Home My Page Projects Code Snippets Project Openings diderot

# SCM Repository

[diderot] Diff of /branches/pure-cfg/TODO
 [diderot] / branches / pure-cfg / TODO

# Diff of /branches/pure-cfg/TODO

revision 763, Thu Apr 7 19:57:56 2011 UTC revision 950, Fri Apr 22 19:27:39 2011 UTC
# Line 1  Line 1
1  Outer products  ========================
2      syntax [DONE]  SHORT TERM =============  (for curvature-based VR)
3      typechecking  ========================
IL and codegen
4
5  Add M dot v, v dot M, and M dot N  IL support for higher-order tensor values (matrices, etc).
6        tensor construction [DONE]
7        tensor indexing [DONE]
8        tensor slicing
9        verify that hessians work correctly [DONE]
10
11  Identity matrix [DONE up to code generation]  Outer products [DONE]
12
13  trace [DONE]  Add M dot v, v dot M, and M dot N [DONE]
14
15  expand trace in mid to low translation  Identity matrix [DONE]
16
17  extend norm (|exp|) to all tensor types  Zero tensor [DONE]
18
19  Infix dot product and cross product [DONE]  trace [DONE]
20
21  lerp on scalars and vectors [DONE]  extend norm (|exp|) to tensor[3,3] [DONE]
22
23    Matrix addition, subtraction, and scaling  [DONE]
24
25  Use ∇⊗ etc syntax  Use ∇⊗ etc. syntax
26      syntax [DONE]      syntax [DONE]
27      typechecking      typechecking
28      IL and codegen      IL and codegen
# Line 28  Line 33
33          value construction: '{' e1 ',' … ',' en '}'          value construction: '{' e1 ',' … ',' en '}'
34          indexing: e '{' e '}'          indexing: e '{' e '}'
35
36  IL support for higher-order tensor values (matrices, etc).  Infix dot product and cross product [DONE]
37      tensor construction
38      tensor indexing and slicing  lerp on scalars and vectors [DONE]
39
40    Infix "^" operator for pow() [DONE]
41
42    Code generation support for 1D image data, such as RGBA transfer functions [DONE]
43
44    ==============================
45    other SHORT TERM =============  (including needed for LIC)
46    ==============================
47
48    test/uninit.diderot:
49    documents need for better compiler error messages when output variables
50    are not initialized; the current messages are very cryptic
51
52    determinant ("det") for tensor[3,3]
53
54    extend norm (|exp|) to all tensor types [DONE for vectors and matrices]
55
56    expand trace in mid to low translation
57
58    value-numbering optimization
59
60  Add type aliases for color types  Add type aliases for color types
61      rgb = real{3}      rgb = real{3}
62      rgba = real{4}      rgba = real{4}
63
64    ==============================
65    MEDIUM TERM ================== (including needed for streamlines & tractography)
66    ==============================
67
68    Want: allow X *= Y, X /= Y, X += Y, X -= Y to mean what they do in C,
69    provided that X*Y, X/Y, X+Y, X-Y are already supported.
70    Nearly every Diderot program would be simplified by this.
71
72    Want: ability to apply "normalize" to a field itself, e.g.
73      field#1(2)[2] V = normalize(Vimg ⊛ ctmr);
74    so that V(x) = normalize((Vimg ⊛ ctmr)(x)).
75    (For now, it would be okay for the differentability of V to be 0;
76    later we can teach Diderot the quotient rule of derivatives.)
77    Having this would simplify expression of standard LIC method, and
78    would also help express other vector field expressions that arise
79    in vector field feature exraction
80    Related to this and possibly simpler:
82      field#0(2)[] Vlen = |Vimg ⊛ bspln3|;
83    to get a scalar field of vector length, or
84      field#2(2)[] F = Fimg ⊛ bspln3;
85      field#0(2)[] Gmag = |∇F|;
86    to get a scalar field of gradient magnitude, or
87      field#2(2)[] F = Fimg ⊛ bspln3;
88      field#0(2)[] Gmsq = ∇F•∇F;
89    to get a scalar field of squared gradient magnitude, which is simpler
90    to differentiate.  Again, though, it would be such a benefit to have
91    these expressions, without any differentiability.  Hence my indication
92    of "field#0" for these above.
93
94    ability to declare a field in such a way so that probe
95    positions are *always* clamped to the support of "inside";
96    there are many cases where this is the sensible behavior
97
98    ability to emit/track/record variables into dynamically re-sized
99    runtime buffer
100
101    evals & evecs for symmetric tensor[3,3]
102
103    tensor fields: convolution on general tensor images
104
105    Use of Teem's "hest" command-line parser for getting
106    any input variables that are not defined in the source file
107
108    Save Diderot output to nrrd, instead of "mip.txt"
109      For grid of strands, save to similarly-shaped array
110      For list of strands, save to long 1-D (or 2-D for non-scalar output) list
111      For ragged things (like tractography output), will need to save both
112        complete list of values, as well as list of start indices and lengths
113        to index into complete list
114
115    ==============================
116    other MEDIUM TERM ============ (needed for particles)
117    ==============================
118
119    run-time birth and death of strands
120
121    "initially" supports lists
122
123    "initially" supports lists of positions output from
124    different initalization Diderot program
125
126    spatial data structure that permits strands' queries of neighbors
127
128    proper handling of stabilize method
129
130    test/vr-kcomp2.diderot: Add support for code like
131
132            (F1 if x else F2)@pos
133
134    This will require duplication of the continuation of the conditional (but we should only
135    duplicate over the live-range of the result of the conditional.
136
137    ==============================
138    other MEDIUM TERM ============
139    ==============================
140
141    want: warnings when "D" (reserved for differentiation) is declared as
142    a variable name (get confusing error messages now)
143
144    support for Python interop and GUI
145
146    ==============================
147    LONG TERM ====================
148    ==============================
149
150    Better handling of variables that determines the scope of a variable
151    based on its actual use, instead of where the user defined it.  So,
152    for example, we should lift strand-invariant variables to global
153    scope.  Also prune out useless variables, which should include field
154    variables after the translation to mid-il.
155
156    co- vs contra- index distinction
157
158    add ":" for tensor dot product (contracts out two indices
159    instead of one like •), valid for all pairs of tensors with
160    at least two indices
161
162    some indication of tensor symmetry
163    (have to identify the group of index permutations that are symmetries)
164
165    dot works on all tensors
166
167    outer works on all tensors
168
169    Einstein summation notation
170
171    "tensor comprehension" (like list comprehension)
172
173    Python/ctypes interface to run-timez
174
175    ======================
176    BUGS =================
177    ======================
178
179    test/fields.diderot:
180    documents various bugs/questions associated with field expressions
181
182    test/mip-cam.diderot:
183        // generated C code looks like "(float)p_ui_0--0.5e0f"
184        //real rayU = 2.0*camUmax*(real(ui) - -0.5)/real(imgResU) - camUmax;
185
186    test/vr-kcomp.diderot:
187        // HEY (scoping BUG): the strand parameters (e.g. ui, vi) ...
188
189    test/zslice2.diderot:
190    // but is this not valid syntax for creating field in one shot?
191    //field#2(3)[] F = (load("../data/zimg112.nrrd")) ⊛ bspln3;
192    // It is valid syntax, but there is a bug in the conversion from HighIL to MidIL
193
194    test/zslice2.diderot:
195    // HEY (bug) bspln5 leads to problems ...
196    //  uncaught exception Size [size]
197    //    raised at c-target/c-target.sml:47.15-47.19
198    //field#4(3)[] F = img ⊛ bspln5;

Legend:
 Removed from v.763 changed lines Added in v.950