DAX Query generation against MDX cube resulting in terrible performance

[mtassin]

We've got an MDX Cube with about 10 billion rows, partitioned into 1 month blocks of about 20 million rows.

What we've noticed though is a query generated via Excel vastly outperforms the same desired results generated out of Power BI Desktop (Report server edition).  The only thing we can surmise is that the DAX query is just less efficient than the MDX query for the same result set.

Seeing as Excel has a fairly decent (not perfect) MDX query generator, why doesn't Power BI borrow that for MDX cubes instead of the translation of DAX to MDX that's apparnetly murdering our performance?

Is there a way to get Power BI to generate MDX queries instead?

[Anonymous]

Can you explain how can you use 20billion rows? We struggle when we try to import even less than 1million from BW. What kind of database source are you using?

I would really appreciate it to understand if we are doing the wrong approach.

Thanks!