Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • parallel error

    hi,
    I installed the parallel package for stata mp and am trying to execute a command using parallel:reshape long


    i first set the clusters and get the following result:

    parallel setclusters 4
    N Clusters: 4
    Stata dir: C:\Program Files (x86)\Stata15/StataMP-64.exe



    however , i keep getting the following error after entering the parallel reshape long command, or any other command as well:
    -----------------------------------------------------------------------
    Parallel Computing with Stata (by GVY)
    Clusters : 4
    pll_id : kiaua6hsh1
    Running at : \\sts.ad.ualberta.ca\UofA\Home\SPH\vishal\My Documents
    Randtype : datetime
    Waiting for the clusters to finish...
    cluster 0001 has finished with a connection error -601- (timeout) (see more)...
    --Break--
    r(1);
    -------------------------------------------------------------------------------------

    i am wondering if it's a stata dir issue but i couldnt change the directory.

    any thoughts?

    thanks
    vishal

  • #2
    With this kind of program specific problem, you may need to contact the author. I, and I suspect many of those active on Statalist, simply pay for Stata's own parallel (i.e., multiprocessor) capabilities.

    Comment


    • #3
      What do you get if you run:
      Code:
      which parallel
      
      sysuse bplong.dta
      sort patient
      parallel initialize 2
      parallel, by(patient): by patient: egen max_bp = max(bp)
      Report code and result using the code tags, see FAQ 12.3 How to use CODE delimiters.

      Comment


      • #4
        I add a comment on Stata/MP vs the parallel package:

        Stata/MP: One Stata session doing the same task using several processors/cores with shared memory on one computer (node). Stata/MP provides symmetric multiprocessing (SMP) using the Open/MP API.

        parallel: Severall Stata sessions, running independent tasks on one or more computers (nodes).

        One example of using -parallel- could be running the same estimation, e.g. mlogit, on 60 imputed datasets.

        (The above is different from parallel computing on a computer cluster: running simultaneous computations of the same task
        across networks of computers with "globally addressable" memory, using MPI or hybrid OpenMP/MPI.)
        Last edited by Bjarte Aagnes; 02 Jun 2019, 13:21.

        Comment


        • #5
          Vishal Sharma
          The key line is:

          Code:
          Running at : \\sts.ad.ualberta.ca\UofA\Home\SPH\vishal\My Documents
          This indicates that the instance is not running locally but is running and/or writing data to a network location. The connection error could be one of a thousand different things. If you can run the command locally try that instead of trying to connect to multiple servers.

          Comment


          • #6
            Dear all

            I am getting the same error as @Vishal and I am running the script below locally:

            Code:
            clear
            cd "/Applications/Stata"
            parallel setclusters 2, statapath(/Applications/Stata)
            sysuse auto, clear
            parallel bs, rep(1000): regress mpg weight gear foreign
            This is the message I get:

            Code:
            . parallel bs, rep(1000): regress mpg weight gear foreign
            --------------------------------------------------------------------------------
            Parallel Computing with Stata (by GVY)
            Clusters   : 2
            pll_id     : 2jggphu610
            Running at : /Applications/Stata
            Randtype   : datetime
            Waiting for the clusters to finish...
            cluster 0001 has finished with a connection error -601- (timeout) (see more)...
            cluster 0002 has finished with a connection error -601- (timeout) (see more)...
            --Break--
            r(1);

            Does anyone have an idea why?

            Thanks!

            Comment

            Working...
            X