Persistent Realities » Home

Site
News
Files

Visual Basic
Strings
Math
General
Properties
Memory
Methods

Inline ASM-VB
Strings
Math
General
Memory

Single vs Multi dim arrays

Single arrays are much faster.
Probably the more dimensions, the faster the single dimension array is in comparison. If I increase the number of iterations, the difference would probably get bigger as well.

Declarations: Private myArray1(10000) As Long Private myArray2(100, 100) As Long

Code:
Public Sub TestOne()
    Dim lngX As Long, lngY As Long
    For lngY = 0 To 100 - 1
        For lngX = 0 To 100 - 1
            myArray2(lngX, lngY) = 1
        Next lngX
    Next lngY
End Sub

Public Sub TestTwo()
    Dim lngX As Long, lngY As Long, lngCalc As Long
    '//Same as using X,y
    For lngY = 0 To 100 - 1
        lngCalc = (lngY * 100)  '//Normal multi dim width
        For lngX = 0 To 100 - 1
            myArray1((lngX + lngCalc)) = 1
        Next lngX
    Next lngY
End Sub

Single % faster than Multi	Single (sec)	Multi (sec)
29.6%	0.896483	0.691485
28.6%	1.050448	0.817123
15.6%	0.960848	0.831025
14.5%	0.998089	0.871410
47.6%	1.062180	0.719528

User contributed notes:

Author: Tom (hurendo_kun at hotmail dot com)	Date: 16:06 08/06/2005
This is true for older chipsets, but it doesn't seem to hold for newer ones. A similar profile on my system with one-, two-, and three-dimensional arrays yielded comparable results for each. (However, this may be attributed to my dimensions, which were powers-of-two. Using powers-of-two provides better memory alignment and thus yields faster performance.)