Lithography based semiconductor chip making
in less than 15 nm node is less certain
In electronics industry there are few people who knows
trends in every domain of the industry and the wisdom they
share is valuable, one such person is Dr. Walden C. Rhines,
Chairman and CEO of Mentor Graphics. He is respected for
clear macro-level view of the complete industry and micro-level
view of EDA software industry. In an exclusive meeting with
him, I asked few questions on the broad trends in the industry.
Here below is the Q& A interaction with him.
Q. These days in EDA industry, there is a trend of
offering more of an integration services along with design
software package, how you see this trend?
Dr. Walden C. Rhines: Very true -- that companies like
Mentor Graphics provide Solutions to problems, the difference
probably between Mentor and others is -- we help our customers
put together best in class solutions, we deal with open
standards, we make it easy for innovating startup companies
to integrate their tools in our flow, we work with them
to ensure that our customer get the best flow and, we ensure
that what we do is a best in class so that we can put together
complete design solution.
Q. So if there is startup in India with its own product,
you are ready to accept its product if it fits into your
kind of a product setup, and sell it along your product
or something like that?
Dr. Walden C. Rhines: We will always help them to interface
into our tools and our flows so that they can be part of
solutions for the customer.
Q. Does the technology advancement in EDA domain
create more jobs or cut jobs?
Dr. Walden C. Rhines: It always grows jobs. It's true that
in recent times during the recession, the number of jobs
in EDA actually flattened and even declined little, but
whenever the technology innovation creates the need for
new tools, new methodologies and new technology -- and every
time the new node is created -- right now we are ramping
up 28nm and lot of our R&D focuses on 20nm and that
will create new products and the revenue for us -- and off
course new jobs (create) for the people who do the development
and many of them are right here in India.
Q. When the semicondcutor industry is moving to 28nm
and less, it is becoming extremely challenging for them
to develop chips at that nodes, is EDA industry also face
same level of challenge in developing chip design software
for shrinking geometries?
Dr. Walden C. Rhines: Everyones job is tough, we have different
challenges when we go to new technology nodes, as we now
trasiting to 28nm and as we shortly to 20 nm -- because
millions of dollars of R&D to do the basic work.
Q.So you lead ahead of chip makers in delivering
them the software?
Dr. Walden C. Rhines: Typically we work in partnership as
they develop their process we do the tools in parallel.
Per example we have a joint venture with IBM on the developement
of their 28nm process , and we work closely together developing
the software that will help them to resolution enhancement
they need for their process and develop tools for our customers
who will be able to use in order to take advantages of that
technology.
Q. These days if you look at the chip design, its
more of a software rather than a hardware, the level of
software is so high, how does this impact the learning habits
of engineers?
Dr. Walden C. Rhines: Traditional EDA design is been focused
on hardware. In 1996, Mentor begin a major step to broaden
the solution to provide solutions to both the hardware and
software problems -- just as decade earlier we have expanded
the focus from digital to analog to mixed signal. With the
last 15 years, we expanded our solutions for embedded software
development, embedded software verification, the use of
the open source community to help to support our customers,
and also working in partner ship with companies like Freescale
Semiconductor, Netlogic and others.
Q. What should the colleges teach for changing environment?
Dr. Walden C. Rhines: Just because the development of EDA
tools is a software discipline it does't mean you don't
have to know about device physics. So, I think in general-eduation,
it is important to have an understanding of both the physical
reality as well as software development techniques.
Q. So the core of subject should remain same!
Dr. Walden C. Rhines: I think core engineering training
we look for is much the same, except for the specific skill
that are required now.
Q. Do you foresee change in the top order of global
semiconductor vendor ranking?
Dr. Walden C. Rhines: I actually regularly give a presentation
on how the top 10 semiconductor companies change, and the
principle driving force is being the leader in whatever
the new technology is evolving in a decade. The last decade
really seen a lot of fables seminconductor companies. Every
dacade is something new.
On the emerging model of fablite!
