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

SCM Repository

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

Diff of /branches/tash/TODO

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

revision 1140, Fri May 6 14:16:00 2011 UTC revision 1165, Mon May 9 22:02:04 2011 UTC
# Line 1  Line 1 
1  NOTE: GLK's approximate ranking of 5 most important tagged with  NOTE: GLK's approximate ranking of 8 most important tagged with
2  [GLK:1], [GLK:2], ...  [GLK:1], [GLK:2], ...
3    
4  ==============================  ========================
5  other SHORT TERM =============  (including needed for LIC)  SHORT TERM ============= (*needed* for streamlines & tractography)
6  ==============================  ========================
   
 Add a clamp function, which takes three arguments; either three scalars:  
   clamp(x, minval, maxval)  = max(minval, min(maxval, x))  
 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).  
7    
8  Level of differentiability in field type should be statement about how  [GLK:3] Add sequence types (needed for evals & evecs)
 much differentiation the program *needs*, rather than what the kernel  
 *provides*.  The needed differentiability can be less than or equal to  
 the provided differentiability.  
   
 [GLK:1] Add sequence types (needed for evals & evecs)  
9      syntax      syntax
10          types: ty '{' INT '}'          types: ty '{' INT '}'
11          value construction: '{' e1 ',' … ',' en '}'          value construction: '{' e1 ',' … ',' en '}'
12          indexing: e '{' e '}'          indexing: e '{' e '}'
13    
14  IL support for higher-order tensor values (matrices, etc).  [GLK:4] evals & evecs for symmetric tensor[2,2] and
15      tensor construction [DONE]  tensor[3,3] (requires sequences)
     tensor indexing [DONE]  
     tensor slicing  
     verify that hessians work correctly [DONE]  
16    
17  Use ∇⊗ etc. syntax  ability to emit/track/record variables into dynamically re-sized
18      syntax [DONE]  runtime buffer
     typechecking  
     IL and codegen  
   
 test/uninit.diderot:  
 documents need for better compiler error messages when output variables  
 are not initialized; the current messages are very cryptic  
   
 determinant ("det") for tensor[3,3]  
19    
20  expand trace in mid to low translation  tensor fields: convolution on general tensor images
21    
22  value-numbering optimization  ========================
23    SHORT-ISH TERM ========= (to make using Diderot less annoying to
24    ========================  program in, and slow to execute)
25    
26    value-numbering optimization [DONE, but needs more testing]
27    
28    [GLK:1] Add a clamp function, which takes three arguments; either
29    three scalars:
30      clamp(lo, hi, x)  = max(lo, min(hi, x))
31    or three vectors of the same size:
32      clamp(lo, hi, [x,y])  = [max(lo[0], min(hi[0], x)),
33                               max(lo[1], min(hi[1], y))]
34    This would be useful in many current Diderot programs.
35    One question: clamp(x, lo, hi) is the argument order used in OpenCL
36    and other places, but clamp(lo, hi, x) is much more consistent with
37    lerp(lo, hi, x), hence GLK's preference
38    
39  Add type aliases for color types  [GLK:2] Proper handling of stabilize method
     rgb = real{3}  
     rgba = real{4}  
40    
41  ==============================  allow "*" to represent "modulate": per-component multiplication of
42  MEDIUM TERM ================== (including needed for streamlines & tractography)  vectors, and vectors only (not tensors of order 2 or higher).  Once
43  ==============================  sequences are implemented this should be removed: the operation is not
44    invariant WRT basis so it is not a legit vector computation.
45    
46  [GLK:1] evals & evecs for symmetric tensor[3,3] (requires sequences)  implicit type promotion of integers to reals where reals are
47    required (e.g. not exponentiation "^")
48    
49  [GLK:2] Save Diderot output to nrrd, instead of "mip.txt"  [GLK:5] Save Diderot output to nrrd, instead of "mip.txt"
50    For grid of strands, save to similarly-shaped array    For grid of strands, save to similarly-shaped array
51    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
52    For ragged things (like tractography output), will need to save both    For ragged things (like tractography output), will need to save both
53      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
54      to index into complete list      to index into complete list
55    
56  [GLK:3] Use of Teem's "hest" command-line parser for getting  [GLK:6] Use of Teem's "hest" command-line parser for getting
57  any input variables that are not defined in the source file  any input variables that are not defined in the source file
58    
59  [GLK:4] ability to declare a field so that probe positions are  [GLK:7] ability to declare a field so that probe positions are
60  *always* "inside"; with various ways of mapping the known image values  *always* "inside"; with various ways of mapping the known image values
61  to non-existant index locations.  One possible syntax emphasizes that  to non-existant index locations.  One possible syntax emphasizes that
62  there is a index mapping function that logically precedes convolution:  there is a index mapping function that logically precedes convolution:
# Line 76  Line 65 
65    F = bspln3 ⊛ (img ◦ mirror)    F = bspln3 ⊛ (img ◦ mirror)
66  where "◦" or "∘" is used to indicate function composition  where "◦" or "∘" is used to indicate function composition
67    
68  extend norm (|exp|) to all tensor types [DONE for vectors and matrices]  Level of differentiability in field type should be statement about how
69    much differentiation the program *needs*, rather than what the kernel
70  ability to emit/track/record variables into dynamically re-sized  *provides*.  The needed differentiability can be less than or equal to
71  runtime buffer  the provided differentiability.
   
