OPTIMIZE SUMX MEASURE PERFORMANCE PLEASE HELP!

[cdawidow]

Hi guys.  I have a basic calculation to model new sales ( New Proposed Sales) based on implementing a product discount ( Proposed Discount) across my customer base and their respective sales history ( based on average quantity sold).  Because I have to iterate across each customers respective average sales data to accomplish my measure, I used SUMX.  The formula is shown below.  The formual works but it is too clunky and I want to optimize it so the memory use is minimized perhaps using virtual tables but I am running into a brick wall and cannot seem to get the answer apart from using SUMX.  Any help would be appreciated !

[Mohammad_Refaei]

It will help a lot if you share a file with random data.

In general, the if statement in your measure will force the formula engine to operate and will generate a CallBackDataID which impacts the performance.

I don't understand the significance of the Sales[PI], if you can't avoid the condition, try to evaluate it in a variable.

[lbendlin]

Are you familiar with DAX Studio?  Use it to analyze the performance of this measure (and all its upstream measures) and then decide if it needs optimizing or not.