Dr. Walden C. Rhines: Fablite is evolving at some time,
so now half of the top ten semiconductors are in fact even
fables or fablite. One fables is qualcomm and 4 are fablite,
but you could argue there are some strange things going
on, one of the people point to Apple Computers which was
a company that simply purchased all its components and now
getting into the area of Integrated Circuit design and processor
design. So, I think the change that is occured, because
companies need to leverage in different ways. Apple which
produces enormously high volume of consumer products, that
they find designing own processor is an important way to
capture more of the value and differentiator they require.
Q. In today's world if a startup has to scale up
to billion dollar company, what are ingredients for success?
Dr. Walden C. Rhines: Most startup companies atleast dream
of becoming billion $ companies. What's differnce about
what you do early on to get there, I would say, The most
important thing is in the foundation of the company you
built in cultural values that can transcend different sizes.
You will go through up and downs, but if per example, if
your company has basic priniciples of integrity, the way
they treat the customer and how they view their business
that allows them to scale and develop new technologies and
capabilities required. If you have a culture that does't
scale then off course, you are limited and you have probably
stop short of a billion and may be a 100 million.
Q. In India, do you foresee the market for R&
D design services offered by companies such as Wipro will
continue for long time? Do you see this business model as
a good business model?
Dr. Walden C. Rhines: I think design services by independent
companies are a very good thing. It allows a degree of specialization
and it allows people to become very effecient in the core
sense of the design. For EDA companies, I don't think it
need to be core business. In Mentor case, we find it is
not desirable to compete with our customers, we really don't
do much turn key design. The services we provide is for
teaching people like how to better use our tools and we
leave it to the Wipros of the world to do the actual design
execution for the companies who would outsource that capability.
Q. For a region like India, do you see a value in
electronic hardware manufacturing, so that the phones and
computers can be made locally instead of importing since
significant market growth is happening now?
Dr. Walden C. Rhines: I think certainly that the option
is available to India. You have resources, you have people
that are educated and innovative, but I think the great
thing about the technology today is, you dont have to. In
a business that scales worldwide, we can have certain regions
of world or certain companies that intensly manufacture,
foxconn, per example, an 80 billion$ company that largely
just build things, where as you have design companies here
in India that do nothing but design, and because we are
virtual world, we can specaialize, so, you don't have to
do everything in every region and you can choose the one
that you have greatest advantage.
Q. Does how long this present lithography based chip
making technology using silicon will continue?
Dr. Walden C. Rhines: Well, we clearly have a roadmap to
20 nm going out to the point that how the basic production
will work, the technology to 15 nm is quite clear although
it involves some cost challenges around, the next turn is
a little less certain, because the lithography becomes a
problem. Whether we need a change beyond that is not clear,
traditioinally we go find a way to go beyond but at some
point we will have to go to something like extreme UV, EUV,
or some other approach, if we want to continue to shrink
feature sizes. Now I might note we don't necessarily have
to go on this linear path to feature size shrinking, we
can still get cost reduction by stacking dies or growing
design in 3rd dimension, so we have many variables in hands
and that one we choose will be the one which is cost-effective
to move ahead, and the linear shrink; Moore's law kind of
a thing is not necessarily the most effecient way to get
it.
Q. You have been in the industry for 25/30 years,
has the present technology reached the imagination you had,
or has it exceeded?
Dr. Walden C. Rhines: I think every one in the industry
is amazed at what's been accomplished. All of us had visions
for greatness of technology and many great things happened
and we still surprise ourselves by the breadth of applications
that have emerged. The things we could have not thought
off, way to use the technologies, but also, how far did
it pushed, How many times we said this technology will come
to end, and yet we find some innovative ways to go further.
Its great thing about the semiconductor industry.
Q. Finally about your Indian office? Is it expanding
more than what you have anticipated?
Dr. Walden C. Rhines: I tend to believe that expansion
in India is inevitable, Indian economy is quite strong but
more to the point we go where we find balance we find excellent
engineering talent in India and that's what bring us here
in India. Even though on salary basis, salary diffrences
are getting smaller and smaller, I recommend to the companies
don't go to new sites because of low cost because eventually
there won't be low cost. The reason is you can find people
you can't find in other places and India is a rich resource
in that sense and that's what we continue to grow and expect
to grow in near future.
