<s>
Exascale	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
1018	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
exaFLOPS	O
)	O
"	O
;	O
it	O
is	O
a	O
measure	O
of	O
supercomputer	B-Architecture
performance	O
.	O
</s>
<s>
Exascale	B-General_Concept
computing	I-General_Concept
is	O
a	O
significant	O
achievement	O
in	O
computer	O
engineering	O
:	O
primarily	O
,	O
it	O
allows	O
improved	O
scientific	O
applications	O
and	O
better	O
prediction	O
accuracy	O
in	O
domains	O
such	O
as	O
weather	O
forecasting	O
,	O
climate	B-Application
modeling	I-Application
and	O
personalised	O
medicine	O
.	O
</s>
<s>
Exascale	B-General_Concept
also	O
reaches	O
the	O
estimated	O
processing	O
power	O
of	O
the	O
human	O
brain	O
at	O
the	O
neural	O
level	O
,	O
a	O
target	O
of	O
the	O
Human	O
Brain	O
Project	O
.	O
</s>
<s>
There	O
has	O
been	O
a	O
race	O
to	O
be	O
the	O
first	O
country	O
to	O
build	O
an	O
exascale	B-General_Concept
computer	I-General_Concept
,	O
typically	O
ranked	O
in	O
the	O
TOP500	B-Operating_System
list	O
.	O
</s>
<s>
In	O
2022	O
,	O
the	O
world	O
's	O
first	O
public	O
exascale	B-General_Concept
computer	I-General_Concept
,	O
Frontier	B-Device
,	O
was	O
announced	O
.	O
</s>
<s>
,	O
it	O
is	O
the	O
world	O
's	O
fastest	B-Operating_System
supercomputer	I-Operating_System
.	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>
Whilst	O
a	O
distributed	B-Architecture
computing	I-Architecture
system	O
had	O
broken	O
the	O
1	O
exaFLOPS	O
barrier	O
before	O
Frontier	B-Device
,	O
the	O
metric	O
typically	O
refers	O
to	O
single	O
computing	O
systems	O
.	O
</s>
<s>
Supercomputers	B-Architecture
had	O
also	O
previously	O
broken	O
the	O
1	O
exaFLOPS	O
barrier	O
using	O
alternative	O
precision	O
measures	O
;	O
again	O
these	O
do	O
not	O
meet	O
the	O
criteria	O
for	O
exascale	B-General_Concept
computing	I-General_Concept
using	O
the	O
standard	O
metric	O
.	O
</s>
<s>
It	O
has	O
been	O
recognised	O
that	O
HPLinpack	O
may	O
not	O
be	O
a	O
good	O
general	O
measure	O
of	O
supercomputer	B-Architecture
utility	O
in	O
real	O
world	O
application	O
,	O
however	O
it	O
is	O
the	O
common	O
standard	O
for	O
performance	O
measurement	O
.	O
</s>
<s>
It	O
has	O
been	O
recognized	O
that	O
enabling	O
applications	O
to	O
fully	O
exploit	O
capabilities	O
of	O
exascale	B-General_Concept
computing	I-General_Concept
systems	O
is	O
not	O
straightforward	O
.	O
</s>
<s>
Developing	O
data-intensive	O
applications	O
over	O
exascale	B-General_Concept
platforms	O
requires	O
the	O
availability	O
of	O
new	O
and	O
effective	O
programming	O
paradigms	O
and	O
runtime	O
systems	O
.	O
</s>
<s>
The	O
Folding	B-Application
@home	I-Application
project	O
,	O
the	O
first	O
to	O
break	O
this	O
barrier	O
,	O
relied	O
on	O
a	O
network	O
of	O
servers	O
sending	O
pieces	O
of	O
work	O
to	O
hundreds	O
of	O
thousands	O
of	O
clients	O
using	O
a	O
client	B-Operating_System
–	I-Operating_System
server	I-Operating_System
model	I-Operating_System
