Changes in / [bf09f55:62d7601] in sasmodels


Ignore:
Files:
1 added
9 edited

Legend:

Unmodified
Added
Removed

  • doc/genmodel.py

    r745b7bb rb866abf  
    33import sys, os, math, re 
    44import numpy as np 
     5import matplotlib 
     6matplotlib.use('Agg') 
    57import matplotlib.pyplot as plt 
    68sys.path.insert(0, os.path.abspath('..')) 

  • doc/guide/index.rst

    r2e66ef5 rc0d7ab3  
    99   intro.rst 
    1010   install.rst 
     11   gpu_setup.rst 
    1112   pd/polydispersity.rst 
    1213   resolution.rst 

  • doc/guide/install.rst

    rf8a2baa rc0d7ab3  
    5353This will allow you to edit the files in the package and have the changes 
    5454show up immediately in python the next time you load your program. 
    55  
    56 OpenCL Installation 
    57 ******************* 
    58 *Warning! GPU devices do not in general offer the same level of memory 
    59 protection as CPU devices. If your code attempts to write outside allocated 
    60 memory buffers unpredicatable behaviour may result (eg, your video display 
    61 may freeze, or your system may crash, etc). Do not install OpenCL drivers 
    62 without first checking for known issues (eg, some computer manufacturers 
    63 install modified graphics drivers so replacing these may not be a good 
    64 idea!). If in doubt, seek advice from an IT professional before proceeding 
    65 further.* 
    66  
    67 Check if you have OpenCL already installed 
    68 ========================================== 
    69  
    70 **Windows** 
    71  
    72 The following instructions are based on 
    73 http://web.engr.oregonstate.edu/~mjb/cs475/DoIHaveOpenCL.pdf 
    74  
    75 * Go to: Start -> Control Panel -> System & Security -> Administrative Tools 
    76 * Double Click on Computer Managment 
    77 * Click on Device Manager 
    78 * Click open Display Adapters 
    79 * Right-click on available adapter and select Properties 
    80 * Click on Driver 
    81 * Go to Driver Details 
    82 * Scroll down and see if OpenCL is installed (look for OpenCL*.dll files) 
    83  
    84 **Mac OSX** 
    85  
    86 For OS X operating systems higher than 10.6 OpenCL is shipped along with 
    87 the system. 
    88  
    89 However, OpenCL has had a rocky history on Macs. Apple provide a useful 
    90 compatibility table at https://support.apple.com/en-us/HT202823 
    91  
    92  
    93 Installation 
    94 ============ 
    95  
    96 **Windows** 
    97  
    98 Depending on the graphic card in your system, drivers 
    99 can be obtained from different sources: 
    100  
    101 * NVIDIA: https://developer.nvidia.com/opencl 
    102 * AMD: http://developer.amd.com/tools-and-sdks/opencl-zone/ 
    103  
    104  
    105 **Mac OSX** 
    106  
    107 N/A 
    108  
    109 You cannot download OpenCL driver updates for your Mac. They are packaged 
    110 with the normal quarterly OS X updates from Apple. 
    111  
    112  
    113 .. note:: 
    114     Intel provides OpenCL drivers for Intel processors at 
    115     https://software.intel.com/en-us/articles/opencl-drivers 
    116     These can sometimes make use of special vector instructions across multiple 
    117     processors, so it is worth installing if the GPU does not support double 
    118     precision. You can install this driver alongside the GPU driver for NVIDIA 
