What is Quantum Computing? – EEs Talk Tech Electrical Engineering Podcast #15

What is Quantum Computing? – EEs Talk Tech Electrical Engineering Podcast #15


hey everybody welcome to another episode
of EEs Talk Tech my name is Mike Hoffman and with me as always I’m Daniel Bogdanoff and today we have Lee Barford and Lee is a quantum computing expert
here at keysight so today you know Mike and I really know nothing about quantum
computing so we’re very very excited to have Lee on here we’re going to talk about
what is quantum computing Lee is in beautiful Reno but he will also be
floating somewhere over there so we have him on the phone today so Lee thank you
for being here can you tell us a little bit about yourself and sure first of all
I’d like to say I I consider myself very much a quantum computing (and quantum computer) beginner I’ve
been learning about it and and and helping guide some parts of keysight
into this into this into this business into keysight helping the real experts
make some advances in in quantum computing and quantum computers. my background is actually in
computer science despite the name of your of your podcast I have a bachelor’s
in computer science from Temple University in Philadelphia and a
master’s in PhD in computer science from Cornell however during the next during
the 30 years that I’ve been with keysight in its predecessor tis to
predecessor companies almost all that time I’ve been involved in coming up
with novel kinds of software that support other kinds of engineers most of
that time certainly all the keysight time has been in in in in in jin
software modeling techniques simulation techniques for electrical engineers but during for
example my part of my HP Labs time I was involved in new kinds of software to
help mechanical engineers and manufacturing engineers
yeah I got involved in quantum computing and this sort of goes into so
what’s some of the read what what some of the importance of it of quantum
computing is likely could be going forward
to back up for the what what’s the the motivation for it as you probably as you
may have noticed clock rates in CPUs it all kinds of digital processors stopped
going up in about 2006 and and that was because of heating limits that were
brought about that that it was no longer at that point it was no longer possible
to increase the clock rates and maintain any sort of reasonable thermal power
dissipation so the processor manufactures and I’m getting back to
quantum computing from this the quanta the processor manufacturers realized
that they had to have more parallelism in fact since about 2006 at 2007 up
until about a year or two ago most of my work for keysight involved power i’m
taking advantage of that parallelism and and and and teaching and enabling
keysight engineers to improve our product performance by taking advantage
of the parallelism rather than the clock speed available in current in in in
microprocessors for the last decade microprocessors GPU yeah yeah i interned
for a chipset manufacturer in high school right of that 2005-2006 range and i did
a lot of benchmark testing in the lab and it was right when multi-core
processors were starting to get popular and all we did was benchmarking and it
was you could tell certain software’s were taking advantage of the parallel
process and some were basically just using one core because it’s kind of a
symbiosis you have the hardware but you have to program for that hardware it’s
not done automatically which is anneka and then there additional additional fee
such as most processors now have vector instructions that can do multiple
integer floating-point operations per cycle if you have a per instruction if
you have multiple operations ganged up in a single register there’s the use of
of graphics processors as basically vector and matrix machines that can
accelerate all sorts of signal processing and matrix map in addition to
the graphics or which they were originally intended now thank us thank
you thank you video games but also also thank you physicists that went to the
graphics chip companies and said look look we can do this neat stuff with your
with with your chip please open up the interfaces so that we can do that and
then and the graphics chip companies saw a market there and complied and and and
and provide a Bitcoin mining I think they use graphic cards but that kind of
Sally this kind of block chain style calculations because there’s so much
faster yeah yeah so yeah this is with it and so the physicists of heavy use in
physics heavy use in heavy use in well heavy use for all kinds of simulation
including including electronics including electronic simulation heavy
use in in in biotech also well-suited well suited for all those things as I
might mention late a mention later we come back to it they’re also very good
on the sorts of computation you need to simulate quantum computers which has
which which which is an important thing because they’re quite expensive to build
and yet there’s a a lot of the experimentation that’s happening is with
simulating them including simulating them with with the assistance of GPUs so
you can simulate bigger bigger ones and why that’s interesting I’ll get back to
anyway we’re going down the path of having more and more parallel computers
with every tick-tock of the of the cycles of improved
including feature size shrink in semiconductors however when we get down
to right now depending on who you believe there are parts out there being
made at ten maybe seven nanometers none of the manufacturers are very specific
about what that actually means but let’s say that actually means that the actual
there’s an actual feature size in the digital transistor that say seven
nanometers today that means in not too many more years will be heading below
five nanometers and at that point there aren’t very many unit cells of silicon
left a unit cell silicon is about half a nanometer so at five nanometers were
down to about five unit cells I don’t know what that is 35 40 atoms something
like that of silicon and then you’ve got the dopant atoms in there at that point
quantum you’ve got some fuel enough atoms that you’re close to or at the
point where quantum mechanical effects are going to disturb the electronics so
