►
From YouTube: Caravel SoC Learnings
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Final
speaker
of
the
day
is
Muhammad
qasim,
who
is
the
CTO
of
e-fabulous?
So
thanks
for
joining
us
and
apologize
for
the
delay
and
time
here,
but
looking
forward
to
your
talk,
thank
you.
A
B
And
let
me
I'm
gonna
I
think
I
can
still
share
the
screen.
A
B
Okay,
so
so
my
name
is
Muhammad
qasim
I'm,
gonna
I'm,
just
co-founder
of
the
CTO
D
fabulous
I'm
I'm,
going
to
first
on
elaborate
on
the
name
of
the
title
of
presentation,
because
some
of
you
have
seen
some
of
the
links
or
the
posts
are
asking
what
is
Caravel,
but
this
is
doing
the
only
engineering
related
to
the
infrastructure
and
the
simplification
working
with
Google
on
the
programs
for
the
open
source
chip
design.
So
that's
my
back.
B
My
background
is
Hardcore
electronics
and
Chip
design
for
devices
that
already
that
now
no
longer
exist,
except
on
eBay
as
used
devices
because
I
have
lifted.
You
know
that
world
probably
12
years
ago
to
start
the
company,
so
I
used
to
work
at
Texas.
Instruments
I
was
an
analog
mix
signal
designer
designing
all
the
analog
on
chips
like
home
app,
you
know
pick
anything
outside
the
digital
infrastructure
and
that
would
be
I
was
responsible
for
it.
B
Initially,
I
was
responsible
for
one
of
them
and
then
I
had
a
great
team
of
Engineers
that
we
worked
on
the
risk
as
we
go
forward
before
I
leave,
TI
I
also-
and
this
picture
I've
probably
seen
it
before
I
I
do
buy
things
to
understand
where
they
are
and
how
they're
designed,
and,
of
course
you
know-
probably
many
of
you
will
be
doing
that
so
on
a
little
bit
about
you
know
the
context
of
the,
why
you
know
we're
doing
what
we're
doing
so,
especially,
for
example,
within
the
chips
Alliance
at
the
end
of
the
day.
B
What
we
want
to
do.
We
need
to
get
to
a
position
where
we
have
an
open
source
chip
design
that
is
available
in
a
way
that
is
commercially
usable,
similar
to
what
Linux
has
arrived
to,
and
in
order
to
do
that
there
is
a
there's
a
lot,
a
lot
of
steps.
But
when
you
do
that,
think
about
in
the
hardware
perspective
you're
going
to
you
you're
targeting
devices
like
what
I
have
here.
So
imagine
you
know
we.
There
are
many
many
devices.
B
Typically,
these
things
happen
with
larger
companies
or
larger
companies
that
have
Capital
and
the
to
we
want
to
convert
that
into
some
sort
of
a
process
that
allows
a
low
volume
product
lab.
B
Now,
if
you
look
at
this
picture
of
translation,
is
that
on
the
left
is
the
smartphone
unlimited
or
maybe
hundreds
of
millions
of
units
or
for
one
chip
or
you
know,
or
our
derivatives
and
then
on
the
right
hand,
side
the
orange
a
lot
of
applications
that
don't
require
or
don't
exist
for
larger
volume,
because
of
just
the
market
is
Niche,
but
they
need
customization.
They
need
to
to
connect
I'm,
sorry
to
optimize
the
the
compute
and
power
to
these
applications
versus
just
having
a
standard.
B
Typically
historically,
this
hasn't
happened
because
and
the
it
was
very
expensive
to
think
about
customization
for
a
small
or
low
volume
product.
So
you
know
in
order
to
get
there,
we
need
more
custom
A6
for
low
volume
devices
and
who's
going
to
design
them
as
a
typo.
Here,
that's
an
S,
so
we
need
to
see
a
lot
more
designers
to
be
able
to
to
achieve
that
then
actually
go
in
the
future.
B
So
the
process
of
Chip
design
needs
to
be
simplified
in
a
way
that
you're,
you
know,
I,
think
being
a
hardcore
semiconductor
person.
I,
don't
think
we're
going
to
achieve
a
thousand
X
by
teaching
everybody
to
to
become
an
expert
in
ic
design,
and
things
have
never
worked
that
way
in
the
past.
