Home My Page Projects Code Snippets Project Openings diderot
Summary Activity Tracker Tasks SCM

SCM Repository

[diderot] Diff of /web/htdocs/index.html
ViewVC logotype

Diff of /web/htdocs/index.html

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 374, Sun Oct 3 18:56:32 2010 UTC revision 375, Mon Oct 4 12:45:40 2010 UTC
# Line 18  Line 18 
18  We are particularly interested in a class of algorithms that are programmed  We are particularly interested in a class of algorithms that are programmed
19  in terms of <em>continuous</em> scalar, vector, and tensor fields that  in terms of <em>continuous</em> scalar, vector, and tensor fields that
20  are reconstructed from the image data.  are reconstructed from the image data.
21    Our goals are to provide a high-level mathematical programming model for
22    these algorithms, while also providing high-performance implementations
23    on a variety of parallel hardware platforms.
24    </p>
26    <h3>Language overview</h3>
27    <p>
28    The following is an overview of our current preliminary design for Diderot.
29    We are building a compiler for this design and expect that the design will
30    evolve as we get experience with the implementation.
31    Also, the design is conservative, in that it does not provide all of the features that
32    we plan to provide (<i>e.g.</i>, actor-actor interactions).
33    </p>
35    <h4>Types</h4>
36    <p>
37    The main type of computational value in Diderot is a <em>tensor</em>, which includes
38    reals (0-order tensors), vectors, and matrices.
39    In addition, Diderot provides booleans, integers, and strings.
40    Diderot also has three <em>abstract</em> types:
41    </p>
42    <dl>
43      <dt>images</dt>
44        <dd>are used to represent the data being analyzed, as well as other array data, such as transfer
45    functions.
46        </dd>
47      <dt>kernels</dt>
48        <dd>are separable convolution kernels</dd>
49      <dt>fields</dt>
50        <dd>are an abstraction of functions from 1D, 2D, or 3D space to some tensor type.
51        A field is defined by convolving an image with a kernel.
52        </dd>
53    </dl>
55    <h4>Program structure</h4>
56    <p>A Diderot program is organized into three logical sections:</p>
57    <ol>
58      <li>Global declarations define global values, such as fields, as well as the inputs to the program.</li>
59      <li>Actor definitions define the computational agents that implement the program</li>
60      <li>Initialization defines the initial set of actors and their structure</li>
61    </ol>
63    <h4>Actors</h4>
64    <p>
65    An <em>actor</em> represents a mostly autonomous computation with local state, which includes
66    their <em>position</em> in world space.
67    An actor definition consists of declared state variables and methods.
68    All actors must have an <em>update</em> method and may optionally have a <em>stabilize</em>
69    method.
70    </p>
72    <h4>Execution model</h4>
73    <p>The Diderot execution model is <em>bulk synchronous</em>.
74    At each iteration, the update methods of all active actors are invoked, resulting in a new
75    configuration.
76  </p>  </p>
78  <h3>People</h3>  <h3>People</h3>
# Line 35  Line 90 
90  </p>  </p>
92  <hr />  <hr />
93  Last modified: September 30, 2010.  Last modified: October 4, 2010.
94  <hr />  <hr />
96  </body>  </body>

Removed from v.374  
changed lines
  Added in v.375

ViewVC Help
Powered by ViewVC 1.0.0