Home My Page Projects Code Snippets Project Openings diderot

# SCM Repository

[diderot] Diff of /branches/tash/TODO
 [diderot] / branches / tash / TODO

# Diff of /branches/tash/TODO

trunk/TODO revision 1120, Thu May 5 08:18:26 2011 UTC branches/vis12/TODO revision 1685, Sun Jan 22 15:23:36 2012 UTC
# Line 1  Line 1
1  NOTE: GLK's approximate ranking of 5 most important tagged with  ***************************************************
2  [GLK:1], [GLK:2], ...  ***************************************************
3    THIS TODO HAS BEEN MOVED TO THE DIDEROT WIKI:
4
5  ==============================  http://diderot-wiki.cs.uchicago.edu/index.php/Todo
other SHORT TERM =============  (including needed for LIC)
==============================
6
7  Add a clamp function, which takes three arguments; either three scalars:  PLEASE USE THAT PAGE TO UPDATE PROBLEMS AND PROGRESS
8    clamp(x, minval, maxval)  = max(minval, min(maxval, x))  ***************************************************
9  or three vectors of the same size:  ***************************************************
clamp([x,y], minvec, maxvec)  = [max(minvec[0], min(maxvec[0], x)),
max(minvec[1], min(maxvec[1], y))]
This would be useful in many current Diderot programs.
One question: clamp(x, minval, maxval) is the argument order
used in OpenCL and other places, but clamp(minval, maxval, x)
would be more consistent with lerp(minout, maxout, x).
10
11  Level of differentiability in field type should be statement about how  NOTE: GLK's approximate ranking of 8 most important tagged with
12  much differentiation the program *needs*, rather than what the kernel  [GLK:1], [GLK:2], ...
13  *provides*.  The needed differentiability can be less than or equal to
14  the provided differentiability.  ========================
15    SHORT TERM ============= (*needed* for streamlines & tractography)
16    ========================
17
18  [GLK:1] Add sequence types (needed for evals & evecs)  [GLK:2] Add sequence types (needed for evals & evecs)
19      syntax      syntax
20          types: ty '{' INT '}'          types: ty '{' INT '}'
21          value construction: '{' e1 ',' … ',' en '}'          value construction: '{' e1 ',' … ',' en '}'
22          indexing: e '{' e '}'          indexing: e '{' e '}'
23
24  IL support for higher-order tensor values (matrices, etc).  [GLK:3] evals & evecs for symmetric tensor[2,2] and
25      tensor construction [DONE]  tensor[3,3] (requires sequences)
tensor indexing [DONE]
tensor slicing
verify that hessians work correctly [DONE]
26
27  Use ∇⊗ etc. syntax  ability to emit/track/record variables into dynamically re-sized
28      syntax [DONE]  runtime output buffer
typechecking
IL and codegen
29
30  test/uninit.diderot:  [GLK:4] tensor fields from tensor images: Initially need at least
31  documents need for better compiler error messages when output variables  convolution on tensor[2,2] and tensor[3,3] (the same component-wise
32  are not initialized; the current messages are very cryptic  convolution as for vectors).
33
34  determinant ("det") for tensor[3,3]  ========================
35    SHORT-ISH TERM ========= (to make using Diderot less annoying to
36    ========================  program in, and slow to execute)
37
38  expand trace in mid to low translation  Allow ".ddro" file extensions in addition to ".diderot"
39
40  value-numbering optimization  Be able to output values of type tensor[2,2] and tensor[3,3];
41    (currently only scalars & vectors).  Want to add some regression tests
42    based on this and currently can't
43
44  Add type aliases for color types  [GLK:1] Proper handling of stabilize method
rgb = real{3}
rgba = real{4}
45
46  ==============================  Convolution on general tensor images (order > 2)
MEDIUM TERM ================== (including needed for streamlines & tractography)
==============================
47
48  [GLK:1] evals & evecs for symmetric tensor[3,3] (requires sequences)  allow "*" to represent "modulate": per-component multiplication of
49    vectors, and vectors only (not tensors of order 2 or higher).  Once
50    sequences are implemented this should be removed: the operation is not
51    invariant WRT basis so it is not a legit vector computation.
52
53  [GLK:2] Save Diderot output to nrrd, instead of "mip.txt"  implicit type promotion of integers to reals where reals are
54    required (e.g. not exponentiation "^")
55
56    [Nick working on this] Save Diderot output to nrrd, instead of "mip.txt"
57    For grid of strands, save to similarly-shaped array    For grid of strands, save to similarly-shaped array
58    For list of strands, save to long 1-D (or 2-D for non-scalar output) list    For list of strands, save to long 1-D (or 2-D for non-scalar output) list
59    For ragged things (like tractography output), will need to save both    For ragged things (like tractography output), will need to save both
60      complete list of values, as well as list of start indices and lengths      complete list of values, as well as list of start indices and lengths
61      to index into complete list      to index into complete list
62
63  [GLK:3] Use of Teem's "hest" command-line parser for getting  [GLK:6] ability to declare a field so that probe positions are
any input variables that are not defined in the source file

