<s>
Burstsort	B-Algorithm
and	O
its	O
variants	O
are	O
cache-efficient	O
algorithms	O
for	O
sorting	B-Algorithm
strings	O
.	O
</s>
<s>
They	O
are	O
variants	O
of	O
the	O
traditional	O
radix	B-Algorithm
sort	I-Algorithm
but	O
faster	O
for	O
large	O
data	B-General_Concept
sets	I-General_Concept
of	O
common	O
strings	O
,	O
first	O
published	O
in	O
2003	O
,	O
with	O
some	O
optimizing	O
versions	O
published	O
in	O
later	O
years	O
.	O
</s>
<s>
Burstsort	B-Algorithm
algorithms	O
use	O
a	O
trie	B-General_Concept
to	O
store	O
prefixes	O
of	O
strings	O
,	O
with	O
growable	B-Data_Structure
arrays	I-Data_Structure
of	O
pointers	O
as	O
end	O
nodes	O
containing	O
sorted	O
,	O
unique	O
,	O
suffixes	O
(	O
referred	O
to	O
as	O
buckets	O
)	O
.	O
</s>
<s>
As	O
the	O
buckets	O
grow	O
beyond	O
a	O
predetermined	O
threshold	O
,	O
the	O
buckets	O
are	O
"	O
burst	O
"	O
into	O
tries	B-General_Concept
,	O
giving	O
the	O
sort	O
its	O
name	O
.	O
</s>
<s>
By	O
dividing	O
the	O
input	O
into	O
buckets	O
with	O
common	O
prefixes	O
,	O
the	O
sorting	B-Algorithm
can	O
be	O
done	O
in	O
a	O
cache-efficient	O
manner	O
.	O
</s>
<s>
Burstsort	B-Algorithm
was	O
introduced	O
as	O
a	O
sort	O
that	O
is	O
similar	O
to	O
MSD	O
radix	B-Algorithm
sort	I-Algorithm
,	O
but	O
is	O
faster	O
due	O
to	O
being	O
aware	O
of	O
caching	O
and	O
related	O
radixes	O
being	O
stored	O
closer	O
to	O
each	O
other	O
due	O
to	O
specifics	O
of	O
trie	B-General_Concept
structure	O
.	O
</s>
<s>
And	O
although	O
asymptotically	O
it	O
is	O
the	O
same	O
as	O
radix	B-Algorithm
sort	I-Algorithm
,	O
with	O
time	O
complexity	O
of	O
(	O
w	O
–	O
word	O
length	O
and	O
n	O
–	O
number	O
of	O
strings	O
to	O
be	O
sorted	O
)	O
,	O
but	O
due	O
to	O
better	O
memory	O
distribution	O
it	O
tends	O
to	O
be	O
twice	O
as	O
fast	O
on	O
big	O
data	B-General_Concept
sets	I-General_Concept
of	O
strings	O
.	O
</s>
