Forum Discussion
Advocate IVFabric Semantic Model Vs SSAS Tabular Cube
I have created fabric semantic model as an alternative to On-Premise tabular SSAS cube. But can see that the Power BI Reports on it are taking too much time to render even when running on F64
Thankyou, nielsvdc, for your response.
Hi Anusha66,
Based on my understanding, slow rendering typically occurs when Fabric semantic models default to Direct Lake and queries fall back to DirectQuery, for example when unsupported features are used or when the data is not cached in memory. In such cases, visuals tend to be slower compared with the in-memory behaviour of SSAS Tabular.Please follow the steps below, which may help improve performance:
- Check for Direct Lake fallback by using Performance Analyzer. If visuals show DirectQuery time, fallback is occurring.
- For the best performance, switch large tables to Import mode if data freshness permits, or adopt a hybrid approach. If you continue to use Direct Lake, ensure the Delta tables are optimised with V-Order or compaction.
- Design the model following best practices: maintain a star schema, remove high-cardinality columns, and apply aggregations.
- Monitor capacity using the Fabric Capacity Metrics app to verify whether the F64 is experiencing CPU or memory pressure.
Additionally, please refer to the links below for further information:
Semantic model modes in the Power BI service - Power BI | Microsoft Learn
Direct Lake overview - Microsoft Fabric | Microsoft Learn
Use Performance Analyzer to Examine Report Performance - Power BI | Microsoft Learn
Understand the metrics app compute page - Microsoft Fabric | Microsoft Learn
Optimization guide for Power BI - Power BI | Microsoft LearnWe hope the information provided helps resolve the issue. Should you have any further queries, please feel free to contact the Microsoft Fabric community.
Thank you.
5 Replies
- nielsvdc
Super User
What are the connection setting for the semantic model? Are you using import mode, direct query or direct lake?
If you want to compare performance between SSAS and a semantic model, you should be using import mode for the tables as data is then loaded into memory of the semantic model, the same as SSAS does.
When you used Power BI Desktop to create your semantic model, the default mode is import mode. When you create the semnatic model through the web browser and initiated it from a Lakehouse or Warehouse, the default mode is direct lake.
Hope this helps. If so, please give a Kudos 👍 and mark as Accepted Solution ✔️.
v-pnaroju-msftCommunity Support
Thankyou, nielsvdc, for your response.
Hi Anusha66,
Based on my understanding, slow rendering typically occurs when Fabric semantic models default to Direct Lake and queries fall back to DirectQuery, for example when unsupported features are used or when the data is not cached in memory. In such cases, visuals tend to be slower compared with the in-memory behaviour of SSAS Tabular.Please follow the steps below, which may help improve performance:
- Check for Direct Lake fallback by using Performance Analyzer. If visuals show DirectQuery time, fallback is occurring.
- For the best performance, switch large tables to Import mode if data freshness permits, or adopt a hybrid approach. If you continue to use Direct Lake, ensure the Delta tables are optimised with V-Order or compaction.
- Design the model following best practices: maintain a star schema, remove high-cardinality columns, and apply aggregations.
- Monitor capacity using the Fabric Capacity Metrics app to verify whether the F64 is experiencing CPU or memory pressure.
Additionally, please refer to the links below for further information:
Semantic model modes in the Power BI service - Power BI | Microsoft Learn
Direct Lake overview - Microsoft Fabric | Microsoft Learn
Use Performance Analyzer to Examine Report Performance - Power BI | Microsoft Learn
Understand the metrics app compute page - Microsoft Fabric | Microsoft Learn
Optimization guide for Power BI - Power BI | Microsoft LearnWe hope the information provided helps resolve the issue. Should you have any further queries, please feel free to contact the Microsoft Fabric community.
Thank you.
- v-pnaroju-msft
Hi Anusha66,
We would like to follow up and see whether the details we shared have resolved your problem. If you need any more assistance, please feel free to connect with the Microsoft Fabric community.
Thank you. - v-pnaroju-msft
Hi Anusha66,
We are following up to see if what we shared solved your issue. If you need more support, please reach out to the Microsoft Fabric community.
Thank you. - v-pnaroju-msft
Hi Anusha66,
We wanted to see if the information we gave helped fix your problem. If you need more help, please feel free to contact the Microsoft Fabric community.
Thank you.
