So I have read about the debate about hyperthreading on a SQL Server or not.
Well, I decided to put it to the test.
I have an ETL process that runs about 8 hours, moving data from a DB2 source, to a staging area on SQL Server.
Then it is moved into an operational data store, and then finally into a data mart.
The system has four 6-core Xeon processors.
I turned off hyperthreading, and the ETL process decreased by more than an hour!
So there it is.. My totally unscientific, unique results!
Your mileage may vary, but at least give it a try on a Xeon processor.
peace
1 comment:
The performance folks actually testing SQL Server at Redmon recently recommended enabling HT for any server that is a Nehalem architecture (or newer). They need to take this path in order for Hyper-V 2012 to be effective and to run SQL Server on Hyper-V.
Post a Comment