February 3, 2017 at 4:21 am
Hi Kerri,
In my understanding setting this option to "auto" is not the same as manually checking every cpu for the affinity.
In the auto modus you might get context switches between the different cpus which can slow you down under very heavy cpu loads as the cpu caches will have to reload the data every time over and over again.
When you check every cpu manual, you will have the same number of available processors but every SQL sheduler bound to a certain cpu and no context switching should happen.
So it would not classify as a bug but rather as by design as you eliminate context swithing by doing so.
Would not improve the throughput on any workload though, but you having large recordsets could explain why data cache cleaning could have you slowed down in the auto modus.
Wkr,
Eddy
February 3, 2017 at 8:09 am
Please note: 5 year old thread, and the person you're replying to hasn't logged into this site in 2 years.
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
February 3, 2017 at 8:26 am
I have been watching the thread and waiting for an answer - thanks for finally posting it. You can view the posts without having to login.
Viewing 3 posts - 16 through 17 (of 17 total)
You must be logged in to reply to this topic. Login to reply