Why Slicer and not Filter

[adik]

I've read all over the internet that IBM recommends the use of Slicers and not Filters when it comes to OLAP/DMR.
My question is why?
The situations where I had to filter the query based on a data item not shown in the report, I managed to handle them perfectly with Filters and the results were as expected. The same results I managed to get via Slicers, but with one major difference: performance.
When using a prompt slicer for example, a report i recently developer ran for 30 mins without showing any results and without generating any SQL. Same report but using Filter ran for 40 seconds and returned the correct results.
So why Slicers?
Btw. I read most of the internet post about this but they did not explain very logically the difference.

[blom0344]

To clear things up, are you using an OLAP  or a DMR model? With DMR a filter would drastically reduce data returned from the database, hence the better performance. But that would only be true if the filter can be added to the SQL generated..

[adik]

This latest report I developed was DMR. But there have been ocasions with TM1 cube for instance where Filter was faster than a Slicer