none
Power Query Performance (CPU Cores) RRS feed

  • Question

  • I've got an query regarding Power Query / Power BI performance on different hardware. I'm interested to know whether Power Query's performance/refresh times would benefit significantly by adding additional CPU cores? I understand that there are numerous factors influencing performance (including source of data/type of transformations etc), but just inquiring on a general basis. For example, if I needed to combine 1GB of CSV files and apply some transformations and it took, say 10 minutes (on a Core I7 - 2 Core / 16 GB ram PC), would the same query speed up significantly by running on an 18 core / 32GB PC? If not, are there any specific hardware combinations that might yield the maximum performance from PQ, just generally (mainly working with excel and csv type files). 

    Thursday, March 22, 2018 2:08 PM

Answers

  • From the experience, there is no situation when adding memory is bad.

    Memory consumedby mashup processes, each process takes up to 256 Mb of memory. May be there is a limited number of such processes (six?), but I do not sure. You can see a good performance rise moving from 8gb oto 16 gb, also as from 16 to 32. For most cases now 32 is ok and for big files is better, than 16.

    Strongly recommend to use SSD when working with files - it can add a lot of performance gain.

    CPU cores - cannot give any comment. Supposingly more cores - more fast each mashup process works


    Maxim Zelensky Excel Inside

    Thursday, March 22, 2018 2:56 PM