one way when you take quantum mechanical effects is that like the uncertainty
it’s like the uncertainty principle it’s like it’s like the uncertainty principle
and like the fact that when you measure when you measure measure when you
measure the when you measure the physical state that the act of
measurement in self will noticeably perturb the state including putting
taking it from an indefinite superimposed quantum state a la
Schrodinger’s cat to a definite state you it means having having superposition
of states well I mentioned Schrodinger’s cat already all right that’s that’s the
Super’s ition of the states the cat the cat alive the cat dead in in digital
electronics that would be a zero state in a one state you have some prop
ability of being in one or the other and the act of actually measuring the bit so
that you can do some other computation with it as you can read it out will
force it definitely into the zero one state even though it was formally and
one or the other and also just just the noise that the the noise and
disturbances that come from the physical roughness at that kind of like that kind
of level and the fact that manufacturing errors that are relatively small will
mean you have a different number of atoms in each replication of that
transistor say but won’t be but but but you don’t have that very many
transistors to start with right so you’re going to have higher and higher
manufacturing variants just because normal tolerance of a couple atoms one
way the other is a big is a big difference in the number of atoms in
that your hands ister yeah your tolerances get quantized exactly exactly
real and and and even though the physicists wouldn’t call that a quantum
mechanical effect but in in the theoretical sense as you notice it still
has to do it still change it’s still it’s still coming it’s coming from the
fact that you’re dealing with quanta you’re dealing with individual atoms and
the count just even the count of them starts to matter so one way to talk
about quantum computing is that it’s just one way of moving the computer
industry past that barrier there are there others out there they’re there
they’re there other approaches and and I’m even less expert in them and so I
won’t go there people for them for how to get past that barrier one way to look
at the quantum computing and the reason for their investments in quantum
computing is it is it is a it is one of the potential ways past that barrier
that basically is that the barrier laws law essentially reaching the limits of
physics moore’s law reaches the limits of physics and so there’s kind of two
things you can do on the one on the one and you can you can try you can try
workarounds and and error correction methods and computational approaches
that can deal with high error rates in the fundamental electronic operations
and do sorts of error correction error correction mechanisms that deal with the
fact that the the that underlying digital electronics are going to be very
noisy but another another another approach is to learn to engineer with
quantum mechanical effects and take advantage of them hmm so expect so
you’re at a party and someone comes up to you and says Lee I’ve never heard of
quantum computing before what is it what is it down grab a drink
how would you go it’s the use of it’s it’s it’s it’s it’s taking advantage of
it’s taking it it’s learn it’s taking advantage of quantum mechanical effects
engineering with them to build a new kind of computers that for certain
problems promise but are not not proven to do better than what we have now even
what we could build at this mot when we when we get to the end of Moore’s Law okay yeah building do you think we will
I don’t know do we will he see the end of Moore’s law it depends on how you
define Moore’s law if you define Moore’s law to mean that digital CMOS shrinks
every shrink severy 18 months or now it’s somewhat longer than that which in
itself is a sign of the at the end of the potential end
yes Moore’s law hat Moore’s law has to end because you can only make a
transistor features so small if on the other hand you use Moore’s law to
colloquially mean electronics continues to advance so that there’s more
a computational capability faster wider bandwidth analog capability the ability
to build larger parts at less cost I don’t see that that’s coming to an end
because there are lots of different there are lots of different experimental
approaches to move forward after the be and by expert I mean being researched at
like universities around the world for moving beyond for moving beyond the end
of digital seem the end of digital CMOS as we have known it and continuing to to
to allow electronics to make the electronics industry make progress
so all of those approaches would need to fail right for an order for for progress
in electronics to stop all those experimental approaches would need to
fail and they’re not all going to fail some of them will succeed right so I
guess the answer is if if you want Moore’s law to fail you can define it as
such that it will definitely fail and it’s your Moore’s law fanboys and it
will continue to to live on well I mean you can think back to the old days of
vacuum tubes right it’s like your could only get so small right and it’s like
boom now we have transistors yes and you could you could kind of say Moore’s law
is independent of the technology being used I think that’s sad yeah yeah if I
know there’s some Moore’s law haters I think again I want to draw this
distinction if you more strictly speaking more meant as law to apply to
digital like digital electron large-scale integration he would have
called it back in the early 60s right and so if you’re being if you’re being
formal right that’s that’s the real meaning of Moore’s Law if you’re being
informal the Moore’s law just meaning electronics gets faster cheaper better
then then you’re right okay awesome so you know we’re actually nearing it’s the
end of what we Exuma considered to be an episode okay you want to take a if we
had a few minutes left is there anything else you’d like to discuss
we talked about you know technologies and challenges leading up to quantum
computing can you give us a little bit of teaser for future episodes because
we’ll definitely have you on another yeah sure I think I think future future