In
anything,
you
simplify
the
process,
make
it
available,
so
you
can
including
driving
cars
or
the
app
app
stores.
What
they
did
is
that
they
simplified
the
process
of
making
a
new
software
product
and
put
it
in
the
marketplace.
B
So,
to
start
with
what
is
carabil,
the
name
was
coming
from.
First
of
all,
it
was
chosen
by
my
colleague
Tim
Edwards
from
igabus,
and
it
it
means
what
it
means,
and
it's
a
it's
a
you
know,
15th
century
ship
with
a
certain
size,
and
in
this
case
it's
called
carrier
chip
and
that
that
where
Caravel
came,
but
it
is
actually
a
chip
as
a
that
is
a
part
of
a
platform
that
makes
a
designer
start
from
the
15th
floor
to
reach
20,
rather
than
starting
to
build
the
chip
from
scratch.
B
So
if
you
look
here
so,
for
example,
in
in
general
designs
versus
highly
customized
for
a
specific
application
like
no
leakies
or
so
iOS
on
the
Chip
And
The
Packaging
are
just
are
just
a
just
necessary
evil.
You
need
to
protect
the
chip
from
ESD.
You
need
to
mechanically
protect
your
chip,
but
the
actual
design
could
be
that
a
part
that
you
would
be
the
part
in
the
middle
here.
B
So
the
way
it
was
designed
is
to
leave
that
10
millimeter
squares
in
the
middle,
and
then
the
designer
would
focus
on
what's
building
what's
in
the
middle,
and
then
we
have
an
automatic
approach
just
to
integrate
it,
and
then
it
goes
to
the
Fab,
and
then
you
get
bored
rather
than
parts
so
makes
it
easy
and
to
actually
test
it,
and
this
board
is
USB
pluggable
and
you
can
run
it
right
away.
The
Carabella
SOC
is
actually
a
riskified
microcontroller
with
a
few.
B
You
know
it's
not
a
super
microcontroller,
but
it
is,
has
a
a
little
bit
of
ram.
It
has
a
peripherals
and
it
has
fully
controlled
by
those
one
of
the
things
we
do
and
we've
done
in
the
past
is
that
the
the
risk
5
code
that
was
actually
here
inside
the
this
area
of
Carvel
was
it
was
a
decor
v32
and
we
have
two
flavors
between
the
fabulous
and
and
and
the
Google
program.
B
We
also
actually
we're
going
to
have
another
flavor,
but
it's
not
necessarily
open
source,
but
it's
going
to
be
based
on
arm.
So
there
are
multiple
flavors
that
can
have.
We
can
change
that
and
work
with
it's
a
node.
It's
actually
a
note
to
say:
I,
don't
have
a
slide
for
it,
but
it's
this
area
of
the
chip
was
built
using
amaranth
and
latex,
and
it
has
been.
B
They
basically
started
from
a
python
Hardware
description
and
we
adapted
the
latex
framework
to
Target.
You
know,
in
addition
to
fpgas
to
Target
Sky,
130
and
gf1
is
now.
B
This
chip
is
available
in
two
two
process:
Technologies
skyr
skywater,
130,
nanometer,
open
source,
BDK
and
Global
Towers
189
meter
MCU,
it's
an
open
pdk
as
well
the
open
source,
Speed
game,
the
that
flavor,
the
the
global
families
version,
is
slightly
different,
slightly
smaller
in
terms
of
the
side
area
here,
but
it's
still
writing
greater
than
nine
millimeter
squares
and
I
and
I
will
eventually
provide
pictures
that
are
actually
or
more
information
on
that.
But
the
shuttle
for
Bloomberg
families
has
been
out
on
many
designs.
B
It
is
so
if
you
look
at
zoom
out
and
look
at
the
whole
cycle,
this
is
just
their
beep,
but
a
big
part
of
the
project
was
actually
enable,
tooling
and
open
source
tools
and
Eda
to
allow
you
to
not
have
any
Show
Stoppers
in
order
to
start
designing
and
that,
actually,
you
can
start
doing
this
process
at
zero
cost
until
you
get
to
manufacturing
and
you
can
start
a
digital
design
or
analog
design
and
they're
different
because
they're
to
get
a
little
chip.
