Here's the old server info using the deprecated Sql_Blitz commands.  Do you know how much quicker SQL 2017 is supposed to be over SQL 2008?

1. This server might be...

Yes, I ran it with the CheckServerInfo flag and here are all the entries for server info.  In my query, I was able to peg all 4 CPUs @100% on...

• This reply was modified 6 months ago by  MichaelT.

Thank you,  I ran blitz and also installed sp_whoisactive (nifty tool).

Here are the main performance related ones:

• Recovery Interval Not Optimal -The database [xxx] has a target recovery interval...

Thanks Federico, I'll give it a try.

I'm running a query that should peg the CPU but it only uses 25% (1 of the 4 CPUs) on both servers and it...

• This reply was modified 6 months ago by  MichaelT.
• This reply was modified 6 months ago by  MichaelT.

I gave you a series of links for IO testing - and I've also gave you a series of questions to ask your VM admin - have these get...

So we ran Crystal DiskMark and it produced number of 2,000 on the SQL Server 2017 and 1,000 on the SQL Server 2008 R2.  These seemed a bit off for...

We use Ola Hallengren's maintenance...

Out of curiosity, our 2017 box shows a CPU cap of 25% with 4 CPUs which leads me to believe that it's not able to use all 4 CPUs and...

I've run the operations many times while looking at Task Manager CPU, Memory, and Disk Usage.

SQL Server 2017

CPU: 25%

Total Bytes: ~2,000,000

SQL Server 2008 R2

CPU: 13%

Total Bytes: ~6,000,000

We're clearly writing more...

Thank you, I've looked...

Thank for your insightful answer - I would have marked it as an answer but I couldn't mark more than one.

What I do when benchmarking my SQL instances is...

• This reply was modified 6 months, 2 weeks ago by  MichaelT.

Thank you! We'll look into those settings.

Just out of curiosity, if the machines are identical in terms of all specs with more RAM for the SQL 2017 should we be...

Thanks for the reply and the information.

It's VMWare.

Disks should be local SSDs (no encryption).

I checked AV (whew) - I thought that could be it but turning Real-Time monitoring off had...

• This reply was modified 6 months, 3 weeks ago by  MichaelT.

UPDATE

I decided to test performance of individual updates by looping through 100,000 rows and updating each one.  No transactions.

The performance degradation for the 1st run was 250% which matches the...

• This reply was modified 6 months, 3 weeks ago by  MichaelT.