Ignore:
Timestamp:
Oct 26, 2016 9:12:33 AM (8 years ago)
Author:
smk78
Branches:
master, ESS_GUI, ESS_GUI_Docs, ESS_GUI_batch_fitting, ESS_GUI_bumps_abstraction, ESS_GUI_iss1116, ESS_GUI_iss879, ESS_GUI_iss959, ESS_GUI_opencl, ESS_GUI_ordering, ESS_GUI_sync_sascalc, costrafo411, magnetic_scatt, release-4.1.1, release-4.1.2, release-4.2.2, ticket-1009, ticket-1094-headless, ticket-1242-2d-resolution, ticket-1243, ticket-1249, ticket885, unittest-saveload
Children:
20cfa23
Parents:
af6de50
Message:

Reviewed new text by PaulK. Minor corrections.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • src/sas/sasgui/perspectives/fitting/media/plugin.rst

    raf6de50 r907186d  
    2727 
    2828the next time SasView is started it will compile the plugin and add 
    29 it to the list of *Plugin Models* in a FitPage. 
     29it to the list of *Customized Models* in a FitPage. 
    3030 
    3131SasView models can be of three types: 
     
    6161- If necessary use underscore to separate words (i.e. barbell not BarBell or 
    6262  broad_peak not BroadPeak) 
    63 - Do not include “modelâ€ᅵ in the name (i.e. barbell not BarBellModel) 
     63- Do not include "model" in the name (i.e. barbell not BarBellModel) 
    6464 
    6565 
     
    486486and it includes orientation parameters. 
    487487 
    488 *form_volume* defines the volume of the shape. As in python models, 
     488*form_volume* defines the volume of the shape. As in python models, it  
    489489includes only the volume parameters. 
    490490 
     
    673673particular problems: 
    674674 
    675   (1) Code is too complex, or uses too much memory.  GPU devices only have a 
     675  **(1) Code is too complex, or uses too much memory.**  GPU devices only have a 
    676676  limited amount of memory available for each processor.  If you run programs 
    677677  which take too much memory, then rather than running multiple values in parallel 
     
    680680  some platforms, or worse, cause the screen to go blank or the system to reboot. 
    681681 
    682   (2) Code takes too long.  Because GPU devices are used for the computer 
     682  **(2) Code takes too long.**  Because GPU devices are used for the computer 
    683683  display, the OpenCL drivers are very careful about the amount of time they 
    684684  will allow any code to run.  For example, on OS X, the model will stop running 
    685   after 5 seconds regardless if the computation is complete.  You may end up 
    686   with only some of your 2-D array defined, with the rest containing random 
     685  after 5 seconds regardless of whether the computation is complete.  You may end up 
     686  with only some of your 2D array defined, with the rest containing random 
    687687  data. Or it may cause the screen to go blank or the system to reboot. 
    688688 
    689   (3) Memory is not *aligned*.  The GPU hardware is specialized to operate on 
    690   multiple values simultaneously.  To keep the GPU simpler the values in memory 
     689  **(3) Memory is not aligned**.  The GPU hardware is specialized to operate on 
     690  multiple values simultaneously.  To keep the GPU simple the values in memory 
    691691  must be aligned with the different GPU compute engines.  Not following these 
    692692  rules can lead to unexpected values being loaded into memory, and wrong answers 
    693693  computed.  The conclusion from a very long and strange debugging session was 
    694694  that any arrays that you declare in your model should be a multiple of four. 
    695   For example 
     695  For example:: 
    696696 
    697697      double Iq(q, p1, p2, ...) 
     
    702702 
    703703The first step when your model is behaving strangely is to set **single=False**. 
    704 This automatically restricts the model to only run on the CPU, or on high end 
    705 GPU cards.  There can still be problems even on high end cards, so you can force 
     704This automatically restricts the model to only run on the CPU, or on high-end 
     705GPU cards.  There can still be problems even on high-end cards, so you can force 
    706706the model off the GPU by setting **opencl=False**.  This runs the model 
    707707as a normal C program without any GPU restrictions so you know that 
     
    945945.. ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ 
    946946 
    947 .. note::  This help document was last changed by Steve King, 10Oct2016 
     947.. note::  This help document was last changed by Steve King, 25Oct2016 
Note: See TracChangeset for help on using the changeset viewer.