We
have
two
different
versions
of
it.
C
B
So
we
in
that
results
we
have
an
offering
that
is
for
startups
and
universities
and
other
systems
that
not
necessarily
want
to
open
source
their
10
millimeters.
They
use
an
open
source
chip,
so
we
have
a
product
called
chip
ignite
we're
offering
multiple
options
in
it.
It's
not
a
part
of
the
Google
program,
like
some
people
asked
for,
for
that's,
not
a
part
of
the
Google
program
yet
or,
and
and
it's
being
discussed
whether
it's
going
to
be
there
or
not.
B
But
but
these
are
caribou
has
become
a
platform
that
you
can
do
many
things,
including
just
you
know,
not
using
some
of
the
resources.
So
it
is
intended
to
enable
the
learning
Innovation
cycle,
which
is
you
design,
something
you
you
test
it
you
learn
it
and
then
repeat:
the
thing
is
in
the
chip
world.
This
is
expensive,
it's
actually
not
cheap.
So
in
order
to
make
it
cheap,
the
Google
you
know
is
a
part
of
is
the
driver
of
an
open,
mpw
program
funding.
C
B
Source
you
complete
for
manufacturing,
that's
actually
a
program
that
has
been
2021
22.
and
it's
ongoing
in
23,
and
thanks
to
all
the
team
and
Google
to
actually
work
with
us
in
the
community.
To
make
that
happen.
One
of
the
things
that
have
happened
also
is
that
we
started
in
2022
and
2020,
rather
with
the
skyline
30
process
now
over
time,
and
because
of
the
success
and
the
number
of
designs
that
were
generated
on
the
sky
130..
B
The
other
found
reason,
including
a
major
Foundry
like
Global
foundries,
saw
the
the
value
in
opening
another
process
with
allowing
the
community
to
design
on
it.
It's
a
very
simple
business
proposition.
The
more
designers
are
using
your
process.
The
more
likely
you
will
hit
a
home
run
is
the
statistical
problem.
Just
you
know,
you
know
you
just
take
them
the
more
more
designers,
more
users,
more
likely
to
have
a
whole
business,
so
that
was
demonstrated
on
the
sky,
water
and
then
at
the
in
the
third
one
in
2023.
B
Built
on
an
infrastructure
funded
by
the
government
for
Sky
130,
and
this
process
is
a
different
process
from
the
ones
on
the
left.
That's
fully
depleted
SOI,
which
has
unique
properties
for
leakage
and
analog
design.
All
of
that
no
NDA
and
you
can
go
to
this
link
to
actually
find
it
or
clone
it.
Recursively
and
you'll
get
everything.
So
that's
a
description,
more
detailed
description
to
of
the
Google
program.
B
You
know
the
users
would
get
five
boards
and
300
CSP
Parts.
Everything
needs
to
be
under
open
source
license
and
one
of
the
things
that
we're
trying
to
drive
with
we're
trying
to
raise
the
quality
of
the
designs.
How
the
quality
here
doesn't
mean
a
design
that
has
a
better
function.
It's
a
design
that
is
working
to
a
certain
spec
is
defined
by
the
designer
and
fully
reproducible
by
other
people.
The
reason
we
do
that
is
that
we
want
instead
of
having
a
pile
of
gds's
or
layouts
or
code.
B
We
would
want
a
full
structure
that
other
people
can
learn
from
and
and
that's
how
people
learn.
That's
how
you
know
when
you
you
want
something
to
download
it.
The
first
thing
you
want
to
do
is
run
it
if
it
works,
you
start
modifying
and
it's
not
limited
to
chip
designers.
Most
of
the
designers
that
we've
seen
are
coming
from
adjacent
backgrounds.
B
A
key
metric
for
this
and
I
stole
that
text,
but
I
added
to
it
Tim
Ansel.
So
it's
actually
had
the
number
of
users
and
contributors
to
everything
that
is
in
the
ecosystem
of
Chip
design.
So
and
I,
say
users
see
a
user
could
be
somebody
who
downloads
something.
But
it's
also
if
the
project
is
actually
people
invest
the
time
to
contribute
into
it.
