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

SCM Repository

[diderot] Annotation of /branches/opencl/TODO
ViewVC logotype

Annotation of /branches/opencl/TODO

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1389 - (view) (download)
Original Path: trunk/TODO

1 : glk 1162 NOTE: GLK's approximate ranking of 8 most important tagged with
2 : jhr 1115 [GLK:1], [GLK:2], ...
3 :    
4 : glk 1156 ========================
5 :     SHORT TERM ============= (*needed* for streamlines & tractography)
6 :     ========================
7 : jhr 1115
8 : glk 1338 [GLK:2] Add sequence types (needed for evals & evecs)
9 : jhr 1115 syntax
10 :     types: ty '{' INT '}'
11 :     value construction: '{' e1 ',' … ',' en '}'
12 :     indexing: e '{' e '}'
13 :    
14 : glk 1338 [GLK:3] evals & evecs for symmetric tensor[2,2] and
15 : glk 1162 tensor[3,3] (requires sequences)
16 :    
17 : glk 1156 ability to emit/track/record variables into dynamically re-sized
18 : glk 1389 runtime output buffer
19 : jhr 1115
20 : jhr 1388 tensor fields: convolution on general tensor images (order > 1)
21 : jhr 1115
22 : glk 1156 ========================
23 : glk 1162 SHORT-ISH TERM ========= (to make using Diderot less annoying to
24 :     ======================== program in, and slow to execute)
25 : jhr 1115
26 : glk 1167 Allow ".ddro" file extensions in addition to ".diderot"
27 :    
28 : glk 1204 Be able to output values of type tensor[2,2] and tensor[3,3];
29 :     (currently only scalars & vectors). Want to add some regression tests
30 :     based on this and currently can't
31 : glk 1167
32 : glk 1338 [GLK:1] Proper handling of stabilize method
33 : jhr 1115
34 : glk 1162 allow "*" to represent "modulate": per-component multiplication of
35 :     vectors, and vectors only (not tensors of order 2 or higher). Once
36 :     sequences are implemented this should be removed: the operation is not
37 :     invariant WRT basis so it is not a legit vector computation.
38 :    
39 :     implicit type promotion of integers to reals where reals are
40 :     required (e.g. not exponentiation "^")
41 :    
42 : glk 1338 [GLK:4] Save Diderot output to nrrd, instead of "mip.txt"
43 : jhr 1115 For grid of strands, save to similarly-shaped array
44 :     For list of strands, save to long 1-D (or 2-D for non-scalar output) list
45 :     For ragged things (like tractography output), will need to save both
46 :     complete list of values, as well as list of start indices and lengths
47 :     to index into complete list
48 :    
49 : glk 1389 [GLK:5] ability to declare a field so that probe positions are
50 : glk 1120 *always* "inside"; with various ways of mapping the known image values
51 :     to non-existant index locations. One possible syntax emphasizes that
52 :     there is a index mapping function that logically precedes convolution:
53 : glk 1162 F = bspln3 ⊛ (img ◦ clamp)
54 : glk 1120 F = bspln3 ⊛ (img ◦ repeat)
55 :     F = bspln3 ⊛ (img ◦ mirror)
56 :     where "◦" or "∘" is used to indicate function composition
57 : jhr 1115
58 : glk 1162 Level of differentiability in field type should be statement about how
59 :     much differentiation the program *needs*, rather than what the kernel
60 :     *provides*. The needed differentiability can be less than or equal to
61 :     the provided differentiability.
62 :    
63 : glk 1156 Use ∇⊗ etc. syntax
64 :     syntax [DONE]
65 :     typechecking
66 :     IL and codegen
67 : jhr 1115
68 : glk 1156 Add type aliases for color types
69 :     rgb = real{3}
70 :     rgba = real{4}
71 : jhr 1115
72 : glk 1389 Revisit how images are created within the language.
73 :     The "load" operator should probably go away, and its strangs
74 :     that strings are there only as a way to refer to nrrd filenames
75 :    
76 : jhr 1115 ==============================
77 : glk 1156 MEDIUM TERM ================== (*needed* for particles)
78 : jhr 1115 ==============================
79 :    
80 :     run-time birth of strands
81 :    
82 :     "initially" supports lists
83 :    
84 : glk 1254 "initially" supports lists of positions output from different
85 :     initalization Diderot program (or output from the same program;
86 :     e.g. using output of iso2d.diderot for one isovalue to seed the input
87 :     to another invocation of the same program)
88 : jhr 1115
