<s>
Zettascale	B-General_Concept
computing	I-General_Concept
refers	O
to	O
computing	O
systems	O
capable	O
of	O
calculating	O
at	O
least	O
"	O
1021	O
IEEE	O
754	O
Double	O
Precision	O
(	O
64-bit	O
)	O
operations	O
(	O
multiplications	O
and/or	O
additions	O
)	O
per	O
second	O
(	O
zettaFLOPS	O
)	O
"	O
.	O
</s>
<s>
It	O
is	O
a	O
measure	O
of	O
supercomputer	B-Architecture
performance	O
,	O
and	O
is	O
a	O
hypothetical	O
performance	O
barrier	O
.	O
</s>
<s>
FLOPS	O
can	O
be	O
recorded	O
in	O
different	O
measures	O
of	O
precision	O
,	O
however	O
the	O
standard	O
measure	O
(	O
used	O
by	O
the	O
TOP500	B-Operating_System
supercomputer	B-Architecture
list	O
)	O
uses	O
64	O
bit	O
(	O
double-precision	O
floating-point	O
format	O
)	O
operations	O
per	O
second	O
using	O
the	O
High	O
Performance	O
LINPACK	O
(	O
HPLinpack	O
)	O
benchmark	O
.	O
</s>
<s>
This	O
forecast	O
looks	O
plausible	O
from	O
the	O
historical	O
point	O
of	O
view	O
as	O
it	O
took	O
some	O
12	O
years	O
to	O
progress	O
from	O
the	O
terascale	O
machines	O
(	O
1012	O
)	O
to	O
petascale	B-General_Concept
systems	O
(	O
1015	O
)	O
and	O
then	O
14	O
more	O
years	O
to	O
move	O
to	O
exascale	B-General_Concept
computers	I-General_Concept
(	O
1018	O
)	O
.	O
</s>
<s>
As	O
Moore	O
's	O
law	O
nears	O
its	O
natural	O
limits	O
,	O
supercomputing	B-Architecture
will	O
face	O
serious	O
physical	O
problems	O
to	O
move	O
from	O
exascale	B-General_Concept
to	O
zettascale	O
systems	O
,	O
making	O
the	O
decade	O
after	O
2020	O
a	O
vital	O
period	O
to	O
develop	O
key	O
high-performance	B-Architecture
computing	I-Architecture
techniques	O
.	O
</s>