network	B-Architecture
architecture	I-Architecture
.	O
</s>
<s>
The	O
first	O
petascale	B-General_Concept
(	O
1015	O
FLOPS	O
)	O
computer	O
entered	O
operation	O
in	O
2008	O
.	O
</s>
<s>
At	O
a	O
supercomputing	B-Architecture
conference	O
in	O
2009	O
,	O
Computerworld	O
projected	O
exascale	B-General_Concept
implementation	O
by	O
2018	O
.	O
</s>
<s>
In	O
June	O
2014	O
,	O
the	O
stagnation	O
of	O
the	O
Top500	B-Operating_System
supercomputer	B-Architecture
list	O
had	O
observers	O
question	O
the	O
possibility	O
of	O
exascale	B-General_Concept
systems	O
by	O
2020	O
.	O
</s>
<s>
Although	O
exascale	B-General_Concept
computing	I-General_Concept
was	O
not	O
achieved	O
by	O
2018	O
,	O
in	O
the	O
same	O
year	O
the	O
Summit	B-Device
OLCF-4	I-Device
supercomputer	I-Device
performed	O
1.8	O
calculations	O
per	O
second	O
using	O
an	O
alternative	O
metric	O
whilst	O
analysing	O
genomic	O
information	O
.	O
</s>
<s>
The	O
team	O
performing	O
this	O
won	O
the	O
Gordon	O
Bell	O
Prize	O
at	O
the	O
2018	O
ACM/IEEE	O
Supercomputing	B-Architecture
Conference	O
.	O
</s>
<s>
The	O
exaFLOPS	O
barrier	O
was	O
first	O
broken	O
in	O
March	O
2020	O
by	O
the	O
distributed	B-Architecture
computing	I-Architecture
network	O
Folding	B-Application
@home	I-Application
coronavirus	O
research	O
project	O
.	O
</s>
<s>
In	O
June	O
2020	O
the	O
Japanese	O
supercomputer	B-Architecture
Fugaku	B-Device
achieved	O
1.42	O
exaFLOPS	O
using	O
the	O
alternative	O
HPL-AI	O
benchmark	O
.	O
</s>
<s>
In	O
2022	O
,	O
the	O
world	O
's	O
first	O
public	O
exascale	B-General_Concept
computer	I-General_Concept
,	O
Frontier	B-Device
,	O
was	O
announced	O
.	O
</s>
<s>
,	O
it	O
is	O
the	O
world	O
's	O
fastest	B-Operating_System
supercomputer	I-Operating_System
.	O
</s>
<s>
In	O
2008	O
,	O
two	O
United	O
States	O
of	O
America	O
governmental	O
organisations	O
within	O
the	O
US	O
Department	O
of	O
Energy	O
,	O
the	O
Office	O
of	O
Science	O
and	O
the	O
National	O
Nuclear	O
Security	O
Administration	O
,	O
provided	O
funding	O
to	O
the	O
Institute	O
for	O
Advanced	O
Architectures	O
for	O
the	O
development	O
of	O
an	O
exascale	B-General_Concept
supercomputer	B-Architecture
;	O
Sandia	O
National	O
Laboratory	O
and	O
the	O
Oak	O
Ridge	O
National	O
Laboratory	O
were	O
also	O
to	O
collaborate	O
on	O
exascale	B-General_Concept
designs	O
.	O
</s>
<s>
In	O
January	O
2012	O
,	O
Intel	O
purchased	O
the	O
InfiniBand	B-Architecture
product	O
line	O
from	O
QLogic	O
for	O
US$125	O
million	O
in	O
order	O
to	O
fulfill	O
its	O
promise	O
of	O
developing	O
exascale	B-General_Concept
technology	O
by	O
2018	O
.	O
</s>
<s>
By	O
2012	O
,	O
the	O
United	O
States	O
had	O
allotted	O
$126	O
million	O
for	O
exascale	B-General_Concept
computing	I-General_Concept
development	O
.	O
</s>
<s>
In	O
February	O
2013	O
,	O
the	O
Intelligence	O
Advanced	O
Research	O
Projects	O
Activity	O
started	O
the	O
Cryogenic	O
Computer	O
Complexity	O
(	O
C3	O
)	O
program	O
,	O
which	O
envisions	O
a	O
new	O
generation	O
