| Author |
Topic |
|
SamC
White Water Yakist
3467 Posts |
 Posted - 2004-01-09 : 08:47:01
|
| I noticed SQL supports PINning a table in memory. I imagine there's not a lot of this going on for a couple of reasones - SQL does a pretty good job managing caching itself, and it would be difficult to identify which tables might benefit overall performance if PINned.I'd be interested in hearing a success story involving PINing, and what led to the idea of PINing the table??Sam |
|
|
X002548
Not Just a Number
15586 Posts |
Posted - 2004-01-09 : 09:28:38
|
| Sam,Just did a search with DBCC PIN, and use robvolk as the memeber...There have been lots of discussions...I think there's even an article about it...My take away is that you need to be careful, because ethe pinned tables reside in memroy (or buffer cache?) and that the amount of data in a pinned table should not be too large...I just stay away from them...I mean if you scan a code table...it's going in to memory anyway...and if you don't use identity to have to join back to the code table, you don't have to worry about it anyway...ok...now for a bigMOIO (My own ignorant opinion)Brett8-) |
 |
|
|
SamC
White Water Yakist
3467 Posts |
Posted - 2004-01-09 : 10:07:53
|
| (1) I should search before I post.(2) See step (1) |
|
|
|
MichaelP
Jedi Yak
2489 Posts |
Posted - 2004-01-09 : 12:28:56
|
| My Sig says it all....Michael<Yoda>Use the Search page you must. Find the answer you will.</Yoda> |
|
|
|
tkizer
Almighty SQL Goddess
38200 Posts |
Posted - 2004-01-09 : 12:32:29
|
| In Yukon, DBCC PINTABLE and DBCC UNPINTABLE will do nothing. The commands will still work, but it won't doing anything. My take on it is that MS doesn't recommend its use so they are getting rid of it but they are allowing the commands to run for backward compatibility. Here's their quote:"These two statements have been made no-ops in SQL Server Yukon"Maybe I misunderstood their comment. Anyone else have a different take on it?Tara |
|
|
|
ehorn
Master Smack Fu Yak Hacker
1632 Posts |
Posted - 2004-01-09 : 13:20:12
|
I had never heard that term used before. Websters The Free On-line Dictionary of Computing defines no-op as:quote:
1. A machine instruction that does nothing (sometimes used in assembler-level programming as filler for data or patch areas, or to overwrite code to be removed in binaries).
I would most definately agree with your assessment.EDIT: Fixed Source |
|
|
|
SamC
White Water Yakist
3467 Posts |
Posted - 2004-01-09 : 13:40:17
|
| Many assembly / machine language instruction sets include the noop instruction.I've seen delay loops written that would have a few noops and a counter in an outside loop.Hey, it worked. |
|
|
|
X002548
Not Just a Number
15586 Posts |
Posted - 2004-01-09 : 15:29:16
|
quote:
Originally posted by ehorn
I had never heard that term used before. Websters The Free On-line Dictionary of Computing defines no-op as:quote:
1. A machine instruction that does nothing (sometimes used in assembler-level programming as filler for data or patch areas, or to overwrite code to be removed in binaries).
I would most definately agree with your assessment.EDIT: Fixed Source
I've seen a lot of developers that that description applies too.Brett8-) |
|
|
|
SamC
White Water Yakist
3467 Posts |
Posted - 2004-01-09 : 15:43:54
|
| Yes. But I've seen it apply to management much more.Don't forget the cover letters on the TPS reports. |
|
|
|
|
PINing