IEEE Solid-States Circuits Magazine - Fall 2020 - 6

A SSOC IATE EDITOR'S VIE W

Marcel Pelgrom

Mastering Complexity

W

When world chess champion Mag-
nus Carlsen was asked what he had
learned during his training sessions
with former world champion Garry
Kasparov, his answer was simple: how
to rapidly analyze a complex chess
position. And, indeed, Carlsen is one
of the fastest to find the crucial move
in an unknown position. An important
aspect of chess analysis is to recog-
nize structures: the specific arrange-
ments of pawns and pieces that
create a solid defense or allow specific
attacking tactics. Experienced players
cherish such structures as they form
the red line in their game.
Similar mechanisms can be rec-
ognized in designing and analyzing
complex technical systems, in both
the hardware and software domains.
In the era where a telephone ex--
change required a as much space as
a three-story building, the only way
to keep control over the design was
to strictly separate functions and
structures. The same is valid for
a " simple " system-on-chip IC with
various input/output channels, a
few processing units, and a diverse
collection of memories. Clear func-
tion definitions and interfaces allow
irrelevant sidelines to be offloaded
when overviewing the system.

Managing Complexity Starts With
Selecting People
Brilliant engineers design simple sys-
tems. Electronics and software mav-
ericks, on the other hand, are a real
danger, as their smart and elegant
Digital Object Identifier 10.1109/MSSC.2020.3021863
Date of current version: 18 November 2020

6	

FA L L 2 0 2 0	

ideas mess up the simplicity of struc-
tions, and even repairs are drowned
tures and turn mastering complexity
in an uncontrollable workload.
into a nightmare. Equally dangerous
The objective comes sometimes
are people with hammers for whom
with the specifications, but mostly
every issue is a nail. These engi-
it is more than that. Finding the
neers, once they've succeeded with
crucial objective can be trivial. For
a trick, apply the same solution to
Miele household appliances, the
every problem.
objective is reliability; for Ferrari
Engineers must be able to adapt
cars, it is appearance; for Apple, it
their way of thinking to the systems
is ease of use. When Apple wanted
they design. That starts with acquir-
to introduce the smartphone, Steve
ing a variety of tools and skills,
Jobs had every proposal brought
which allows for switching easily
into his office. He knew what his com-
from analyzing a control
pany objective was and
loop in the time domain to
how it should translate
Engineers must
its frequency domain or
into a smartphone.
be able to adapt
root-locus view. The real
Every proposed device
their way of
masters of complexity
that did not satisfy this
thinking to the
are able to simplify a sys-
objective was smashed
systems they
tem to its essence with-
against the wall.
design.
out getting distracted.
During the implemen-
They can change their
tation phase, the objec-
perspective and analyze a problem
tive leads and is recognizable in
from various sides. Broad techni-
the backbone of the system. In chip
cal experience allows them to use
design, that often starts with choos-
ideas and methods from other dis-
ing the appropriate technology. It is
ciplines. It is here that they outper-
tempting to continue working on the
form the young engineer. They see
next project using a known technol-
the red line running through the
ogy, disregarding the fact that the
project and, from experience, dis-
project objective is different.
tinguish essential operations from
Engineers have an unhealthy ten-
those that are nice to have.
dency toward economizing their
I mple m e nt i ng a s y ste m a n d
designs, as if the Excel generals in the
choosing its technical structures are
company would be able to spot the
both a consequence of rigorously
waste in hardware or code. Particu-
analyzing the required objective for
larly in the early phase of a system
the system. Forty years of working
implementation, it is crucial to avoid
have taught me that there is always
optimization. Unless there are tech-
only one objective that really mat-
nically compelling reasons, every
ters-the rest is negotiable. A system
structure or core has its own set of
built to serve multiple objectives is,
supporting functions.
by definition, too complex to mas-
Complexity requires an appro-
ter. Extensions, upgrades, modifica-
priate level of administration. In the

IEEE SOLID-STATE CIRCUITS MAGAZINE
IEEE Solid-States Circuits Magazine - Fall 2020

Table of Contents for the Digital Edition of IEEE Solid-States Circuits Magazine - Fall 2020