B
That's
actually
a
commitment
that
actually
means
that
it
has
been
so
valuable
to
someone
that
they
believe
that
they
can
use
it
not
only
use
it
but
also,
and
their
own
contribution
of
you
know
either
find
the
dishes
or
addition
fixing
issues
or
additional
development.
So
these
are
metrics
that
that
are
very
important
and
because
we're
doing
this
project
that
is
completely
new.
It
is
important
to
keep
that
in
on
in
sight
as
we
move
forward.
B
So
now
this
was
a
much
lower
number
last
year,
and
so
we
have,
as
of
today,
we
have
4
700
member
Plus
on
the
slack
Community,
which
you
can
go
sign
up
to
it
and
you
you
can
go
just
when
you
click
on
this
you're
going
to
go
to
the
you're
going
to
get
the
the
the
website
that
you
can
put
your
email
in
there
and
you
will
get
a
a
an
invite
for
the
the
Select
Community
and
it
has
many
Channel
channels
and
it's
everything
is
open
there
and
that's
the
beauty
of
Open
Source,
pdk
and
designs,
so
some
numbers
that
are
kind
of
new.
B
So
this
is
the
number
of
projects
is
increasing
over
time
and
the
also
there's
also
I,
don't
have
the
npw7
here
but
again
and
seven
is
oversubscribed,
and
then
we
have
other
projects
are
coming
from
the
the
the
GF
180.
It's
worth
mentioning
that
the
JF
180
deadline
was
on
December
5th.
We
had
that
about
20
participants
over
40
slots
available
or
it
it
was
over
the
weekend.
It
ended
up
with
90.
B
and
and
of
which
about
70,
completely
clean,
ready
to
tape
out.
C
B
The
lottery
had
to
be
in
enacted,
of
course,
to
because
we
don't
want
to
have
the
bias
any
selection
based
on
judgment
or
of
someone
as
long
as
it's
manufacturable.
This
is
how
Innovation
actually
works.
On
the
other
side,
everything
is
going
in
the
right
direction.
In
terms
of
you
know,
the
pdk
download
one
important
thing
that
happened
over
that
last
two
weeks.
B
This
is
not
a
person
that
he
was
downloading
twice
or
this
is
unique
for
open,
Lane
I,
don't
have
the
data
for
the
pdk
yet,
but
it
is
actually.
It
says
something
that,
within
the
last
two
weeks,
you
had
that
number
of
unique
cloners
of
the
PDQ,
which
is
very
impressive
in
terms
of
the
cover
of
the
open,
Lane
I'm.
Sorry
because
opening
is
actually
a
common
flow,
that's
being
used
for
both
the
GF
180
MCU
and
the
sky
130.
B
yeah
again,
this
picture
probably
seen
it
before,
but
it's
a
representation
of
two
shuttles.
This
is
how
they
get
distributed
and
made
available
to
The
Foundry.
B
Some
of
the
examples
that
are
on
the
shuttles
are
very
creative
and
good
ideas.
You
can
see
I
mentioned
earlier
that,
because
of
the
simplification
of
the
entire
process,
more
designers
are
getting
in,
which
is
an
important
way
around
validation.
Also
for
the
simplification.
We
do
need
more.
So
it's
not
we're
not
in
a
way
to
we.
We
want
to
have
the
ability.
If
you
can
code,
you
can
make
your
own
trip
and
also
optimize
it
with
your
software.
B
Mpw2
came
as
well
now
and
then
that
this
is
new
material
here,
looking
at
them,
this
is
actually
the
shuttle
npw2,
but
I
have
put
the
also,
although
since
the
form
factor
looks
different,
but
it
is
really.
This
is
the
the
the
array
of
the
parts
all
loaded
on
a
a
breakout
board
for
the
board
that
we
have
we
one
of
the
things
we
did
in
epw2.
We
actually
added
to
our
transition
temperature
forcing
station.
B
This
is
at
home
so
and
it's
a
little
bit
noisy,
but
it's
really
great
and
with
a
test
setup
that
you
can
actually
be
recommended
to
the
community
for
others
without
the
temperature
aspect
of
it,
you
can
replicate
it.
We
we.
This
is
the
latest
board
for
mpw2
mpw
for
Carnival
period,
actually
for
Indian.
