How much memory is too much memory for a Power BI Development environment?

[T2]

@Greg_Deckler provided useful advice a few years ago on minimum specs here: https://community.powerbi.com/t5/Desktop/computer-specification/m-p/173966/highlight/true#M75942

Our Power BI team expects to continue working from home for quite some time so we are considering standing up larger VMs on-premises, which in addition to being closer to our data, can be beefier than our current laptops, which are 16GB, Intel Core i5-8350U (1.70GHz, 1896 Mhz, 4 Cores, 8 Logical Processors), HPs running Windows 10 Enterprise.

Assuming we provided the equivalent nominal CPU allocation for the three of us, how much memory is too much memory for an individual user? Will we never hit 32GB memory due to limitations in Power BI itself?

We'd appreciate the community's thoughts on this.

Thanks!

[amitchandak]

@T2 , while performance will depend on the lot may factor.  You should refer some of the best practices to optimize the performance

https://maqsoftware.com/expertise/powerbi/power-bi-best-practices
https://docs.microsoft.com/en-us/power-bi/guidance/power-bi-optimization

https://www.thebiccountant.com/2016/11/08/speed-powerbi-power-query-design-process/
https://www.c-sharpcorner.com/article/power-bi-best-practices-part-1/
https://www.knowledgehut.com/blog/business-intelligence-and-visualization/power-bi-best-practices
https://community.powerbi.com/t5/MBAS-Gallery/Microsoft-Power-BI-The-Do-s-and-Don-ts-of-Power-BI-Rel...
https://community.powerbi.com/t5/MBAS-Gallery/Microsoft-Power-BI-My-Power-BI-report-is-slow-What-sho...

You can allocate virtual memory to windows machines to accommodate big models.

Check is power Bi is the largest consumer of RAM, sometimes browsers are doing so.

[Mariusz]

Hi @T2 

There is no such thing as too much memory for Power BI, if you build huge report with a lot of pages and a lot of visuals and/or big data models then you can quite quickly run out of memory, especially if you have to compare new vs old version of the file.

You can start at 32 and get more allocation later if not enough if its VM it should be easy to expand.

Best Regards,
Mariusz

If this post helps, then please consider Accepting it as the solution.

Please feel free to connect with me.
LinkedIn

[parry2k]

@T2 If you specifically talks about memory, it depends on your dataset size, # of columns with unique value etc. You can check all these details in vertipaq analyzer by connecting your model from DAX studio about the column unique values, which table/column is consuming maximum memory etc  

[T2]

Thanks for the quick reply. As we worked with increasingly larger data sets we became aware of such tools.

However, my question is not about how much memory we are consuming.

My question is, more simply, at what point will having more free memory available NOT make a difference to the Power BI Desktop client. Our machines have 16GB of memory. If we added another 16GB-which I don't know is possible on my particular model-would Power BI be able to take advantage of it all? How about 64GB? I had a 64GB Dell Precision machine before this one but I never got close to pegging it with Power BI.  At the same time, I don't know if Power BI could take advantage of another 32GB of free memory.

That is the understanding we seek.

Thanks, again!