<s>
Massively	B-Operating_System
parallel	I-Operating_System
is	O
the	O
term	O
for	O
using	O
a	O
large	O
number	O
of	O
computer	O
processors	O
(	O
or	O
separate	O
computers	O
)	O
to	O
simultaneously	O
perform	O
a	O
set	O
of	O
coordinated	O
computations	O
in	B-Operating_System
parallel	I-Operating_System
.	O
</s>
<s>
GPUs	B-Architecture
are	O
massively	B-Operating_System
parallel	I-Operating_System
architecture	O
with	O
tens	O
of	O
thousands	O
of	O
threads	O
.	O
</s>
<s>
One	O
approach	O
is	O
grid	B-Architecture
computing	I-Architecture
,	O
where	O
the	O
processing	O
power	O
of	O
many	O
computers	O
in	O
distributed	O
,	O
diverse	O
administrative	B-Protocol
domains	I-Protocol
is	O
opportunistically	O
used	O
whenever	O
a	O
computer	O
is	O
available	O
.	O
</s>
<s>
An	O
example	O
is	O
BOINC	B-Operating_System
,	O
a	O
volunteer-based	B-Operating_System
,	O
opportunistic	O
grid	O
system	O
,	O
whereby	O
the	O
grid	O
provides	O
power	O
only	O
on	O
a	O
best	O
effort	O
basis	O
.	O
</s>
<s>
Another	O
approach	O
is	O
grouping	O
many	O
processors	O
in	O
close	O
proximity	O
to	O
each	O
other	O
,	O
as	O
in	O
a	O
computer	B-Architecture
cluster	I-Architecture
.	O
</s>
<s>
In	O
such	O
a	O
centralized	O
system	O
the	O
speed	O
and	O
flexibility	O
of	O
the	O
interconnect	B-General_Concept
becomes	O
very	O
important	O
,	O
and	O
modern	O
supercomputers	B-Architecture
have	O
used	O
various	O
approaches	O
ranging	O
from	O
enhanced	O
InfiniBand	B-Architecture
systems	O
to	O
three-dimensional	O
torus	B-Operating_System
interconnects	I-Operating_System
.	O
</s>
<s>
The	O
term	O
also	O
applies	O
to	O
massively	B-General_Concept
parallel	I-General_Concept
processor	I-General_Concept
arrays	I-General_Concept
(	O
MPPAs	O
)	O
,	O
a	O
type	O
of	O
integrated	O
circuit	O
with	O
an	O
array	O
of	O
hundreds	O
or	O
thousands	O
of	O
central	B-General_Concept
processing	I-General_Concept
units	I-General_Concept
(	O
CPUs	O
)	O
and	O
random-access	B-Architecture
memory	I-Architecture
(	O
RAM	B-Architecture
)	O
banks	O
.	O
</s>
<s>
These	O
processors	O
pass	O
work	O
to	O
one	O
another	O
through	O
a	O
reconfigurable	O
interconnect	B-General_Concept
of	O
channels	O
.	O
</s>
<s>
By	O
harnessing	O
many	O
processors	O
working	O
in	B-Operating_System
parallel	I-Operating_System
,	O
an	O
MPPA	O
chip	O
can	O
accomplish	O
more	O
demanding	O
tasks	O
than	O
conventional	O
chips	O
.	O
</s>
<s>
MPPAs	O
are	O
based	O
on	O
a	O
software	O
parallel	B-Operating_System
programming	I-Operating_System
model	O
for	O
developing	O
high-performance	O
embedded	O
system	O
applications	O
.	O
</s>
<s>
Goodyear	B-Device
MPP	I-Device
was	O
an	O
early	O
implementation	O
of	O
a	O
massively	B-Operating_System
parallel	I-Operating_System
computer	I-Operating_System
architecture	O
.	O
</s>
<s>
MPP	O
architectures	O
are	O
the	O
second	O
most	O
common	O
supercomputer	B-Architecture
implementations	O
after	O
clusters	O
,	O
as	O
of	O
November	O
2013	O
.	O
</s>
<s>
Data	O
warehouse	O
appliances	O
such	O
as	O
Teradata	O
,	O
Netezza	O
or	O
Microsoft	O
's	O
PDW	O
commonly	O
implement	O
an	O
MPP	O
architecture	O
to	O
handle	O
the	O
processing	O
of	O
very	O
large	O
amounts	O
of	O
data	O
in	B-Operating_System
parallel	I-Operating_System
.	O
</s>