B
So
on
the
top
left,
that's
the
carnival
board,
and
if
you
can,
you
can
see
here
that
we
use
the
flexi
pins
for
the
purpose
of
making
it
easy
to
pop
the
part
and
put
it
back
in
and
the
if
you
get
300
parts
you
can
actually
do
that,
and
the
other
thing
is
that
this
board
was
created
with
a
dual
mode.
One
of
them
is
a
hat
has
a
bolt
into
the
board.
B
That
is
the
board
that
is
in
the
bottom
here,
which
is
the
nuclear
board
if
746
as
a
tester
so
versus
you
can
put
them
together
and
you
can
use
that
as
a
as
a
tester
or
you
can
just
put
it
stand
alone
and
plug
it
into
a
USB
and
you
can
utilize
values.
B
Why
did
we
do
that?
This
was
a
story
that
I,
you
know,
I'm
I'm,
not
putting
enough
details
in
it
here,
but
basically
we
there
was
a
timing
bug
in
mpw2
and
three
and
we
found
the
software
solution.
That
depends
on
the
chip
itself
and
the
data
type,
so
whether
it's
a
data,
the
data,
independent
data,
independent
type
of
timing
issue.
B
So,
but
we
don't
want
to
make
it
that
complex
process
in
the
hands
of
the
user,
because
it's
too
much
and
and
it's
actually
you
know
it-
is
too
much
even
for
an
expert
user.
So
this
nuclear
board
does
two
jobs.
One
of
them
is
to
when
you
put
the
board
on
it,
it
will
actually
find
out
characterize
the
chip.
B
It
will
characterize
the
chip
from
a
functional
perspective
and
then
find
out
the
timing
issue
and
what
and
what
algorithm
was
used
to
solve
the
timing
by
from
the
software
side,
and
it
will
apply
it
and
tell
you
you
know
the
chip
is
good
and
you
can
use
it
in
as
transparently.
You
don't
have
to
use
the
continue
to
use
the
test
board.
You
you
can
once
you
find
out
the
that
the
the
set
of
values
that
you
need
to
use
for
that
chip
it
have
this
timing
issue
is
a
chip
by
chip.
B
So
if
you
have
300
parts,
you
would
want
to
have
a
system
like
this
to
actually
put
it
in.
It
will
immediately
within
seconds,
tell
you
the
part
is
working
and
we
can
and
gives
you
the
values
of
the
software
programming
that
you
can
actually
add
to
make
it
completely
smooth
one
of
the
things
that
jumping
ahead.
B
This
slide
is
I
wanted
to
say
that
it's
actually
old
and
the
reason
I
printed
it
up
here
is
that
a
lot
of
the
stuff
that's
happened
here
is,
it
is,
is
it
was,
was
predicted,
I
didn't
have
Global
foundries
here,
I
mean
I
had
a
next
Foundry
here,
and
this
is
something
that
you
know.
B
We
projected
together
with
Tim
Ansel,
that
this
is
going
to
happen
and
it
happens
and
there's
more
and
more
coming,
but
it
is
important
to
highlight
that
if
we're
going
to
go
talk
about
these
applications
and
solve
real
problems
in
the
world
around
the
world
with
open
source,
IP
or
blocks,
we
need
to
continue
to
do
to
look
at
the
picture
and
see
what
what
Solutions
I
would
actually
add
at
top
level
here
for
actual
applications.
But
we've
been
some
Moon
doing
some
movement
here.
B
Now
these
are
you
know,
I'm
going
to
show
just
a
few
creative
aspects
here
now
our
friend
Tim
I'm,
sorry
madwin
has
created
the
concept
of
tiny
tape,
Health,
which
actually
does
the
it
takes
the
10
millimeter
square,
divided
into
500
pieces
with
a
small
editor
that
actually
do
it,
and
that
actually
is
an
interesting
experiment
when
it
happened,
because
it
shows
that
are
people
that
are
interested
in
doing
that
and
and
actually
have
a
hundred
by
100
100
square
at
100
millimeters.
B
In
as
a
part
of
the
the
chip-
and
these
are
all
addressable
from
from
the
tips
of
when
Matt
gets
the
board
and
they
can
actually
get
it
funded
without
Google,
so
this
is
how
it
would
get
positioned.
