<s>
Explicit	B-Operating_System
Multi-Threading	I-Operating_System
(	O
XMT	O
)	O
is	O
a	O
computer	B-General_Concept
science	I-General_Concept
paradigm	O
for	O
building	O
and	O
programming	O
parallel	B-Operating_System
computers	I-Operating_System
designed	O
around	O
the	O
parallel	B-Operating_System
random-access	I-Operating_System
machine	I-Operating_System
(	O
PRAM	B-Operating_System
)	O
parallel	O
computational	O
model	O
.	O
</s>
<s>
A	O
more	O
direct	O
explanation	O
of	O
XMT	O
starts	O
with	O
the	O
rudimentary	O
abstraction	O
that	O
made	O
serial	B-Device
computing	I-Device
simple	O
:	O
that	O
any	O
single	O
instruction	O
available	O
for	O
execution	O
in	O
a	O
serial	O
program	O
executes	O
immediately	O
.	O
</s>
<s>
The	O
rudimentary	O
parallel	O
abstraction	O
behind	O
XMT	O
,	O
dubbed	O
Immediate	O
Concurrent	B-Operating_System
Execution	O
(	O
ICE	O
)	O
in	O
,	O
is	O
that	O
indefinitely	O
many	O
instructions	O
available	O
for	O
concurrent	B-Operating_System
execution	O
execute	O
immediately	O
.	O
</s>
<s>
A	O
consequence	O
of	O
ICE	O
is	O
a	O
step-by-step	O
(	O
inductive	O
)	O
explication	O
of	O
the	O
instructions	O
available	O
next	O
for	O
concurrent	B-Operating_System
execution	O
.	O
</s>
<s>
Moving	O
beyond	O
the	O
serial	O
von	B-Architecture
Neumann	I-Architecture
computer	I-Architecture
(	O
the	O
only	O
successful	O
general-purpose	O
platform	O
to	O
date	O
)	O
,	O
the	O
aspiration	O
of	O
XMT	O
is	O
that	O
computer	B-General_Concept
science	I-General_Concept
will	O
again	O
be	O
able	O
to	O
augment	O
mathematical	O
induction	O
with	O
a	O
simple	O
one-line	O
computing	O
abstraction	O
.	O
</s>
<s>
The	O
random-access	B-Application
machine	I-Application
(	O
RAM	O
)	O
is	O
an	O
abstract	B-Application
machine	I-Application
model	O
used	O
in	O
computer	B-General_Concept
science	I-General_Concept
to	O
study	O
algorithms	O
and	O
complexity	O
for	O
standard	O
serial	B-Device
computing	I-Device
.	O
</s>
<s>
The	O
PRAM	B-Operating_System
computational	O
model	O
is	O
an	O
abstract	O
parallel	B-Operating_System
machine	I-Operating_System
model	O
that	O
had	O
been	O
introduced	O
to	O
similarly	O
study	O
parallel	O
algorithms	O
and	O
complexity	O
for	O
parallel	B-Operating_System
computing	I-Operating_System
,	O
when	O
they	O
were	O
yet	O
to	O
be	O
built	O
.	O
</s>
<s>
Researchers	O
have	O
developed	O
a	O
large	O
body	O
of	O
knowledge	O
of	O
parallel	O
algorithms	O
for	O
the	O
PRAM	B-Operating_System
model	O
.	O
</s>
<s>
This	O
large	O
body	O
of	O
parallel	O
algorithms	O
knowledge	O
for	O
the	O
PRAM	B-Operating_System
model	O
and	O
their	O
relative	O
simplicity	O
motivated	O
building	O
computers	O
whose	O
programming	O
can	O
be	O
guided	O
by	O
these	O
parallel	O
algorithms	O
.	O
</s>
<s>
Since	O
productivity	O
of	O
parallel	O
programmers	O
has	O
long	O
been	O
considered	O
crucial	O
for	O
the	O
success	O
a	O
parallel	B-Operating_System
computer	I-Operating_System
,	O
simplicity	O
of	O
algorithms	O
is	O
important	O
.	O
</s>
<s>
Multi-core	B-Architecture
computers	O
are	O
built	O
around	O
two	O
or	O
more	O
processor	O
cores	O
integrated	O
on	O
a	O
single	O
integrated	O
circuit	O
die	O
.	O
</s>
<s>
Explicit	B-Operating_System
Multi-Threading	I-Operating_System
(	O
XMT	O
)	O
is	O
a	O
computing	O
paradigm	O
for	O
building	O
and	O
programming	O
multi-core	B-Architecture
computers	O
with	O
tens	O
,	O
hundreds	O
or	O
thousands	O
of	O