89 : glk 1156 Communication between strands: they have to be able to learn each
90 :     other's state (at the previous iteration). Early version of this can
91 :     have the network of neighbors be completely static (for running one
92 :     strand/pixel image computations). Later version with strands moving
93 :     through the domain will require some spatial data structure to
94 :     optimize discovery of neighbors.
95 : jhr 1115
96 : glk 1156 ============================
97 :     MEDIUM-ISH TERM ============ (to make Diderot more useful/effective)
98 :     ============================
99 : jhr 1115
100 : glk 1156 Python/ctypes interface to run-time
101 : jhr 1115
102 : glk 1156 support for Python interop and GUI
103 : jhr 1115
104 : glk 1162 Allow integer exponentiation ("^2") to apply to square matrices,
105 :     to represent repeated matrix multiplication
106 :    
107 : glk 1156 Put small 1-D and 2-D fields, when reconstructed specifically by tent
108 :     and when differentiation is not needed, into faster texture buffers.
109 :     test/illust-vr.diderot is good example of program that uses multiple
110 :     such 1-D fields basically as lookup-table-based function evaluation
111 :    
112 :     extend norm (|exp|) to all tensor types [DONE for vectors and matrices]
113 :    
114 :     determinant ("det") for tensor[3,3]
115 :    
116 : jhr 1115 add ":" for tensor dot product (contracts out two indices
117 :     instead of one like •), valid for all pairs of tensors with
118 :     at least two indices
119 :    
120 : glk 1156 test/uninit.diderot:
121 :     documents need for better compiler error messages when output variables
122 :     are not initialized; the current messages are very cryptic
123 : jhr 1115
124 :     want: warnings when "D" (reserved for differentiation) is declared as
125 :     a variable name (get confusing error messages now)
126 :    
127 : glk 1156 ==============================
128 :     LONG TERM ==================== (make Diderot more interesting/attractive from
129 :     ============================== a research standpoint)
130 : jhr 1115
131 : glk 1389 [GLK:6] Want code-generation working for tensors of order three.
132 :     Order three matters for edge detection in scalar fields (to get
133 :     second derivatives of gradient magnitude), second derivatives
134 :     of vector fields (for some feature extraction), and first
135 :     derivatives of diffusion tensor fields.
136 :    
137 : glk 1156 IL support for higher-order tensor values (matrices, etc).
138 :     tensor construction [DONE]
139 :     tensor indexing [DONE]
140 :     tensor slicing
141 : jhr 1115
142 : glk 1156 Better handling of variables that determines the scope of a variable
143 :     based on its actual use, instead of where the user defined it. So,
144 :     for example, we should lift strand-invariant variables to global
145 :     scope. Also prune out useless variables, which should include field
146 :     variables after the translation to mid-il.
147 :    
148 :     test/vr-kcomp2.diderot: Add support for code like
149 :     (F1 if x else F2)@pos
150 :     This will require duplication of the continuation of the conditional
151 :     (but we should only duplicate over the live-range of the result of the
152 :     conditional.
153 :    
154 : glk 1338 [GLK:7] Want: non-trivial field expressions & functions.
155 : glk 1162 scalar fields from scalar fields F and G:
156 :     field#0(2)[] X = (sin(F) + 1.0)/2;
157 :     field#0(2)[] X = F*G;
158 :     scalar field of vector field magnitude:
159 : glk 1156 image(2)[2] Vimg = load(...);
160 :     field#0(2)[] Vlen = |Vimg ⊛ bspln3|;
161 : glk 1162 field of normalized vectors (for LIC and vector field feature extraction)
162 :     field#2(2)[2] F = ...
163 :     field#0(2)[2] V = normalize(F);
164 :     scalar field of gradient magnitude (for edge detection))
165 : glk 1156 field#2(2)[] F = Fimg ⊛ bspln3;
166 :     field#0(2)[] Gmag = |∇F|;
167 : glk 1162 scalar field of squared gradient magnitude (simpler to differentiate):
168 : glk 1156 field#2(2)[] F = Fimg ⊛ bspln3;
169 :     field#0(2)[] Gmsq = ∇F•∇F;
170 : glk 1162 There is value in having these, even if the differentiation of them is
171 :     not supported (hence the indication of "field#0" for these above)
172 : glk 1156