of	O
superconducting	O
supercomputers	B-Architecture
that	O
operate	O
at	O
exascale	B-General_Concept
speeds	O
based	O
on	O
superconducting	O
logic	O
.	O
</s>
<s>
On	O
29	O
July	O
2015	O
,	O
Barack	O
Obama	O
signed	O
an	O
executive	O
order	O
creating	O
a	O
National	B-General_Concept
Strategic	I-General_Concept
Computing	I-General_Concept
Initiative	I-General_Concept
calling	O
for	O
the	O
accelerated	O
development	O
of	O
an	O
exascale	B-General_Concept
system	O
and	O
funding	O
research	O
into	O
post-semiconductor	O
computing	O
.	O
</s>
<s>
The	O
Exascale	B-General_Concept
Computing	I-General_Concept
Project	O
(	O
ECP	O
)	O
hopes	O
to	O
build	O
an	O
exascale	B-General_Concept
computer	I-General_Concept
by	O
2021	O
.	O
</s>
<s>
On	O
18	O
March	O
2019	O
,	O
the	O
United	O
States	O
Department	O
of	O
Energy	O
and	O
Intel	O
announced	O
the	O
first	O
exaFLOPS	O
supercomputer	B-Architecture
would	O
be	O
operational	O
at	O
Argonne	O
National	O
Laboratory	O
by	O
late	O
2022	O
.	O
</s>
<s>
The	O
computer	O
,	O
named	O
Aurora	B-Device
is	O
to	O
be	O
delivered	O
to	O
Argonne	O
by	O
Intel	O
and	O
Cray	O
(	O
now	O
Hewlett	O
Packard	O
Enterprise	O
)	O
,	O
and	O
is	O
expected	O
to	O
use	O
Intel	O
Xe	O
GPGPUs	O
alongside	O
a	O
future	O
Xeon	O
Scalable	O
CPU	O
,	O
and	O
cost	O
US$600	O
Million	O
.	O
</s>
<s>
On	O
7	O
May	O
2019	O
,	O
the	O
U.S.	O
Department	O
of	O
Energy	O
announced	O
a	O
contract	O
with	O
Cray	O
(	O
now	O
Hewlett	O
Packard	O
Enterprise	O
)	O
to	O
build	O
the	O
Frontier	B-Device
supercomputer	I-Device
at	O
Oak	O
Ridge	O
National	O
Laboratory	O
.	O
</s>
<s>
Frontier	B-Device
is	O
anticipated	O
to	O
be	O
fully	O
operational	O
in	O
2022	O
and	O
,	O
with	O
a	O
performance	O
of	O
greater	O
than	O
1.5	O
exaFLOPS	O
,	O
should	O
then	O
be	O
the	O
world	O
's	O
most	O
powerful	O
computer	O
.	O
</s>
<s>
On	O
4	O
March	O
2020	O
,	O
the	O
U.S.	O
Department	O
of	O
Energy	O
announced	O
a	O
contract	O
with	O
Hewlett	O
Packard	O
Enterprise	O
and	O
AMD	O
to	O
build	O
the	O
El	O
Capitan	O
supercomputer	B-Architecture
at	O
a	O
cost	O
of	O
US$600	O
million	O
,	O
to	O
be	O
installed	O
at	O
the	O
Lawrence	O
Livermore	O
National	O
Laboratory	O
(	O
LLNL	O
)	O
.	O
</s>
<s>
El	O
Capitan	O
was	O
first	O
announced	O
in	O
August	O
2019	O
,	O
when	O
the	O
DOE	O
and	O
LLNL	O
revealed	O
the	O
purchase	O
of	O
a	O
Shasta	O
supercomputer	B-Architecture
from	O
Cray	O
.	O
</s>
<s>
It	O
will	O
use	O
AMD	O
CPUs	O
and	O
GPUs	O
,	O
with	O
4	O
Radeon	O
Instinct	O
GPUs	O
per	O
EPYC	O
Zen	O
4	O
CPU	O
,	O
to	O
speed	O
up	O
artificial	B-Application
intelligence	I-Application
tasks	O
.	O
</s>
<s>
As	O
of	O
November	O
2021	O
,	O
the	O
United	O
States	O
has	O
three	O
of	O
the	O
five	O
fastest	B-Operating_System
supercomputers	I-Operating_System
in	O
the	O
world	O
.	O
</s>
<s>
In	O
Japan	O
,	O
in	O
2013	O
,	O
the	O
RIKEN	O
Advanced	O
Institute	O
for	O
Computational	O
Science	O
began	O
planning	O
an	O
exascale	B-General_Concept
system	O
for	O