processor	O
cores	O
.	O
</s>
<s>
Experimental	O
work	O
published	O
in	O
2011	O
and	O
2012	O
demonstrates	O
significantly	O
greater	O
speedups	O
for	O
advanced	O
PRAM	B-Operating_System
algorithms	O
on	O
XMT	O
prototypes	O
than	O
for	O
the	O
same	O
problems	O
on	O
state-of-the-art	O
multi-core	B-Architecture
computers	O
.	O
</s>
<s>
Work	O
published	O
in	O
2018	O
shows	O
that	O
lock-step	O
parallel	B-Operating_System
programming	I-Operating_System
(	O
using	O
ICE	O
)	O
can	O
achieve	O
the	O
same	O
performance	O
as	O
the	O
fastest	O
hand-tuned	O
multi-threaded	O
code	O
on	O
XMT	O
systems	O
.	O
</s>
<s>
The	O
Explicit	B-Operating_System
Multi-Threading	I-Operating_System
(	O
XMT	O
)	O
computing	O
paradigm	O
integrates	O
several	O
levels	O
of	O
abstraction	O
.	O
</s>
<s>
For	O
example	O
,	O
the	O
WT	O
framework	O
was	O
adopted	O
as	O
the	O
basic	O
presentation	O
framework	O
in	O
the	O
parallel	O
algorithms	O
books	O
(	O
for	O
the	O
PRAM	B-Operating_System
model	O
)	O
and	O
,	O
as	O
well	O
as	O
in	O
the	O
class	O
notes	O
.	O
</s>
<s>
The	O
XMT	O
paradigm	O
can	O
be	O
programmed	O
using	O
XMTC	B-Language
,	O
a	O
parallel	O
multi-threaded	O
programming	O
language	O
which	O
is	O
a	O
small	O
extension	O
of	O
the	O
programming	O
language	O
C	O
.	O
The	O
XMT	O
paradigm	O
include	O
a	O
programmer	O
's	O
workflow	O
that	O
starts	O
with	O
casting	O
an	O
algorithm	O
in	O
the	O
WT	O
framework	O
and	O
proceeds	O
to	O
programming	O
it	O
in	O
XMTC	B-Language
.	O
</s>
<s>
The	O
XMT	O
multi-core	B-Architecture
computer	O
systems	O
provides	O
run-time	O
load-balancing	O
of	O
multi-threaded	O
programs	O
incorporating	O
several	O
patents	O
.	O
</s>
<s>
One	O
of	O
them	O
generalizes	O
the	O
program	B-General_Concept
counter	I-General_Concept
concept	O
,	O
which	O
is	O
central	O
to	O
the	O
von	B-Architecture
Neumann	I-Architecture
architecture	I-Architecture
to	O
multi-core	B-Architecture
hardware	O
.	O
</s>
<s>
Since	O
making	O
parallel	B-Operating_System
programming	I-Operating_System
easy	O
is	O
one	O
of	O
the	O
biggest	O
challenges	O
facing	O
computer	B-General_Concept
science	I-General_Concept
today	O
,	O
the	O
demonstration	O
also	O
sought	O
to	O
include	O
teaching	O
the	O
basics	O
of	O
PRAM	B-Operating_System
algorithms	O
and	O
XMTC	B-Language
programming	O
to	O
students	O
ranging	O
from	O
high-school	O
to	O
graduate	O
school	O
.	O
</s>
<s>
Experimental	O
work	O
reported	O
in	O
for	O
the	O
Maximum	B-Algorithm
flow	I-Algorithm
problem	I-Algorithm
,	O
and	O
in	O
two	O
papers	O
by	O
for	O
the	O
Graph	O
Connectivity	O
(	O
Connectivity	O
(	O
graph	O
theory	O
)	O
)	O
,	O
Graph	O
Biconnectivity	O
(	O
biconnected	O
graph	O
)	O
and	O
Graph	O
Triconnectivity	O
(	O
Triconnected	B-Data_Structure
component	I-Data_Structure
)	O
problems	O
demonstrated	O
that	O
for	O
some	O
of	O
the	O
most	O
advanced	O
algorithms	O
in	O
the	O
parallel	O
algorithmic	O
literature	O
,	O
the	O
XMT	O
paradigm	O
can	O
offer	O
8	O
times	O
to	O
over	O
100	O
times	O
greater	O
speedups	O
than	O
for	O
the	O
same	O
problems	O
on	O
state-of-the-art	O
multi-core	B-Architecture
computers	O
.	O
</s>
<s>
XMT	O
prototyping	O
was	O
culminated	O
in	O
,	O
establishing	O
that	O
lock-step	O
parallel	B-Operating_System
programming	I-Operating_System
(	O
using	O
ICE	O
)	O
can	O
achieve	O
the	O
same	O
performance	O
as	O
the	O
fastest	O
hand-tuned	O
multi-threaded	O
code	O
on	O
XMT	O
systems	O
.	O
</s>