[GLK:4] ability to declare a field so that probe positions are
64  *always* "inside"; with various ways of mapping the known image values  *always* "inside"; with various ways of mapping the known image values
65  to non-existant index locations.  One possible syntax emphasizes that  to non-existant index locations.  One possible syntax emphasizes that
66  there is a index mapping function that logically precedes convolution:  there is a index mapping function that logically precedes convolution:
# Line 76  Line 69
69    F = bspln3 ⊛ (img ◦ mirror)    F = bspln3 ⊛ (img ◦ mirror)
70  where "◦" or "∘" is used to indicate function composition  where "◦" or "∘" is used to indicate function composition
71
72  extend norm (|exp|) to all tensor types [DONE for vectors and matrices]  Level of differentiability in field type should be statement about how
73    much differentiation the program *needs*, rather than what the kernel
74  ability to emit/track/record variables into dynamically re-sized  *provides*.  The needed differentiability can be less than or equal to
75  runtime buffer  the provided differentiability.
76
77  Want: allow X *= Y, X /= Y, X += Y, X -= Y to mean what they do in C,  Use ∇⊗ etc. syntax
78  provided that X*Y, X/Y, X+Y, X-Y are already supported.      syntax [DONE]
79  Nearly every Diderot program would be simplified by this.      typechecking
80        IL and codegen
81
82  Want: non-trivial field expressions & functions:  Add type aliases for color types
83    image(2)[2] Vimg = load(...);      rgb = real{3}
84    field#0(2)[] Vlen = |Vimg ⊛ bspln3|;      rgba = real{4}
to get a scalar field of vector length, or
field#2(2)[] F = Fimg ⊛ bspln3;
field#0(2)[] Gmag = |∇F|;
to get a scalar field of gradient magnitude, or
field#2(2)[] F = Fimg ⊛ bspln3;
field#0(2)[] Gmsq = ∇F•∇F;
to get a scalar field of squared gradient magnitude, which is simpler
to differentiate.  However, there is value in having these, even if
the differentiation of them is not supported (hence the indication
of "field#0" for these above)

Want: ability to apply "normalize" to a field itself, e.g.
field#0(2)[2] V = normalize(Vimg ⊛ ctmr);
so that V(x) = normalize((Vimg ⊛ ctmr)(x)).
Having this would simplify expression of standard LIC method, and
would also help express other vector field expressions that arise
in vector field feature exraction.
85
86  tensor fields: convolution on general tensor images  Revisit how images are created within the language.
87    The "load" operator should probably go away, and its strange
88    that strings are there only as a way to refer to nrrd filenames
89
90  ==============================  ==============================
91  other MEDIUM TERM ============ (needed for particles)  MEDIUM TERM ================== (*needed* for particles)
92  ==============================  ==============================
93
94  [GLK:5] run-time death of strands; test/iso2d.diderot provides  [Lamont working on this] run-time birth of strands
great initial test for "die" command.