episodes can cover things like what sorts of what sorts of technology goes
into a goes into a quantum computer what the current state of experimentation is
where the money where what is it what are some more of the motivations
including where the money is coming from and I think we probably also want to say
how is keysight already already involved in helping those experimentals you make
progress because we’ve kind of you’re saying before we started we’ve been you
know secretly ish involved in quantum computing up till now and you can like
you know redact your resume yes well I think this is a bit of a coming out for
us you know that hey we are here I’d be interested in to knowing what kind of
problems quantum computing is trying to solve and yeah will quantum computing as
we know it today more if its way into consumer electronics right place right
the CMOS based computing we have today or is it purely a specialist super high
end applications yeah and that’s a that’s a really that’s a that’s a that’s
a that’s a that’s a question I can probably talk about but absolutely not
give you a definitive not give you a definitive answer because one of the
things we’ll be talking about doing is that right now all pretty much most
quantum computing experimental approaches require very extreme
environments very high vacuums or very cold temperatures that that you can it’s
hard to see how those approaches are going to be in consumer products yeah
yeah in the last couple minutes we have left you talked about how we have to you
know quantum computing is about using those quantum effects
I’m going to use the term loosely yeah back to our advantage
can you give us a brief overview of what’s that sure not it’s not a
specifics of how people are doing it but what yeah so I think that this the
simplest one to describe is is that of is that of a superposition of states as
a so consider even used any of the terminology yet but in quantum computing
the fundamental storage unit is something called a quantum bit (qubit) and in
the quantum bit it the quantum bit can be in states that are super positions or
mixtures of 1 and 0 if you have a quantum register consisting of so that’s
that’s analogous to your visit Schrodinger cat right it’s alive or dead
with some probability likewise a quantum bit can be a 1 or 0 with some
probability that doesn’t seem useful to me that seems like ah but what happens
if you have a quantum register of n bits and you want to store a number of n bit
data data you can store them all you could store them all with if you if you
want to make them in differentiate it you can store them all with equal
probability and then then then there could be an algorithm for searching
searching all of those and there is such an algorithm can you give us a simple so
like you attend it register and you need the stored 10 bits you know well you
could still yeah but you can store a number of different 10 bit numbers in
the same register and compute on them assuming 10 factorial ones right yeah or
two does it yeah I’m still trying around or potentially assuming that you had
incredibly accurate or physical right in the assuming you had a sufficient
accurate quantum bits sufficiently no noise quantum bits you could store all
the all the two to the tenth possible ten bit not ten bit numbers in that
register usually it’s going to be more interesting how do you read them yeah
that’s that’s one of the trick you can only you the bad news is you can only
read one you can store them all but you can oh if you
have them if you have them all stored with equal probability one way to one
model of quantum computation is just like when you when you open the box and
Schrodinger’s cat is either alive or dead once you look inside once you read
that register you’re only going to get one of the answers and even if you
reread the register so yeah because it would be out there because just like
Schrodinger’s cat it will collapse to being live or dead right shirting yours
cat you close the box and open it again if your first time you looked at it the
cat was dead it’s still dead when you look at it the second time so it’s the
same thing with that register when you read it the second time it’s going to be
whatever read whatever value it you read the first time so how do you get what
you want out of it the answer is like theoretically you’re
trying to cure you’re going to do you’re going to do as much of the computations
as you as you can in with it with the with the computer operating in the
undisturbed unread mode and then only when you when you when the algorithm
requires it or it’s finished do you read it and get an answer so you can
basically like reduce out your formulas and your computations and how it depends
on what that is specific yeah and then you read it out and you like okay that’s
it here’s answer you that’s one that yeah that’s that’s that that’s that’s
one that’s that’s one way it works now typically the answer is it typically the
algorithms are random algorithms so you were you were in you were expecting to
get that have the that the the prop you have the following
properties either the algorithm is going to give you the right answer with such
high probability that you don’t need to check it or the problem was one that
it’s easy to check that you got the right answer so one of the problems that
people are very interested in is factoring very large numbers because
that can be important in cryptography here breaking RSA
is is the two-word explanation okay so I killed me to cut us up oh yeah okay and
you ever said here and we will definitely be picking this up maybe we
start back up talking about security next time yep you know we may even
instead of every other week do some depending on how long we’re yeah here
chatting do a weekly thing instead I really think so Thank You Lee Barford
that was fantastic my brain is missing make sure you subscribe to the EE Tok
tech podcast or the keysight oscilloscopes youtube channel you guys
know the drill if you’re on iTunes or Google Play or stitcher or any of those
the podcast is there please give us a rating give us a thumbs up if you like
this video and subscribe to the keysight oscilloscopes youtube channel Daniel
Bogdanoff Mike Hoffman Lee Barford we will see you next time thanks Cheers

