The problem was... Extended Events (one active session), which logged queries on one - hardly used DB. Incomprehensible...

Switching it off caused above calculation time decline from 40s to less than...

There is something wrong with that instance. HammerDB results confirm the problem. On the same VM I have installed second instance of SQL and guess! No CPU performance issue and...

Look at this, please

https://itpeernetwork.intel.com/how-to-maximise-cpu-performance-for-sql-server-on-windows/#gs.v3qqw4

"maybe you misunderstood my comment - I meant that I saw that you used a loop to test, is this because there are lots of loops in your real code...

Jonathan

I have not tested it separetly yet

Mike

"(...) SQL needs disk as its main resource (...)"

I know that generally but this is not a only one test. I have started collect...

My steps:

1. Created table (TestTable) on Server2 (this "good" one) with 1000 rows: Id (bigint), DateTimeUTC (datetime) populated with data
2. Copied this table to Server1 (same data on both)
3. No...

• This reply was modified 5 years, 11 months ago by everand76.

At this moment I can say, that statistics time shows huge CPU time processing in section 2 - more comments - biger that time. It looks like comented code is...

I know - I sound like I'm crazy, but it's real life. Execution plans are identical- with comments and without. You have right - comments should be ignored, but on...

I don't catch your question (my english) but I mean statements like this:

/*1*/

select t1.DateUTC into #Tmp1 from Table1 t1

/*2*/

select function(t2.DateUTC) DateLocal from #Tmp1 t2

/*3*/

/*

select

t3.Field1

,t3.Field2

,t3.Field4

...

from TableX t3

join TableY t4 on ...

......

*/

If...