So
it's
a
really
showed
a
lot
of
interest
around
the
the
youngest
person
that
got
on
that
or
designed
the
design
is
about
nine
years
old.
B
We
other
examples
of
the
people
who
did
design
the
eye.
They
have
a
program
full
program
for
the
the
so-called
the
IEEE
Peak
program.
It's
a
competition,
so
the
carabelle
is
used
for
multiple
projects,
as
you
can
see
together
this
year,
there's
another
one
that
looks
like
that,
except
that
the
layouts
are
not
are
about
to
be
finished,
but
look
at
the
diversity
of
designs,
and
these
you
know
how
how
the
what
was
this
is.
B
The
22
that
were
selected
for
moving
forward
and-
and
that
actually
is
is
in
is
great
because
it
shows
that
we're
only
not
have
a
digital.
We
also
have
an
analog,
and
we
started
to
thinking
about
enriching
the
library
of
available
for
open
source
blocks
for
everyone,
some
of
the
Silicon
that
came
back
and,
and
some
it
you
know,
depends
on
some
of
it.
B
You
know,
even
though
we
had
some
bugs
and
issues
the
people
were
able
to
make
it
work
this
one,
the
LA,
the
left
ones
were
going
to
be
one
on
the
right
one
from
Chip
ignite,
which
is
the
offering
that
defense
does
for
University
and
startup,
and
you
can
see
this
one
here
is
a
bug
converter,
and
this
is
Jeep
converter
and
it's
available
also
online,
and
these
are
some
of
the
things
that
were
done
between
Matlin
and
and
TNT
or
something
it
shows
again
the
diversity
of
how
people
connect
together
and
do
things
together
that
they
would
otherwise
wouldn't
have
been
possible.
B
Madwin
has
created
a
whole
course
to
that
gets
a
person
from
zero
to
making
a
small
list
a
very
good
work
and
other
examples
I'm
just
going
to
flash
through
them.
You
know
this
was
androgenerberg
from
from
the
very
first
chip
for
the
test
chip
for
the
ram.
Now,
what
we
need
to
make
Caravel
as
a
platform
richer
for
the
open
source
and
we
need
open
source
blocks
that
are
actually
available
to
serve
that
these
products
and
so
real
human
needed
problems.
B
As
I
said,
you
know
it
takes
a
village,
so
it
takes
a
community
a
journey
with
the
community,
and
what
we're
doing
here
is
disrupting
really
a
stale
industry
methodologies
and
the
approach
to
developing
chips.
Our
Hardware
hasn't
really
changed
much
in
a
significant
Delta.
B
It's
a
step
function
for
for
for
you,
for
at
least
you
know
a
few
decades,
so
the
we
need
to
get
into
that
library
and
start
thinking
about
how
do
what
do
we
need
and
to
serve
the
application
sensor,
interfaces
data,
converters,
communication,
there's
a
Bluetooth
device
that
is
being
done
by
University
of
Michigan.
We
need
this
library
to
start
getting
into
building
the
useful
chips
and
a
useful
chip
is
something
that
you
can
actually
put
it
in
a
watch
or
put
it
in
a
water
filter.
A
water
monitor
a
smoke
detector.
B
These
are
things
that
we
need
to
think
about
to
take
this
entire
initiative.
You
know
a
couple
of
levels
up
one
of
the
things
that
you
can
do
also,
then,
on
the
that
that
is
possible
is
actually
you
can.
If
you
don't
get
the
funding
from
Google,
because
some
community
members
were
concerned
or
not
concerned
did.
A
B
Have
that
option
because
they
didn't
get
on
the
lottery
ticket
they
can
submit
for
the
next
shuttle,
but
they
needed
to
control
some
that
some
of
the
time.
So
again
it
takes
a
village
in
this
way
crowdfunding.
So
we
tried
a
couple
examples:
one
on
our
side.
It's
called
clear
and
that
was
an
embedded
fpga
inside
Caravel
and
yeah.
It
was
funded
and
it's
in
manufacturing
now,
but
and
then
we
have
another
example
on
crowd,
Supply
that
it's
called
Maverick.
B
That
is
done
by
another
team
that
is
using
Caravel
for
ft8
receiver.
