Can Power BI Desktop speed up on more CPU Core PC?

[yoshihirok]

I want to know the speed up on more CPU Core PC.

There are a lot of PC which have many CPU Core.

Example,

Intel Core i9-7900X have 10 cores and 20 threads.

Intel Core i7-8700 have 6 cores and 12 threads

Now, I am using Power BI Desktop on Surface Pro 4 with Intel Core i5-6300U which have 2 cores and 4 threads

So, Can Power BI Desktop speed up on more CPU cores PC?

Regards,

Yoshihiro Kawabata

[je777999]

I am searching for an answer to this question as well, please see the answer below 😃

Power BI Desktop – Does More CPU Help?

YES, CPU cores & threads help in these scenarios:

1. Data refresh (Power Query):

   • Power Query (M engine) can parallelize data loading and transformations across multiple threads.

   • More cores/threads mean faster data load, especially when dealing with multiple data sources or complex transformations.

2. Model processing (DAX engine):

   • VertiPaq (the in-memory engine) uses parallel processing when loading, compressing, and indexing data.

   • Large datasets or complex models benefit from more threads during refresh.

3. Visual rendering:

   • Some visuals and DAX measures are processed in parallel, so more threads can help here too.

BUT, there are caveats:

• DAX queries themselves can only benefit so much — some operations are single-threaded or have low parallelism.

• There's diminishing returns beyond 6-8 cores for typical desktop use.

• RAM and disk I/O (SSD vs HDD) often become the bottleneck before CPU does.

---

🧾 Power BI Report Builder (Paginated Reports)

Paginated reports are very different:

• They're SQL-driven, often calling stored procedures or direct queries.

• CPU helps during rendering (e.g., PDF export, complex layout generation).

• But the real bottlenecks are:

  • Query execution time (SQL Server performance, not your machine)

  • Report complexity

  • Network latency (if remote data sources)

So, more cores may help locally when rendering complex paginated reports, but most of the heavy lifting is server-side if you're hitting a database.

Hope this helps! Please like if this answer solved your issue/question too 😃

[kpmaster2k]

Hi , the most important thing when loading information is the latency of the ram. I have a server with 48 threads, but it win by loading 2 million records a core i7 6700, because the memory has a lower latency, when making large aggregations can influence, but at the time of loading wins the latency of the memory in power bi desktop.

[Anonymous]

I doubt that RAM latency would have any significant effect on Power BI processing.   It is more likely due to CPU frequency or maybe network.   The 6700 runs at 4Ghz whereas high core servers tend to run at lower speeds (typically around 2 to 2.5 Ghz so nearly half) to keep the heat down.

Power BI does indeed make some use of multiple processors.   You can see this in Task Manager as with only one report open you will see multiple CefSharp.BrowserSubprocess and Microsoft Mashup Evaluation Containers running.   I have not tested this thoroughly but doubt it is fully processing in parallel but rather assigning different processes to different tasks (e.g. one per visualisation or query perhaps - others for miscellaneous tasks that may be mostly idle).   If I am right in that assumption then the benefits of having multiple cores would depend on how complex your report is.   So, for example, a single query/single visualisation report may not see much benefit say above 2 cores.

[v-yulgu-msft]

Hi @yoshihirok,

Actually, to run Power BI desktop, we should meet the minimum hardware requirements. If you install Power BI desktop on a machine with a more powered CPU, it might speed up when running desktop and executing a report.

Regards,
Yuliana Gu

[yoshihirok]

Hi @v-yulgu-msft

Thank you for your reply.

and I want to know how many CPU core used by Power BI Desktop' some features.

Features, like Query, DAX, R scripts, visuals.

Scenario:
  Importing from over 100,000,000 lines CSV file to Power BI Desktop data model at Query.

  The machine have 16 memory, SSD storage, and currently over 20 minutes for import.

  I hope to reduce the importing time.

Regards,

Yoshihiro Kawabata

[gtrani]

Hi @yoshihirok 

Did you find out a solution for your question?
I have the same question now.

Tks!