173 : jhr 1195 Introduce region types (syntax region(d), where d is the dimension of the
174 :     region. One useful operator would be
175 :     dom : field#k(d)[s] -> region(d)
176 :     Then the inside test could be written as
177 :     pos ∈ dom(F)
178 :     We could further extend this approach to allow geometric definitions of
179 :     regions. It might also be useful to do inside tests in world space,
180 :     instead of image space.
181 :    
182 : glk 1162 co- vs contra- index distinction
183 : glk 1156
184 : glk 1162 Permit field composition:
185 : glk 1156 field#2(3)[3] warp = bspln3 ⊛ warpData;
186 :     field#2(3)[] F = bspln3 ⊛ img;
187 :     field#2(3)[] Fwarp = F ◦ warp;
188 : glk 1162 So Fwarp(x) = F(warp(X)). Chain rule can be used for differentation.
189 :     This will be instrumental for expressing non-rigid registration
190 :     methods (but those will require co-vs-contra index distinction)
191 : glk 1156
192 : glk 1155 Allow the convolution to be specified either as a single 1D kernel
193 :     (as we have it now):
194 :     field#2(3)[] F = bspln3 ⊛ img;
195 :     or, as a tensor product of kernels, one for each axis, e.g.
196 :     field#0(3)[] F = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img;
197 : glk 1212 This is especially important for things like time-varying fields
198 :     and the use of scale-space in field visualization: one axis of the
199 :     must be convolved with a different kernel during probing.
200 :     What is very unclear is how, in such cases, we should notate the
201 :     gradient, when we only want to differentiate with respect to some
202 :     subset of the axes. One ambitious idea would be:
203 : glk 1162 field#0(3)[] Ft = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img; // 2D time-varying field
204 : glk 1204 field#0(2)[] F = lambda([x,y], Ft([x,y,42.0])) // restriction to time=42.0
205 :     vec2 grad = ∇F([x,y]); // 2D gradient
206 : glk 1155
207 : glk 1204 Tensors of order 3 (e.g. gradients of diffusion tensor fields, or
208 :     hessians of vector fields) and order 4 (e.g. Hessians of diffusion
209 :     tensor fields).
210 :    
211 : glk 1162 representation of tensor symmetry
212 : jhr 1115 (have to identify the group of index permutations that are symmetries)
213 :    
214 :     dot works on all tensors
215 :    
216 :     outer works on all tensors
217 :    
218 : glk 1204 Help for debugging Diderot programs: need to be able to uniquely
219 :     identify strands, and for particular strands that are known to behave
220 :     badly, do something like printf or other logging of their computations
221 :     and updates.
222 :    
223 :     Permit writing dimensionally general code: Have some statement of the
224 :     dimension of the world "W" (or have it be learned from one particular
225 :     field of interest), and then able to write "vec" instead of
226 :     "vec2/vec3", and perhaps "tensor[W,W]" instead of
227 :     "tensor[2,2]/tensor[3,3]"
228 :    
229 :     Traits: all things things that have boilerplate code (especially
230 :     volume rendering) should be expressed in terms of the unique
231 :     computational core. Different kinds of streamline/tractography
232 :     computation will be another example, as well as particle systems.
233 :    
234 : jhr 1115 Einstein summation notation
235 :    
236 :     "tensor comprehension" (like list comprehension)
237 :    
238 : glk 1204 Fields coming from different sources of data:
239 :     * triangular or tetrahedral meshes over 2D or 3D domains (of the
240 :     source produced by finite-element codes; these will come with their
241 :     own specialized kinds of reconstruction kernels, called "basis
242 :     functions" in this context)
243 :     * Large point clouds, with some radial basis function around each point,
244 :     which will be tuned by parameters of the point (at least one parameter
245 :     giving some notion of radius)
246 :    
247 : jhr 1115 ======================
248 :     BUGS =================
249 :     ======================
250 :    
251 :     test/zslice2.diderot:
252 :     // HEY (bug) bspln5 leads to problems ...
253 :     // uncaught exception Size [size]
254 :     // raised at c-target/c-target.sml:47.15-47.19
255 :     //field#4(3)[] F = img ⊛ bspln5;
256 : glk 1336

root@smlnj-gforge.cs.uchicago.edu
ViewVC Help
Powered by ViewVC 1.0.0