Forum Discussion
Reset Performance analyzer- For consistent duration
- 2 months ago
The reason the timings vary is the engine cache. The first run after opening the file is a cold cache, and every run after that is warm, so the numbers drift between attempts. Power BI Desktop's Performance Analyzer does not have a built-in clear cache button.
The cleanest workflow is to open DAX Studio, connect to your PBIX, and run the "Clear Cache" command before each measurement. Then switch back to Power BI Desktop and click "Refresh visuals" in the Performance Analyzer pane. That gives you a consistent cold cache baseline every time. If you prefer not to use DAX Studio, closing and reopening the PBIX between runs has the same effect, just slower.
I also recommend running each test 3 to 5 times and using the median rather than a single number, since query timing always has some natural variance even with a cleared cache.
If this helped, a thumbs up and accepting the solution would be appreciated.
Best regards,
Shai Karmani - 2 months ago
Hello RDS2026,
You usually will not see the exact same duration every time.
Things like cache, background activity, available system resources, and DirectQuery/network delays can all cause the timings to vary a bit.
What I normally do is:
Restart Power BI Desktop first
Ignore the first run
Run the same action a few times
Compare the average instead of one result
That usually gives a better idea whether the report changes actually improved performance.
Docs:
Use Performance Analyzer to examine report performance - 2 months ago
Use DAX Studio instead of Performance Analyzer. Connect DAX Studio to your model, run the same query multiple times using Server Timings, and clear the cache between runs (right-click the model in DAX Studio → Clear Cache). DAX Studio gives you precise Storage Engine and Formula Engine timings that are more consistent than Performance Analyzer.
Hey RDS2026,
Agree with Shai_Karmani , Olufemi7 , cengizhanarslan - the root cause is cache behavior. Cold vs warm cache will always give different numbers. No amount of refreshing fixes this without clearing cache first.
Just to summarize the flow: DAX Studio > Clear Cache > back to Performance Analyzer > Refresh Visuals. Repeat for every test run, and compare medians not individual numbers. Link to Microsoft's own guidance on Performance Analyzer: Use Performance Analyzer to Examine Report Performance - Power BI | Microsoft Learn
Though, it doesn't offer a built-in cache reset, which is exactly why DAX Studio is necessary here.
Hope this helps!
Best,
Harshit