    119     or AMD. 
    120  
    121  
    122 GPU Selection 
    123 ************* 
    124  
    125 sasmodels evaluations can run on your graphics card (GPU) or they can run 
    126 on the processor (CPU). In general, calculations performed on the GPU will run faster. 
    127  
    128 To run on the GPU, your computer must have OpenCL drivers installed. 
    129 For information about OpenCL installation see this 
    130 :ref:`opencl-installation` guidance. 
    131  
    132 Where the model is evaluated is a little bit complicated. 
    133 If the model has the line *single=False* then it requires double precision. 
    134 If the GPU is single precision only, then it will try running via OpenCL 
    135 on the CPU.  If the OpenCL driver is not available for the CPU then 
    136 it will run as a normal program on the CPU. 
    137  
    138 For models with a large number of parameters or with a lot of code, 
    139 the GPU may be too small to run the program effectively. 
    140 In this case, you should try simplifying the model, maybe breaking it 
    141 into several different modules so that you don't need *IF* statements in your code. 
    142 If it is still too big, you can set *opencl=False* in the model file and 
    143 the model will only run as a normal program on the CPU. 
    144 This will not usually be necessary. 
    145  
    146 Device Selection 
    147 ================ 
    148 If you have multiple GPU devices you can tell SasView which device to use. 
    149 By default, SasView looks for one GPU and one CPU device 
    150 from available OpenCL platforms. 
    151  
    152 SasView prefers AMD or NVIDIA drivers for GPU, and prefers Intel or 
    153 Apple drivers for CPU. Both GPU and CPU are included on the assumption that CPU 
    154 is always available and supports double precision. 
    155  
    156 The device order is important: GPU is checked before CPU on the assumption that 
    157 it will be faster. By examining ~/sasview.log you can see which device SasView 
    158 chose to run the model. 
    159  
    160 **If you don't want to use OpenCL, you can set** *SAS_OPENCL=None* 
    161 **in your environment settings, and it will only use normal programs.** 
    162  
    163 If you want to use one of the other devices, you can run the following 
    164 from the python console in SasView:: 
    165  
    166     import pyopencl as cl 
    167     cl.create_some_context() 
    168  
    169 This will provide a menu of different OpenCL drivers available. 
    170 When one is selected, it will say "set PYOPENCL_CTX=..." 
    171 Use that value as the value of *SAS_OPENCL*. 
    172  
    173 Compiler Selection 
    174 ================== 
    175 For models run as normal programs, you may need to specify a compiler. 
    176 This is done using the SAS_COMPILER environment variable. 
    177 Set it to *tinycc* for the tinycc compiler, *msvc* for the 
    178 Microsoft Visual C compiler, or *mingw* for the MinGW compiler. 
    179 TinyCC is provided with SasView so that is the default. 
    180 If you want one of the other compilers, be sure to have it available 
    181 in your *PATH* so SasView can find it! 
    182  
    183  
    184 *Document History* 
    185  
    186 | 2017-05-17 Paul Kienzle 

