Multiple TEMDB files, for multiple CPU's

  • Hey All,

    We are about roll out a new a/p and are going to try having multiple tempdb files.

    The machine spec is: 128GB RAM, 4* SAN volumes (sys db's, sys logs, user db's, user log's) and 4* Quad-Core processors (16 Cores).

    We already have a similar cluster which only has 1 tempdb file and do not see any contention or performance issues.

    Question is: Do we have 1 temdb file per physical processor (i.e., 4 quad-core cpu's) or do we have 1 file per core? There are articles out there that suggest either or but no-one has definitively said 'do it this way'.

    Any suggestions or past experience advise gratefully accepted.

    Cheers,

    Adam.

    Adam Zacks-------------------------------------------Be Nice, Or Leave

  • Taken from: http://technet.microsoft.com/en-us/library/cc966534.aspx

    Best practises suggest you:

    Consider configuration of TEMPDB database

    Make sure to move TEMPDB to adequate storage and pre-size after installing SQL Server.

    Performance may benefit if TEMPDB is placed on RAID 1+0 (dependent on TEMPDB usage).

    For the TEMPDB database, create 1 data file per CPU, as described in #8 below.

    Lining up the number of data files with CPU’s has scalability advantages for allocation intensive workloads.

    It is recommended to have .25 to 1 data files (per filegroup) for each CPU on the host server.

    This is especially true for TEMPDB where the recommendation is 1 data file per CPU.

    Dual core counts as 2 CPUs; logical procs (hyperthreading) do not.

    There are alot of resources and some pretty good reads regarding the subject.

    http://msdn.microsoft.com/en-us/library/cc966545.aspx

    I personally test with one data file per cpu, and then 1 per core.

    -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
    This thing is addressing problems that dont exist. Its solution-ism at its worst. We are dumbing down machines that are inherently superior. - Gilfoyle

  • Thanks. Will look forward to a good read 🙂

    Adam Zacks-------------------------------------------Be Nice, Or Leave

  • I would start with 0.25 files/core, all on the same drive then, if there is allocation contention, add more files, if there is IO contention move the files onto separate LUNs.

    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

    We walk in the dark places no others will enter
    We stand on the bridge and no one may pass

Viewing 4 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic. Login to reply