2020	O
,	O
intended	O
to	O
consume	O
less	O
than	O
30	O
megawatts	O
.	O
</s>
<s>
In	O
2014	O
,	O
Fujitsu	O
was	O
awarded	O
a	O
contract	O
by	O
RIKEN	O
to	O
develop	O
a	O
next-generation	O
supercomputer	B-Architecture
to	O
succeed	O
the	O
K	B-Device
computer	I-Device
.	O
</s>
<s>
The	O
successor	O
is	O
called	O
Fugaku	B-Device
,	O
and	O
aims	O
to	O
have	O
a	O
performance	O
of	O
at	O
least	O
1	O
exaFLOPS	O
,	O
and	O
be	O
fully	O
operational	O
in	O
2021	O
.	O
</s>
<s>
In	O
2015	O
,	O
Fujitsu	O
announced	O
at	O
the	O
International	O
Supercomputing	B-Architecture
Conference	O
that	O
this	O
supercomputer	B-Architecture
would	O
use	O
processors	O
implementing	O
the	O
ARMv8	O
architecture	O
with	O
extensions	O
it	O
was	O
co-designing	O
with	O
ARM	O
Limited	O
.	O
</s>
<s>
It	O
was	O
partially	O
put	O
into	O
operation	O
in	O
June	O
2020	O
and	O
achieved	O
1.42	O
exaFLOPS	O
(	O
fp16	O
with	O
fp64	O
precision	O
)	O
in	O
HPL-AI	O
benchmark	O
making	O
it	O
the	O
first	O
ever	O
supercomputer	B-Architecture
that	O
achieved	O
1	O
exaFLOPS	O
.	O
</s>
<s>
Named	O
after	O
Mount	O
Fuji	O
,	O
Japan	O
's	O
tallest	O
peak	O
,	O
Fugaku	B-Device
retained	O
the	O
No	O
.	O
</s>
<s>
1	O
ranking	O
on	O
the	O
Top	B-Operating_System
500	I-Operating_System
supercomputer	B-Architecture
calculation	O
speed	O
ranking	O
announced	O
on	O
November	O
17	O
2020	O
,	O
reaching	O
a	O
calculation	O
speed	O
of	O
442	O
quadrillion	O
calculations	O
per	O
second	O
,	O
or	O
0.442	O
exaFLOPS	O
.	O
</s>
<s>
As	O
of	O
June	O
2022	O
,	O
China	O
had	O
two	O
of	O
the	O
Top	O
Ten	O
fastest	B-Operating_System
supercomputers	I-Operating_System
in	O
the	O
world	O
.	O
</s>
<s>
According	O
to	O
the	O
national	O
plan	O
for	O
the	O
next	O
generation	O
of	O
high	B-Architecture
performance	I-Architecture
computers	I-Architecture
and	O
the	O
head	O
of	O
the	O
school	O
of	O
computing	O
at	O
the	O
National	O
University	O
of	O
Defense	O
Technology	O
(	O
NUDT	O
)	O
,	O
China	O
was	O
supposed	O
to	O
develop	O
an	O
exascale	B-General_Concept
computer	I-General_Concept
during	O
the	O
13th	O
Five-Year-Plan	O
period	O
(	O
2016	O
–	O
2020	O
)	O
which	O
would	O
enter	O
service	O
in	O
the	O
latter	O
half	O
of	O
2020	O
.	O
</s>
<s>
The	O
government	O
of	O
Tianjin	O
Binhai	O
New	O
Area	O
,	O
NUDT	O
and	O
the	O
National	O
Supercomputing	B-Architecture
Center	O
in	O
Tianjin	O
are	O
working	O
on	O
the	O
project	O
.	O
</s>
<s>
After	O
Tianhe-1	B-Device
and	O
Tianhe-2	B-Device
,	O
the	O
exascale	B-General_Concept
successor	O
is	O
planned	O
to	O
be	O
named	O
Tianhe-3	O
.	O
</s>
<s>
In	O
2011	O
,	O
several	O
projects	O
aiming	O
at	O
developing	O
technologies	O
and	O
software	O
for	O
exascale	B-General_Concept
computing	I-General_Concept
were	O
started	O
in	O
the	O
European	O
Union	O
.	O
</s>
<s>
The	O
CRESTA	O
project	O
(	O
Collaborative	O
Research	O
into	O
Exascale	B-General_Concept
Systemware	O
,	O
Tools	O
and	O
Applications	O
)	O
,	O
the	O
DEEP	O
project	O
(	O
Dynamical	O
