"Fossies" - the Fresh Open Source Software Archive

Member "dx-4.4.4/doc/README_SMP" (25 May 1999, 3057 Bytes) of package /linux/misc/old/dx-4.4.4.tar.gz:


As a special service "Fossies" has tried to format the requested text file into HTML format (style: standard) with prefixed line numbers. Alternatively you can here view or download the uninterpreted source code file.

    1 *
    2 * README for IBM Visualization Data Explorer for Symetric Multiprocessors 
    3 *
    4 *
    5   
    6   
    7 The information contained in this file consists of hints for using the
    8 Data Explorer SMP for symmetric multiprocessors.
    9   
   10 ========================================================================
   11   
   12 CONTENTS
   13     To go directly to an item, search for the item number beginning
   14     at the left margin.
   15  
   16     1. The Partition module
   17     2. Partitioning an already-partitioned field 
   18     3. The default amount of memory used
   19     4. The default number of processors used
   20     5. The difference between SMP and distributed 
   21  
   22 ========================================================================
   23  
   24  
   25 1. The Partition module 
   26  
   27 With Data Explorer SMP, the Partition module should be used to
   28 create a composite field from the imported data. 
   29 Only if the data set is partitioned 
   30 will Data Explorer SMP exploit the multiple processors available.  
   31 Note that you can run partitioned data equally well on both
   32 uniprocessor and multiprocessor machines. 
   33  
   34 2. Partitioning an already-partitioned field 
   35  
   36 Once a field is partitioned it will not be repartitioned if it 
   37 is passed through the Partition module again. Thus to maximize
   38 the use of the available processors, you may want 
   39 to increase the number of partitions (the "n" paramter for 
   40 Partition) from the default value if you will be operating on 
   41 only a subset of the data later on. Note however, that over-partitioning
   42 a field leads to extra overhead.
   43  
   44 3. The default amount of memory used
   45  
   46 By default on a MP machine, less memory will be used than on
   47 a UP machine. This is under the assumption that other users
   48 may be more likely to be using an MP machine. The amount of
   49 memory used can be changed with the -memory command line option
   50 to dx.  Usage("memory",0) will show you how much memory is being 
   51 used.
   52  
   53 4. The default number of processors used
   54  
   55 For an MP machine, if the number of processors (n) is 4 or greater,
   56 the default number of processors used is n/2, rounded down. If the
   57 number of processors is 2 or 3, then the default number of processors
   58 used is 2. The number of processors used can be changed with the
   59 -processors command line option to dx.
   60  
   61 5. The difference between SMP and distributed 
   62  
   63 When you run parallel on an SMP machine, intramodule parallelism
   64 is used. For example, the Isosurface module will operate in
   65 parallel on different portions of a partitioned field. When
   66 you use the distributed method of execution, different modules
   67 run on different machines; for example, Isosurface might be
   68 running on one machine while MapToPlane is running on another.
   69 There is overhead involved with using distributed computation, as
   70 the data must be transferred between machines.
   71  
   72   
   73 ========================================================================
   74  
   75 The following trademarks apply to this information:
   76  
   77 IBM is a registered trademark of International Business Machines Corporation. 
   78 IBM Visualization Data Explorer is a registered trademark of International
   79     Business Machines Corporation