Re: RE: SQL Performance question

Thank you John. Yes, the tempdbs created with "T" option and the create temp with no log. I compared all the other statistics information. I found the bufwrite happened on load data into temp table and create index on temp table. From sqltrace The original sql take 0.180462442569 second The sql with temp table take 0.111681263591 second . create temp table 0.043224455991 create index 0.036193310316 select 0.032263497284 Thank you.

Sorry I did not come up with a solution rather a question that how you calculated these times (execution time of table create,populate) with that precision? It will help me to benchmark execution time of two process executed in fraction of second.