( ESNUG 442 Item 10 ) ------------------------------------------- [03/24/05]
Subject: ( ESNUG 441 #2 ) Mentor Comments on the "DFT Compiler MAX" News
> For the Synopsys test rivals (Mentor, Cadence, LogicVision, etc.), how
> do you see this DFT Compiler MAX announcement?
>
> I'm really stumped here and would like your thoughts on this.
>
> - John Cooley
> ESNUG/DeepChip.com Holliston, MA
From: Greg Aldrich <greg_aldrich=user domain=mentor spot calm>
Hi John,
Since you asked how we see this announcement, here are a few comments.
It sounds like the major "new" thing Synopsys is announcing is another
embedded compression scheme that tries to address the problems of their
previous embedded compression solutions. I'd like to speak to a few of
the specific statements they made as part of this announcement.
First, when Synopsys claims that:
"Existing compression techniques ... require significant expertise
and are difficult to integrate into the implementation flow"
They must be referring to either their own compression techniques
(DBIST and XDBIST) or other competitors. Mentor's TestKompress product
generates standard RTL (which can even be done before a gate-level
netlist is available), so it easily integrates into any implementation
flow. In fact, some of our customers have stated that it fits into the
Synopsys synthesis flow better than the Synopsys solution.
Also, TestKompress is as easy to implement as standard scan and ATPG
since the flow is almost identical. There is no change required to the
core design and, from the tester and manufacturing side, the tests and
the application of the tests look just like standard scan patterns (just
much smaller pattern sets).
Synopsys also states that what differentiates their product is that it
does not require changes to the netlist for X-bounding and that it enables
scan-based diagnostics. TestKompress was architected from the beginning
with these requirements in mind. Since its introduction almost 4 years
ago, TestKompress embedded compression has not required any change to
the core design even when implementing very high compression ratios of
100X or more. It achieves the same high coverage as standard ATPG
without the addition of test points, and it does not require X-bounding
since we have patented technology to deal with unknown states. We also
understood that direct diagnostics was critical for any compression
solution and TestKompress supports direct scan-based diagnostics of both
scan chain and logic failures today.
It's nice to see that Synopsys has realized what Mentor stated almost
4 years ago when we introduced TestKompress embedded compression (i.e.
that test compression is becoming mandatory.) The main difference is
that Mentor Graphics' strategy for addressing the rapidly growing test
compression market appears to be opposite that of Synopsys. We
introduced the concept of embedded test compression in 2001, only after
years of development and technical verification including silicon
validation with our partners. We believe that well-planned, stable
products are necessary for users who will be trying new technology for
the first time. This has resulted in rapid customer adoption with well
over 200 tape-outs, and an estimated 500 million production devices to
date. Along the way we have also continued to make significant
improvements in a number of areas, including additional compression,
performance, modularity, and improved failure diagnosis. However, we
made it a point to keep a stable architecture and methodology in place
over the past four years.
Synopsys, on the other hand, in response to the success of TestKompress,
has fielded several different test compression solutions in the past
3 years. We believe that this is contrary to what users want, which
is an EDA vendor that releases and stands by proven, reliable products.
Synopsys has chosen to pre-announce their latest product before it is
ready for release in September, putting their customers in need of a
compression solution in an awkward situation. For upcoming designs, do
they use a currently released compression product that will be replaced
by another in a few months or do they wait and take a chance on a yet
unproven product?
- Greg Aldrich
Mentor Graphics Wilsonville, OR
Index
Next->Item
|
|
