Re: Defining block size during runtime

From: Paul H. Hargrove (PHHargrove_at_lbl_dot_gov)
Date: Fri Jul 24 2009 - 22:20:04 PDT

  • Next message: Gary Funck: "Re: Defining block size during runtime"
    I am able to run on a Cray XT4 w/o problems.
    So I may try to reproduce your problem, could you tell me what compiler 
    version are you using?
    sainath l wrote:
    > Hi,
    > Thank you very much for answering my questions Paul. And extremely 
    > sorry for not providing the "gettime.h" file. Will make sure that I 
    > provide all the related files from next time.
    > The code is running fine in an smp X4600 SMP node with 16 procs.
    > But it is not running in XT 4.
    > when I run it in XT 4 the code breaks during the first iteration. the 
    > first iteration does not complete. the printf after the upc_free(B) 
    > command does not execute.
    > What do you think might be the reason for this? I get the same error 
    > message which I had pasted in my earlier post.
    > Thanks in Advance.
    > Cheers,
    > sainath
    > On Fri, Jul 24, 2009 at 11:00 PM, Paul H. Hargrove <PHHargrove_at_lbl_dot_gov 
    > <mailto:PHHargrove_at_lbl_dot_gov>> wrote:
    >     I have run your code (I needed to provide a gettime.h) and did not
    >     see any errors.  I tried on both an x86 cluster with myrinet and
    >     on a CrayXT.
    >     To answer your question: I don't believe that use of this data
    >     structure will cause any performance penalty for the collectives,
    >     since the structure is just a "trick" for indexing the block of
    >     data.  Additionally, static-vs-dynamic allocation of memory should
    >     not have an effect on the collectives performance either.
    >     --Paul
    >     sainath l wrote:
    >         Hi paul,
    >         I have attached my code. The first iteration runs till the
    >         deallocation part and then the code breaks.
    >         *** Caught a fatal signal: SIGSEGV(11) on node 0/16
    >         _pmii_daemon(SIGCHLD): PE 0 exit signal Segmentation fault
    >         [NID 26]Apid 315852: initiated application termination
    >         Also I would be very happy to know, if I want to write
    >         micro-benchmarks for the collectives will using this
    >         datastructure be of any problem ? (overhead incurred by using
    >         this datastructure)  Or should I just declare arrays
    >         statically and use them.
    >         In practice, in general, are the source and destination
    >         variables of collective operations dynamically allocated ?  If
    >         yes will that degrade the perfromance.
    >         Thank you very much.
    >         Cheers,
    >         Sainath
    Paul H. Hargrove                          PHHargrove_at_lbl_dot_gov
    Future Technologies Group                 Tel: +1-510-495-2352
    HPC Research Department                   Fax: +1-510-486-6900
    Lawrence Berkeley National Laboratory     

  • Next message: Gary Funck: "Re: Defining block size during runtime"