7 thoughts on “What is Quantum Computing? – EEs Talk Tech Electrical Engineering Podcast #15

  1. Like to know more about basics of processing and correlation with the digital computers. What kinds of computations or or processing it is really useful. Real world examples of how they can work with classical computers to overcome the limitations. (e.g. we know that FPGA or GPU cannot completely replace CPU but how powerful they are together)

  2. Very good video except the audio quality of Lee. But it's not that bad.
    I would love to hear/see something about power electronics.
    Something like measuring the reverse recovery losses of SiC-diodes for high currents (kA) to choose a proper heat sink/cooling system or general the differences between SiC and Si.
    The Control and drives for big motors (or alternators in power plants) and what is needed to do that.
    But probably this isn't Keysight's field of researches. Anything about power electronics would be nice 🙂
    Keep up the good work!

  3. Daniel, I sure wish I could win an Oscilloscope during "Scope Month" but because I live in Florida I guess that's not going to happen and with your prices, a KEYSIGHT SCOPE is a pipe dream. I was double disappointed last year on the last day of "Scope month" finding out because of living in my state I was disqualified and the second disappointment was KEYSIGHT telling me they would get back with me but never did .

  4. I’m extremely satisfied after watching this video , I’ve been interested in quantum mechanics and it’s properties as well as the quantum computing. I’m thinking about going into a computer engineering job…

  5. Incredibly interesting, make sure to create a playlist on the channel for the podcast in case you already don't have one

Leave a Reply

Your email address will not be published. Required fields are marked *