September 4, 2009 at 4:14 pm
Jan Van der Eecken (9/4/2009)
Anyhow, strange that the sore muscles I had from Wednesday's session seem to have magically (almost) disappeared after another session tonight, and not an easy one for that matter. Coach must have been in the army in his younger days. Is that sudden recovery normal?
Yeah. The best cure for sore muscles is more of what made them sore. Tomorrow, however, may be a different story.
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
September 4, 2009 at 4:45 pm
GilaMonster (9/4/2009)
Jan Van der Eecken (9/4/2009)
Anyhow, strange that the sore muscles I had from Wednesday's session seem to have magically (almost) disappeared after another session tonight, and not an easy one for that matter. Coach must have been in the army in his younger days. Is that sudden recovery normal?Yeah. The best cure for sore muscles is more of what made them sore. Tomorrow, however, may be a different story.
Sounds very comforting, thanks. Maybe I should just plan on staying in bed tomorrow.
September 4, 2009 at 4:59 pm
Jan Van der Eecken (9/4/2009)
GilaMonster (9/4/2009)
Jan Van der Eecken (9/4/2009)
Anyhow, strange that the sore muscles I had from Wednesday's session seem to have magically (almost) disappeared after another session tonight, and not an easy one for that matter. Coach must have been in the army in his younger days. Is that sudden recovery normal?Yeah. The best cure for sore muscles is more of what made them sore. Tomorrow, however, may be a different story.
Sounds very comforting, thanks. Maybe I should just plan on staying in bed tomorrow.
Nope, definitely get up and about. Staying in bed will REALLY make you stiff and sore!
Jack Corbett
Consultant - Straight Path Solutions
Check out these links on how to get faster and more accurate answers:
Forum Etiquette: How to post data/code on a forum to get the best help
Need an Answer? Actually, No ... You Need a Question
September 4, 2009 at 5:06 pm
Was kidding, Jack. No way my family would allow me a day off in the first place.
September 4, 2009 at 9:44 pm
Going way back to the very first post on this thread, the answer is a resounding [font="Arial Black"]YES![/font]
http://www.sqlservercentral.com/Forums/Topic783201-146-1.aspx?
--Jeff Moden
Change is inevitable... Change for the better is not.
September 5, 2009 at 1:19 am
Jeff Moden (9/4/2009)
Going way back to the very first post on this thread, the answer is a resounding [font="Arial Black"]YES![/font]http://www.sqlservercentral.com/Forums/Topic783201-146-1.aspx?
That is impressive. I'd like to think he hit 'submit' too soon, but I'm probably deluding myself there.
Awesome question. Points to you for replying sans-pork.
Paul White
SQLPerformance.com
SQLkiwi blog
@SQL_Kiwi
September 5, 2009 at 9:21 am
Just read in Grant's book, Chapter 8, that:
To limit contention with the operating system, I will usually set the max degree of parallelism setting to one less than the number on the server.
I'm wondering if this is sockets or cores, since SQL Server sees each core as a separate processor. Any insights?
Jack Corbett
Consultant - Straight Path Solutions
Check out these links on how to get faster and more accurate answers:
Forum Etiquette: How to post data/code on a forum to get the best help
Need an Answer? Actually, No ... You Need a Question
September 5, 2009 at 10:01 am
Jack Corbett (9/5/2009)
Just read in Grant's book, Chapter 8, that:To limit contention with the operating system, I will usually set the max degree of parallelism setting to one less than the number on the server.
I'm wondering if this is sockets or cores, since SQL Server sees each core as a separate processor. Any insights?
That seems an odd recommendation - if you really want to leave a whole core for the OS you would use the cpu and i/o affinity masks. You might sensibly use the affinity and affinity64 masks if you have lots of cores, or are affinitizing one to network processing, for example.
Setting server MAXDOP only limits the runnable threads per executing request (it is perfectly possible for a request to have more parallel threads than server MAXDOP, or cores for that matter). If more than one session is running a parallel query, chances are that all schedulers will have runnable workers - and in general this means that all cores will be involved too, regardless of server MAXDOP.
