Multi Dimensional Array speed optimization

[JMulligan]

I'm coding for an application that has a function like this:

decimal totalValue;
for(int i=0; i<xArrayLength; i++)
{
   for(int j=0; j<yArrayLength;j++)
   {
       totalValue += currentValue[i,j] //currentValue is modular
   }
}
return totalValue;

xArrayLength and yArrayLength aren't large, but the function can get called millions of times. I've noticed a significant amount of time in my program is spent fetching values from two dimensional arrays like this.

My question: Does swapping the looping order of i and j change the speed of fetching values from the array? Would it affect the speed if the arraylengths were bigger? Is there anything else I can do to speed up the constant access to multi dimensional arrays in scenarios like this?