Put small 1-D and 2-D fields, reconstructed by tent, into
faster texture buffers

run-time birth of strands
95
96  "initially" supports lists  "initially" supports lists
97
98  "initially" supports lists of positions output from  "initially" supports lists of positions output from different
99  different initalization Diderot program  initalization Diderot program (or output from the same program;
100    e.g. using output of iso2d.diderot for one isovalue to seed the input
101    to another invocation of the same program)
102
103    [Lamont working on this] Communication between strands: they have to
104    be able to learn each other's state (at the previous iteration).
105    Early version of this can have the network of neighbors be completely
106    static (for running one strand/pixel image computations).  Later
107    version with strands moving through the domain will require some
108    spatial data structure to optimize discovery of neighbors.
109
110    ============================
111    MEDIUM-ISH TERM ============ (to make Diderot more useful/effective)
112    ============================
113
114    [GLK:5] Want code-generation working for tensors of order three.
115    Order three matters for edge detection in scalar fields (to get
116    second derivatives of gradient magnitude), second derivatives
117    of vector fields (for some feature extraction), and first
118    derivatives of diffusion tensor fields.
119
120  spatial data structure that permits strands' queries of neighbors  Python/ctypes interface to run-time
121
122  proper handling of stabilize method  support for Python interop and GUI
123
124  test/vr-kcomp2.diderot: Add support for code like  Allow integer exponentiation ("^2") to apply to square matrices,
125    to represent repeated matrix multiplication
126
127          (F1 if x else F2)@pos  Put small 1-D and 2-D fields, when reconstructed specifically by tent
128    and when differentiation is not needed, into faster texture buffers.
129    test/illust-vr.diderot is good example of program that uses multiple
130    such 1-D fields basically as lookup-table-based function evaluation
131
132  This will require duplication of the continuation of the conditional  extend norm (|exp|) to all tensor types [DONE for vectors and matrices]
133  (but we should only duplicate over the live-range of the result of the
134  conditional.  determinant ("det") for tensor[3,3]
135
136  add ":" for tensor dot product (contracts out two indices  add ":" for tensor dot product (contracts out two indices
137  instead of one like •), valid for all pairs of tensors with  instead of one like •), valid for all pairs of tensors with
138  at least two indices  at least two indices
139
140  ==============================  test/uninit.diderot:
141  other MEDIUM TERM ============  documents need for better compiler error messages when output variables
142  ==============================  are not initialized; the current messages are very cryptic
143
144  want: warnings when "D" (reserved for differentiation) is declared as  want: warnings when "D" (reserved for differentiation) is declared as
145  a variable name (get confusing error messages now)  a variable name (get confusing error messages now)
146
support for Python interop and GUI

