They'll
have booth space at DAC in 2007 (although it looks like they still will be
doing what you, John, like to call "kidnap" customer meetings off site.)
Overall, I'm optimistic about 2007. It's got to be better than 2006 was.
- Gary Smith
Gary Smith EDA Santa Clara, CA
Index
|
|