Of course Windows is normally free (if the affinity masks are zero) to move SQLOS threads around - so two or more schedulers could end up on the same core if needs be. I should mention that if the cpu affinty mask on an 8 core machine is set to 0xFF then each scheduler is fixed to a core and Windows won't move them around (this is not usually desirable).
So, for lots of reasons, it seems an odd practice - but no worse than many I guess.
I tend to set server maxdop to 1 - partly because most queries don't benefit, partly because the optimizer isn't brilliant at deciding when to use parallelism, but mostly because I prefer to hand-tune queries that do benefit from parallelism and add an OPTION (MAXDOP 4) to override the server setting. I use '4' for NUMA-related reasons, and also because the incremental benefit from parallelism tails off dramatically after about that number of processors. I reserve the right to break my own rules on occasion of course.
Paul
edit: just to add, all bets are off on a powerful server dedicated to data warehousing or decision support. Most queries then benefit from unrestricted parallelism (up to say 16 cpus/cores).
Paul White
SQLPerformance.com
SQLkiwi blog
@SQL_Kiwi
September 6, 2009 at 7:03 pm
Sigh.... my Rebels beat Tiger High, but it wasn't pretty.
Hope you weren't emotionally invested in the game, Alvin. 😉
__________________________________________________
Against stupidity the gods themselves contend in vain. -- Friedrich Schiller
Stop, children, what's that sound? Everybody look what's going down. -- Stephen Stills
September 7, 2009 at 7:38 am
I just got long-winded and gave a lecture here about how execution plans differ depending on the size and nature of the data. If anybody would care to correct any egregious errors or omissions, I would appreciate it.
__________________________________________________
Against stupidity the gods themselves contend in vain. -- Friedrich Schiller
Stop, children, what's that sound? Everybody look what's going down. -- Stephen Stills
September 7, 2009 at 9:10 pm
Question: In SSC, is the proper acronym "LOL" (laughing out loud) or "SOM" (spewing on monitor) ??
__________________________________________________
Against stupidity the gods themselves contend in vain. -- Friedrich Schiller
Stop, children, what's that sound? Everybody look what's going down. -- Stephen Stills
September 7, 2009 at 9:52 pm
Posted my second-ever question here
Anyone have any ideas?
BTW - definitely SOM not LOL 🙂
Paul White
SQLPerformance.com
SQLkiwi blog
@SQL_Kiwi
September 8, 2009 at 12:53 am
Nope no idea, i Prefer ROFLMAO, which is of course. rolling on floor laughing my arse off.
--------------------------------------------------------------------------------------
[highlight]Recommended Articles on How to help us help you and[/highlight]
[highlight]solve commonly asked questions[/highlight]
Forum Etiquette: How to post data/code on a forum to get the best help by Jeff Moden[/url]
Managing Transaction Logs by Gail Shaw[/url]
How to post Performance problems by Gail Shaw[/url]
Help, my database is corrupt. Now what? by Gail Shaw[/url]
September 8, 2009 at 3:41 am
Only in South Africa - http://pigeonrace2009.co.za/pigeon-race-2009
My money's on the pigeon.
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
September 8, 2009 at 3:54 am
GilaMonster (9/8/2009)
Only in South Africa - http://pigeonrace2009.co.za/pigeon-race-2009My money's on the pigeon.
only if it has go-faster stripes on its wings 😀
--------------------------------------------------------------------------------------
[highlight]Recommended Articles on How to help us help you and[/highlight]
[highlight]solve commonly asked questions[/highlight]
Forum Etiquette: How to post data/code on a forum to get the best help by Jeff Moden[/url]
Managing Transaction Logs by Gail Shaw[/url]
How to post Performance problems by Gail Shaw[/url]
Help, my database is corrupt. Now what? by Gail Shaw[/url]
Viewing 15 posts - 7,876 through 7,890 (of 66,712 total)
You must be logged in to reply to this topic. Login to reply