Is it slower on server or when you have it open in Tableau Desktop? The version where you included the extract would likely be faster on your desktop because the data is local (on your hard drive/RAM) where as if you are using a published data source, Tableau desktop is querying server for data over the network, so it depends on your network speeds.
If it is less performant access the view on server, then I am not sure what is going on, unless the published data source and included data source are not identical for some reason.
Is your Tableau server instance a single node or multi-node install?
hope this helps
Thank you for the fast reply, but the situation is the following.
Both reports are absolutely the same and are published on server.
The only difference is that one is using as datasource extract published together with the report and the other is using same datasource, but published as datasource on tableau server and loading of a dashboard is approximately three times slower on second report.
I am using latest version of tableau server and origin of data is MS SQL 2012.
Here's how to make a performance recording on Server to inspect where the difference is happening:
It's possible to publish a data source without an extract and that would be expected to be slower. Here's some more info on publishing data sources:
If you have a lot of quick-filters on this dashboard, that could be the difference. If you publish with an embedded data source, Tableau optimizes for the quick filters you have displayed, but doesn't do that if you use a published data source. Tamas has some excellent detail around this: Tableau Performance: Why Embedded Data Source is Faster than Published -Databoss
Adding to Tony's . Check out if you have nested if conditions in the formula's try to reduce the complexity and perform then in SQL Server . Reduce the color coding ,sheets ,Old data ...avoid too many filters .