jagged arrayz

If you care about speed then you always prefer jagged arrays over multi-dimensional arrays. The .NET runtime support (jitter and CLR) makes a strong distinction between vectors and arrays. A vector is a special case of an array, one dimension with the starting index at 0. It is very heavily optimized inside the CLR, pulling every possible trick. And the jitter, giving strong odds for array bounds checking elimination.

A jagged array is a vector of vectors, you get all of the runtime optimizations when you use them. Multi-dimensional arrays don't. The speed difference is about a factor of two.

Only ever consider using multi-dimensional arrays if you favor the syntax convenience and have verified that array indexing is not on your program's critical path.

Here is a simple example:

Let's say for some reason you need to factorise integers. For small integers it might make sense to create a lookup table in order to speed things up. That would be a jagged array.

int[][] factors = new int[][]

{

null, // 0 doesn't have prime factors

new int[] { }, // 1 is the trivial case

new int[] { 2 },

new int[] { 3 },

new int[] { 2, 2 },

new int[] { 5 },

new int[] { 2, 3},

new int[] { 7 },

new int[] { 2, 2, 2 },

new int[] { 3, 3 },

new int[] { 2, 5 },

// ... and so on

};

int[] PrimeFactorsOf(int value)

{

if (value < factors.Length)

{

return factors[value];

}

else

{

// do the work

}

5 Replies

Know the answer? Post it — somebody with the same question will find it here.

Ankur Jain12y ago
Accepted answer
If you care about speed then you always prefer jagged arrays over multi-dimensional arrays. The .NET runtime support (jitter and CLR) makes a strong distinction between vectors and arrays. A vector is a special case of an array, one dimension with the starting index at 0. It is very heavily optimized inside the CLR, pulling every possible trick. And the jitter, giving strong odds for array bounds checking elimination.

A jagged array is a vector of vectors, you get all of the runtime optimizations when you use them. Multi-dimensional arrays don't. The speed difference is about a factor of two.

Only ever consider using multi-dimensional arrays if you favor the syntax convenience and have verified that array indexing is not on your program's critical path.

Here is a simple example:

Let's say for some reason you need to factorise integers. For small integers it might make sense to create a lookup table in order to speed things up. That would be a jagged array.

int[][] factors = new int[][]
{
null, // 0 doesn't have prime factors
new int[] { }, // 1 is the trivial case
new int[] { 2 },
new int[] { 3 },
new int[] { 2, 2 },
new int[] { 5 },
new int[] { 2, 3},
new int[] { 7 },
new int[] { 2, 2, 2 },
new int[] { 3, 3 },
new int[] { 2, 5 },
// ... and so on
};

int[] PrimeFactorsOf(int value)
{
if (value < factors.Length)
{
return factors[value];
}
else
{
// do the work
}
}
Ankur Jain12y ago
thanx...please mark it as answer if its helpful...
Najna Abdulla12y ago
oky i got idea :)
Najna Abdulla12y ago
but in real senario where it is used in applications?
Ankur Jain12y ago
Hi!

please go through following link..
http://www.dotnetperls.com/jagged-array