  • doc/guide/magnetism/magnetism.rst

    r990d8df r59485a4  
    8787    The values of the 'Up_frac_i' and 'Up_frac_f' must be in the range 0 to 1. 
    8888 
    89 .. note:: 
    90     This help document was last changed by Steve King, 02May2015 
     89*Document History* 
    9190 
    92 * Document History * 
    93  
     91| 2015-05-02 Steve King 
    9492| 2017-05-08 Paul Kienzle 

  • doc/guide/plugin.rst

    r30b60d2 r3048ec6  
    7878     at the start of the model documentation and as a tooltip in the SasView GUI 
    7979 
    80 - a **short discription**: 
     80- a **short description**: 
    8181   - this is the **description** string in the *.py* file 
    8282   - this is a medium length description which appears when you click 
    … …  
    233233 
    234234**name = "mymodel"** defines the name of the model that is shown to the user. 
    235 If it is not provided, it will use the name of the model file, with '_' 
    236 replaced by spaces and the parts capitalized.  So *adsorbed_layer.py* will 
    237 become *Adsorbed Layer*.  The predefined models all use the name of the 
    238 model file as the name of the model, so the default may be changed. 
     235If it is not provided it will use the name of the model file. The name must 
     236be a valid variable name, starting with a letter and contains only letters, 
     237numbers or underscore.  Spaces, dashes, and other symbols are not permitted. 
    239238 
    240239**title = "short description"** is short description of the model which 
    … …  
    298297- **"name"** is the name of the parameter shown on the FitPage. 
    299298 
     299  - the name must be a valid variable name, starting with a letter and 
     300    containing only letters, numbers and underscore. 
     301 
    300302  - parameter names should follow the mathematical convention; e.g., 
    301303    *radius_core* not *core_radius*, or *sld_solvent* not *solvent_sld*. 
    … …  
    366368    dispersion. 
    367369 
     370Some models will have integer parameters, such as number of pearls in the 
     371pearl necklace model, or number of shells in the multi-layer vesicle model. 
     372The optimizers in BUMPS treat all parameters as floating point numbers which 
     373can take arbitrary values, even for integer parameters, so your model should 
     374round the incoming parameter value to the nearest integer inside your model 
     375you should round to the nearest integer.  In C code, you can do this using:: 
     376 
     377    static double 
     378    Iq(double q, ..., double fp_n, ...) 
     379    { 
     380        int n = (int)(fp_n + 0.5); 
     381        ... 
     382    } 
     383 
     384in python:: 
     385 
     386    def Iq(q, ..., n, ...): 
     387        n = int(n+0.5) 
     388        ... 
     389 
     390Derivative based optimizers such as Levenberg-Marquardt will not work 
     391for integer parameters since the partial derivative is always zero, but 
     392the remaining optimizers (DREAM, differential evolution, Nelder-Mead simplex) 
     393will still function. 
    368394 
    369395Model Computation 
    … …  
    391417     \approx \frac{\sum_{i=1}^n G(x_i) I(q,x_i)}{\sum_{i=1}^n G(x_i) V(x_i)} 
    392418 
    393 That is, the indivdual models do not need to include polydispersity 
     419That is, the individual models do not need to include polydispersity 
    394420calculations, but instead rely on numerical integration to compute the 
    395421appropriately smeared pattern.   Angular dispersion values over polar angle 
    … …  
    411437The parameters *par1, par2, ...* are the list of non-orientation parameters 
    412438to the model in the order that they appear in the parameter table. 
    413 **Note that the autogenerated model file uses** *x* **rather than** *q*. 
     439**Note that the auto-generated model file uses** *x* **rather than** *q*. 
    414440 
    415441The *.py* file should import trigonometric and exponential functions from 
    … …  
    935961across multiple parameters can be very slow). 
    936962 
    937 If your model has 2D orientational calculation, then you should also 
     963If your model has 2D orientation calculation, then you should also 
    938964test with:: 
    939965 

  • sasmodels/conversion_table.py

    rd3e3f756 r505d0ad  
    1111account for that. 
    1212 
    13 Usage: 
    14 <old_Sasview_version> : { 
    15     <new_model_name> : [ 
    16         <old_model_name> , 
    17         { 
    18             <new_param_name_1> : <old_param_name_1>, 
    19             ... 
    20             <new_param_name_n> : <old_param_name_n> 
    21         } 
    22     ] 
    23 } 
     13Usage:: 
     14 
     15    <old_Sasview_version> : { 
     16        <new_model_name> : [ 
     17            <old_model_name> , 
     18            { 
     19                <new_param_name_1> : <old_param_name_1>, 
     20                ... 
     21                <new_param_name_n> : <old_param_name_n> 
     22            } 
     23        ] 
     24    } 
    2425 
    2526Any future parameter and model name changes can and should be given in this 

  • sasmodels/custom/__init__.py

    r1a6cd57 r0f48f1e  
    3131        if fullname in sys.modules: 
    3232            del sys.modules[fullname] 
     33        if path.endswith(".py") and os.path.exists(path) and os.path.exists(path+"c"): 
     34            # remove automatic pyc file before loading a py file 
     35            os.unlink(path+"c") 
    3336        module = imp.load_source(fullname, os.path.expanduser(path)) 
    34         #os.unlink(path+"c")  # remove the automatic pyc file 
    3537        return module 
    3638 

  • sasmodels/kernelpy.py

    r883ecf4 r3b32f8d  
    3434        self.info = model_info 
    3535        self.dtype = np.dtype('d') 
     36        logging.info("load python model " + self.info.name) 
    3637 
    3738    def make_kernel(self, q_vectors): 
    38         logging.info("creating python kernel " + self.info.name) 
    3939        q_input = PyInput(q_vectors, dtype=F64) 
    4040        kernel = self.info.Iqxy if q_input.is_2d else self.info.Iq 

  • sasmodels/sasview_model.py

    r9644b5a r9f8ade1  
    1616import logging 
    1717from os.path import basename, splitext, abspath, getmtime 
    18 import thread 
     18try: 
     19    import _thread as thread 
     20except ImportError: 
     21    import thread 
    1922 
    2023import numpy as np  # type: ignore 
Note: See TracChangeset for help on using the changeset viewer.