What may increase the latency 3-4 times for SQL FCI the database?

We have SQL Server Basic Availability Groups (BAG) configured to run an SQL database running over local Intel SSD drive. I was asked to move a database to SQL Server Failover Cluster Instance (FCI) to increase the SQL Server performance: run the database on top of HA virtual drive powered by software-defined storage (DataCore, StorMagic, HP StoreVirtual VSA). From my experience hyperconverged VSAN should double read operations thus the SQL IO latency (for reads from the database) should decrease twice.

So two scenarios were benchmarked: SQL BAG and SQL FCI. For those two cases, the maximum server memory of 512 GB RAM was set to SQL Server to exclude caching and perform fair read operations from the database table.

Management Studio SQL and SQLQueryStress were used for testing purposes. The SQL statement is `SELECT TOP (500000) ... FROM...