Averaging Filter SDLT Sample

Hi all

I am not able to understand the output we receive from the Averaging filer

with SDLT , the output is:

The time taken in number of ticks is 774874281
The time taken just for filter operation is 55768898
 

And with Serial Sample, output is much better:

The time taken in number of ticks is 766142320
The time taken just for filter operation is 51793809

It is clear from the response that Serial Sample is performing better, so why SDLT?

Can you direct me to correct and detailed doc or explain me, why someone should prefer Intel SDLT? Even i do not see any difference in output image.

Thank you