Python/ctypes interface to run-time

==============================
LONG TERM ====================
147  ==============================  ==============================
148    LONG TERM ==================== (make Diderot more interesting/attractive from
149    ==============================  a research standpoint)
150
151    IL support for higher-order tensor values (matrices, etc).
152        tensor construction [DONE]
153        tensor indexing [DONE]
154        tensor slicing
155
156  Better handling of variables that determines the scope of a variable  Better handling of variables that determines the scope of a variable
157  based on its actual use, instead of where the user defined it.  So,  based on its actual use, instead of where the user defined it.  So,
# Line 162  Line 159
159  scope.  Also prune out useless variables, which should include field  scope.  Also prune out useless variables, which should include field
160  variables after the translation to mid-il.  variables after the translation to mid-il.
161
162    test/vr-kcomp2.diderot: Add support for code like
163            (F1 if x else F2)@pos
164    This will require duplication of the continuation of the conditional
165    (but we should only duplicate over the live-range of the result of the
166    conditional.
167
168    [GLK:7] Want: non-trivial field expressions & functions.
169    scalar fields from scalar fields F and G:
170      field#0(2)[] X = (sin(F) + 1.0)/2;
171      field#0(2)[] X = F*G;
172    scalar field of vector field magnitude:
174      field#0(2)[] Vlen = |Vimg ⊛ bspln3|;
175    field of normalized vectors (for LIC and vector field feature extraction)
176      field#2(2)[2] F = ...
177      field#0(2)[2] V = normalize(F);
178    scalar field of gradient magnitude (for edge detection))
179      field#2(2)[] F = Fimg ⊛ bspln3;
180      field#0(2)[] Gmag = |∇F|;
181    scalar field of squared gradient magnitude (simpler to differentiate):
182      field#2(2)[] F = Fimg ⊛ bspln3;
183      field#0(2)[] Gmsq = ∇F•∇F;
184    There is value in having these, even if the differentiation of them is
185    not supported (hence the indication of "field#0" for these above)
186
187    Introduce region types (syntax region(d), where d is the dimension of the
188    region.  One useful operator would be
189            dom : field#k(d)[s] -> region(d)
190    Then the inside test could be written as
191            pos ∈ dom(F)
192    We could further extend this approach to allow geometric definitions of
193    regions.  It might also be useful to do inside tests in world space,
195
196  co- vs contra- index distinction  co- vs contra- index distinction
197
198  some indication of tensor symmetry  Permit field composition:
199      field#2(3)[3] warp = bspln3 ⊛ warpData;
200      field#2(3)[] F = bspln3 ⊛ img;
201      field#2(3)[] Fwarp = F ◦ warp;
202    So Fwarp(x) = F(warp(X)).  Chain rule can be used for differentation.
203    This will be instrumental for expressing non-rigid registration
204    methods (but those will require co-vs-contra index distinction)
205
206    Allow the convolution to be specified either as a single 1D kernel
207    (as we have it now):
208      field#2(3)[] F = bspln3 ⊛ img;
209    or, as a tensor product of kernels, one for each axis, e.g.
210      field#0(3)[] F = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img;
211    This is especially important for things like time-varying fields
212    and the use of scale-space in field visualization: one axis of the
213    must be convolved with a different kernel during probing.
214    What is very unclear is how, in such cases, we should notate the
215    gradient, when we only want to differentiate with respect to some
216    subset of the axes.  One ambitious idea would be:
217      field#0(3)[] Ft = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img; // 2D time-varying field
218      field#0(2)[] F = lambda([x,y], Ft([x,y,42.0]))     // restriction to time=42.0
220
221    representation of tensor symmetry
222  (have to identify the group of index permutations that are symmetries)  (have to identify the group of index permutations that are symmetries)
223
224  dot works on all tensors  dot works on all tensors
225
226  outer works on all tensors  outer works on all tensors
227
228    Help for debugging Diderot programs: need to be able to uniquely
229    identify strands, and for particular strands that are known to behave
230    badly, do something like printf or other logging of their computations
232
233    Permit writing dimensionally general code: Have some statement of the
234    dimension of the world "W" (or have it be learned from one particular
235    field of interest), and then able to write "vec" instead of
236    "vec2/vec3", and perhaps "tensor[W,W]" instead of
237    "tensor[2,2]/tensor[3,3]"
238
239    Traits: all things things that have boilerplate code (especially
240    volume rendering) should be expressed in terms of the unique
241    computational core.  Different kinds of streamline/tractography
242    computation will be another example, as well as particle systems.
243
244  Einstein summation notation  Einstein summation notation
245
246  "tensor comprehension" (like list comprehension)  "tensor comprehension" (like list comprehension)
247
248    Fields coming from different sources of data:
249    * triangular or tetrahedral meshes over 2D or 3D domains (of the
250      source produced by finite-element codes; these will come with their
251      own specialized kinds of reconstruction kernels, called "basis
252      functions" in this context)
253    * Large point clouds, with some radial basis function around each point,
254      which will be tuned by parameters of the point (at least one parameter
255      giving some notion of radius)
256
257  ======================  ======================
258  BUGS =================  BUGS =================
259  ======================  ======================
260
// HEY (BUG?) shouldn't it be a type error to load this 2-D array of
// 2-vectors into a 2-D *scalar* field?  Instead, get:
//   uncaught exception Fail [Fail: Error in compiling lic.diderot]
//     raised at driver/main.sml:31.39-31.76