ExaScale	B-General_Concept
Entry	O
Platform	O
)	O
,	O
and	O
the	O
project	O
Mont-Blanc	O
.	O
</s>
<s>
A	O
major	O
European	O
project	O
based	O
on	O
exascale	B-General_Concept
transition	O
is	O
the	O
MaX	O
(	O
Materials	O
at	O
the	O
Exascale	B-General_Concept
)	O
project	O
.	O
</s>
<s>
The	O
Energy	O
oriented	O
Centre	O
of	O
Excellence	O
(	O
EoCoE	O
)	O
exploits	O
exascale	B-General_Concept
technologies	O
to	O
support	O
carbon-free	O
energy	O
research	O
and	O
applications	O
.	O
</s>
<s>
On	O
28	O
September	O
2018	O
,	O
the	O
European	O
High-Performance	B-Architecture
Computing	I-Architecture
Joint	O
Undertaking	O
(	O
EuroHPC	O
JU	O
)	O
was	O
formally	O
established	O
by	O
the	O
EU	O
.	O
</s>
<s>
The	O
EuroHPC	O
JU	O
aims	O
to	O
build	O
an	O
exascale	B-General_Concept
supercomputer	B-Architecture
by	O
2022/2023	O
.	O
</s>
<s>
In	O
March	O
2023	O
the	O
government	O
of	O
the	O
United	O
Kingdom	O
announced	O
it	O
would	O
invest	O
₤900	O
million	O
in	O
the	O
development	O
of	O
an	O
exascale	B-General_Concept
computer	I-General_Concept
.	O
</s>
<s>
In	O
June	O
2017	O
,	O
Taiwan	O
's	O
National	O
Center	O
for	O
High-Performance	B-Architecture
Computing	I-Architecture
initiated	O
the	O
effort	O
towards	O
designing	O
and	O
building	O
the	O
first	O
Taiwanese	O
exascale	B-General_Concept
supercomputer	B-Architecture
by	O
funding	O
construction	O
of	O
a	O
new	O
intermediary	O
supercomputer	B-Architecture
based	O
on	O
a	O
full	O
technology	O
transfer	O
from	O
Fujitsu	O
corporation	O
of	O
Japan	O
,	O
which	O
is	O
currently	O
building	O
the	O
fastest	O
and	O
most	O
powerful	O
A.I.	B-Application
</s>
<s>
based	O
supercomputer	B-Architecture
in	O
Japan	O
.	O
</s>
<s>
Additionally	O
,	O
numerous	O
other	O
independent	O
efforts	O
have	O
been	O
made	O
in	O
Taiwan	O
with	O
the	O
focus	O
on	O
the	O
rapid	O
development	O
of	O
exascale	B-General_Concept
supercomputing	B-Architecture
technology	O
,	O
such	O
as	O
Foxconn	O
Corporation	O
which	O
recently	O
designed	O
and	O
built	O
the	O
largest	O
and	O
fastest	B-Operating_System
supercomputer	I-Operating_System
in	O
all	O
of	O
Taiwan	O
.	O
</s>
<s>
This	O
new	O
Foxconn	O
supercomputer	B-Architecture
is	O
designed	O
to	O
serve	O
as	O
a	O
stepping	O
stone	O
in	O
research	O
and	O
development	O
towards	O
the	O
design	O
and	O
building	O
of	O
a	O
state	O
of	O
the	O
art	O
exascale	B-General_Concept
supercomputer	B-Architecture
.	O
</s>
<s>
In	O
2012	O
,	O
the	O
Indian	O
Government	O
proposed	O
to	O
commit	O
US	O
$2.5	O
billion	O
to	O
supercomputing	B-Architecture
research	O
during	O
the	O
12th	O
five-year	O
plan	O
period	O
(	O
2012	O
–	O
2017	O
)	O
.	O
</s>
<s>
Additionally	O
,	O
it	O
was	O
later	O
revealed	O
that	O
India	O
plans	O
to	O
develop	O
a	O
supercomputer	B-Architecture
with	O
processing	O
power	O
in	O
the	O
exaFLOPS	O
range	O
.	O
</s>
<s>
These	O
supercomputers	B-Architecture
will	O
use	O
indigenously	O
developed	O
microprocessors	O
by	O
C-DAC	O
in	O
India	O
.	O
</s>
