|
|
Main
topic
Corporate
profile
Warning on
misnomers
What Design Industry?
Targeted Market
Segments
Market analysis
You want a System-on-Chip?
A Practice of
active cooperation with all innovative actors
A Strategy of
Value Enhancing Differentiators
Growth through
innovation in Value Proposition
About service,
comparison is no reason
Growth from
a nurtured corporate culture
Major growth
programs
Our Corporate
Governance
You want a System-on-Chip?
What makes such a fledging SIP sector difficult to analyze is the
ambiguity of most competitors' positioning: most Custom Design Centers
and some Fabless IC suppliers also pretend to act as SIP providers,
while some key players are bought and discarded unexpectedly (Pivotal,
Packet Engines). Anything can be loosely called IP. The decisive factor,
for cautious consideration, is the difference in investment cost with
duration, between some available piece of SIP and a true ViC macrocell,
properly productized for fast delivery.
Typically, IDM’s may claim to have a large portfolio of SIP,
but without productization into ViC which could not be justifiably
performed as long as it is limited to internal usage, with the inherent
risk in any design that was only proven faultless within a single environment,
as opposed to a true ViC thoroughly specified and tested with respect
to any potential context of use or configuration of SoC!
Also, the present deverticalization of the semiconductor industry is
leading to an individualization of actors into new categories:
Silicon Foundries have stirred the emergence of Fabless (ASSP) Application
Specific Standard Product Suppliers, already distinct from Fabless
(ASIC) Application Specific IC suppliers. Now SoC or System Integrators
are emerging as distinct from IC Design Center, while System Designers
as distinct from internal OEM designers… Meanwhile successful
IDM's concentrate their efforts on an efficient combination of IC fabrication
and SoC integration, and they tend to externalize both SIP development
and ViC procurement.
To make matters subtler, four inconsistent definitions of a SoC are
encountered:
- an IC with more than 500,000 active elements (transistors…)
- a complex microcontroller, i.e. with at least one microprocessor
core, demanding for software
- an IC single in its application so as to make useless any (PCB)
Printed Circuit Board
- an IC with at least one externally-sourced ViC…
A "Mixed signal SoC Enabler" like DOLPHIN may accept either
of the last two definitions, as long as they concentrate on CMOS processes
(bulk or SOI). It is estimated now that two thirds of SoC's shall be
with mixed signals within four years.
The DOLPHIN definition of a SoC hence is pragmatic, namely whatever
complex IC designed hierarchically from Virtual Components, giving
sense to its twofold market targets as SIP Editor:
- strategically enabling the low-end SoC market with some essential
ViC, around legacy 8/16 bit microprocessors with their enabling
technologies
in Development Platforms (low-power embedded memories, SUCCESS™ cosimulation,
MEMS generators…)
- opportunistically providing the high-end SoC market with the
same ViC, but as peripheral controllers, even if they look like
commodity there
next to high-speed embedded memories, AD/DA converters,
with VHDL-AMS simulation… even though CAD frameworks still
are more fashionable there.
Next paragraph:
A Practice of active
cooperation with all innovative actors