B
Then
he
is
that
that's
what
all
I
have
for
now.
Yeah
I
I
wanted
to
I
wish
I
had
more
opportunity
to
add
more
details,
but
I
related
for
questions.
A
No,
that's
great
thanks,
so
much
Muhammad,
so
we
have
time
for
a
question
or
two
and
then
we
do
need
to
get
to
our
board
meeting
so
take
the
microphone
on
the
back
here.
C
I
think
that
might
be
the
youngest
person
to
ever
tape
out
every
nine
years
old
hi.
It's
really
cool
I
was
wondering
if
there's
any
plans
for
larger
than
10
millimeter
squared
area
for
open
source
projects.
B
So
let
me
I'll
I'll
just
share
the
screen
real,
quick
to
our
visual.
It's
already
shared,
it's
still
shared
right
Okay.
So
so
we
are
looking
at
things
more
thing,
but
this
is
something
with
for
the
Google
shuttle
we
obviously
it's
Google
funded.
So
Google
has
been
you
know
a
lot
of
say
in
it,
but
we're
considering
the
main
thing
in
giving
the
cost
down
and
make
it
possible
and
easiest
to
actually
get
into
that
tile
side.
B
So
there
is
a
consideration
that
literally
an
idea
on
a
napkin
we
haven't
put
it
yet
is
to
maybe
have
a
a
double
area:
double
wide
double
height
or
double
or
quad.
Okay,
the
reason
we
don't
do
this
right
away,
because
we
don't
do
we
don't
have
a
table
team.
This
is
done
by
a
machine.
Okay,
it
gets
organized,
it
gets
flushed,
it
gets
into
The
Foundry.
So
if
we're
gonna
do
have
four,
that's
a
that's
an
outlier
in
the
process,
an
automation
problem.
B
So
it's
not
just
about
removing
a
couple
layouts
and
putting
a
bigger
one,
because
we
don't
do
that.
We
want
to
be
able
to
do
that
for
hundreds
and
of
shuttles,
and
you
know,
or
maybe
thousands
or
something
and
we
can
just
move
the
layout.
Is
there
a
plan
not
on
the
schedule?
Is
there
a
consideration
based
on
what
we
see
from
the
community.
A
C
Hi
Muhammad,
how
did
this
is
David
kellett,
so
you
had
the
the
easier
design
path
and
the
expert
design
path
so
on
the
easier
design
path.
What
can
I
do
with
your
capability
here
that
I
can't
do
with
a
small
fpga.
B
C
Where
I
could
go
to
expert
mode
use
the
whole
frame
and
I'm
talking
about
the
kind
of
the
top
the
top
row?
Yes,
oh.
B
B
There
are
no
chips
that
actually
do
anything
that
the
phone
wants.
The
application
processor
didn't
exist,
so
it
would.
It
had
to
be
low
power
because
and
it
had
to
support
gaming
later
so
the
choice,
if
you
can
do
it
in
apga,
for
a
product
and
it's
Unique
and
you
can
sell
that
product,
probably
and
I,
make
it
viable.
Then
why
do
I
need?
But
people
do
an
Asic
for
potentially
foreign
Factor
like
instead
of,
for
example,
production
or
power,
maybe
costs
and
meals
to
the
cases.
We
have
a
really
high
value
foreign.
C
B
C
Hey
this
should
be
an
easy
one.
Do
you
have
a
basically
like
a
top
10
of
your
wish
list
for
IP
blocks
that
you
need
for
the
pdk.
B
I
started
I,
both
actually
teammates
and
I
started
so,
but
we
have
I
have
a
list
of
the
IP
existing
IP
and
I
have
a
list
of
wanted
IP
on
a
GitHub
Ripple
I
want
to
publish
it
to
make
it
a
public
repository.
Oh
that's
my
fault,
but
now
that
I
made
a
call
to
action.
I'd
like
to
do
that.
Thank
you
for
the
question.
Thank
you.
A
A
And
thank
you
everyone
for
attending
today
and
hanging
in
there
with
us.
I
know
we
ran
way
over
time
today,
but
all
the
talks
were
very
good.
So
thank
you
and
for
signing
off
here
from
Sunnyvale.