Windows 2008's Magic Performance Button

Windows 2008\'s Magic Performance Button

You really need to change this setting if you are on Windows 2008:


http://jeffstevenson.karamazovgroup.com/2011/06/windows-2008s-magic-performance-button.html

Re: Windows 2008\'s Magic Performance Button

That is a cool post. What kind of observable improvement did you see?

Re: Windows 2008\'s Magic Performance Button

[ QUOTE ]
That is a cool post. What kind of observable improvement did you see?

[/ QUOTE ]

As the articles says: about 70% on a couple of UBE's by only changing the SQL Server. I'm sure more could be gained by changing the enterprise server also.

I wish I had more time to fully test and document but was unable to do so.

Perhaps someone else could.

Re: Windows 2008\'s Magic Performance Button

Here are some metrics for R03B707 A/R Integrity

Prior to any config change: 2 hours 4 min
After config change to DB server: 1 hour 31 min
After config change to DB and Ent server: 54 minutes

Craig

Re: Windows 2008\'s Magic Performance Button

[ QUOTE ]
Here are some metrics for R03B707 A/R Integrity

Prior to any config change: 2 hours 4 min
After config change to DB server: 1 hour 31 min
After config change to DB and Ent server: 54 minutes

Craig

[/ QUOTE ]

Craig,

Did I get our numbers wrong for UBE's?

Re: Windows 2008\'s Magic Performance Button

The numbers I posted were for a different UBE where I had the metrics after the Enterprise Server change as well, so I'm sorry for any confusion.

The one you mentioned was a "simple" UBE that was based over several AB tables with no ER logic. That one increased from 10 minutes to 3 minutes. I don't have the numbers after the ES change but I don't expect much since the UBE logic is almost nothing. Other more complex UBEs did incur the performance gain from the ES change as would be expected.

The point is, it makes a huge impact. Precisely quantifying will be different from UBE to UBE. I did get an email from a delighted analyst who said one UBE went from 1 hours 12 minutes to 4 minutes 32 seconds. So that's even better than what your blog says!

Also, I'm authoring another post about a performance issue with Address Book Privacy. As you know, that came into play with our troubleshooting sessions.

Keep up the great work,
Craig

Re: Windows 2008\'s Magic Performance Button

[ QUOTE ]
The numbers I posted were for a different UBE where I had the metrics after the Enterprise Server change as well, so I'm sorry for any confusion.

The one you mentioned was a "simple" UBE that was based over several AB tables with no ER logic. That one increased from 10 minutes to 3 minutes. I don't have the numbers after the ES change but I don't expect much since the UBE logic is almost nothing. Other more complex UBEs did incur the performance gain from the ES change as would be expected.

The point is, it makes a huge impact. Precisely quantifying will be different from UBE to UBE. I did get an email from a delighted analyst who said one UBE went from 1 hours 12 minutes to 4 minutes 32 seconds. So that's even better than what your blog says!

Also, I'm authoring another post about a performance issue with Address Book Privacy. As you know, that came into play with our troubleshooting sessions.

Keep up the great work,
Craig

[/ QUOTE ]

Thanks for the clarification Craig. Looking forward to the Address Book Privacy write up.