This article lists answers to frequently asked questions about query acceleration in Fabric Data Warehouse.
Does query acceleration require application changes?
No. Query acceleration operates within Fabric Data Warehouse and doesn't require application changes, query rewrites, schema changes, reformatting, data movement, or additional infrastructure to manage.
Does it accelerate every query?
No. Only eligible operations and workloads are accelerated. Some parts of a query might qualify for query acceleration, and some might not. However, when query acceleration is active, all queries are billed under a query acceleration meter. For more information, see Billing and utilization reporting in Fabric Data Warehouse.
How is query acceleration for Fabric Data Warehouse billed?
Query acceleration uses the same business model as the rest of Fabric and is billed using Capacity Units (CU). Query acceleration uses a separate meter. This meter consumes CUs at a higher rate than standard Fabric Data Warehouse workloads. Once enabled, all queries are billed through the Query acceleration meter.
How does query acceleration differ from traditional scaling?
Traditional approaches to improving analytics performance often involve strategies to increase compute resources, add more CPU cores, rewrite queries, or create additional aggregates or optimization structures.
Query acceleration introduces another dimension of optimization by using specialized hardware to process eligible operations more efficiently. Rather than replacing CPUs, accelerators complement them.
Does query acceleration work with Power BI?
Yes. Power BI scenarios that generate SQL (Direct Query) against Fabric Data Warehouse can benefit when eligible queries are accelerated.
Does query acceleration improve the performance of DAX?
Query acceleration accelerates eligible SQL execution within Fabric Data Warehouse. DAX itself is not accelerated directly. Benefits may be observed when DAX queries generate eligible warehouse SQL operations.