 Want: allow X *= Y, X /= Y, X += Y, X -= Y to mean what they do in C,  
 provided that X*Y, X/Y, X+Y, X-Y are already supported.  
 Nearly every Diderot program would be simplified by this.  
72    
73  Want: non-trivial field expressions & functions:  Use ∇⊗ etc. syntax
74    image(2)[2] Vimg = load(...);      syntax [DONE]
75    field#0(2)[] Vlen = |Vimg ⊛ bspln3|;      typechecking
76  to get a scalar field of vector length, or      IL and codegen
   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.  
77    
78  tensor fields: convolution on general tensor images  Add type aliases for color types
79        rgb = real{3}
80        rgba = real{4}
81    
82  ==============================  ==============================
83  other MEDIUM TERM ============ (needed for particles)  MEDIUM TERM ================== (*needed* for particles)
84  ==============================  ==============================
85    
 Put small 1-D and 2-D fields, when reconstructed specifically by tent  
 and when differentiation is not needed, into faster texture buffers.  
 test/illust-vr.diderot is good example of program that uses multiple  
 such 1-D fields basically as lookup-table-based function evaluation  
   
86  run-time birth of strands  run-time birth of strands
87    
88  "initially" supports lists  "initially" supports lists
# Line 124  Line 90 
90  "initially" supports lists of positions output from  "initially" supports lists of positions output from
91  different initalization Diderot program  different initalization Diderot program
92    
93  spatial data structure that permits strands' queries of neighbors  Communication between strands: they have to be able to learn each
94    other's state (at the previous iteration).  Early version of this can
95    have the network of neighbors be completely static (for running one
96    strand/pixel image computations).  Later version with strands moving
97    through the domain will require some spatial data structure to
98    optimize discovery of neighbors.
99    
100    ============================
101    MEDIUM-ISH TERM ============ (to make Diderot more useful/effective)
102    ============================
103    
104  proper handling of stabilize method  Python/ctypes interface to run-time
105    
106  test/vr-kcomp2.diderot: Add support for code like  support for Python interop and GUI
107    
108          (F1 if x else F2)@pos  Allow integer exponentiation ("^2") to apply to square matrices,
109    to represent repeated matrix multiplication
110    
111  This will require duplication of the continuation of the conditional  Alow X *= Y, X /= Y, X += Y, X -= Y to mean what they do in C,
112  (but we should only duplicate over the live-range of the result of the  provided that X*Y, X/Y, X+Y, X-Y are already supported.
113  conditional.  Nearly every Diderot program would be simplified by this.
114    
115    Put small 1-D and 2-D fields, when reconstructed specifically by tent
116    and when differentiation is not needed, into faster texture buffers.
117    test/illust-vr.diderot is good example of program that uses multiple
118    such 1-D fields basically as lookup-table-based function evaluation
119    
120    expand trace in mid to low translation
121    
122    extend norm (|exp|) to all tensor types [DONE for vectors and matrices]
123    
124    determinant ("det") for tensor[3,3]
125    
126  add ":" for tensor dot product (contracts out two indices  add ":" for tensor dot product (contracts out two indices
127  instead of one like •), valid for all pairs of tensors with  instead of one like •), valid for all pairs of tensors with
128  at least two indices  at least two indices
129    
130  ==============================  test/uninit.diderot:
131  other MEDIUM TERM ============  documents need for better compiler error messages when output variables
132  ==============================  are not initialized; the current messages are very cryptic
133    
134  want: warnings when "D" (reserved for differentiation) is declared as  want: warnings when "D" (reserved for differentiation) is declared as
135  a variable name (get confusing error messages now)  a variable name (get confusing error messages now)
136    
 support for Python interop and GUI  
   
 Python/ctypes interface to run-time  
   
 ==============================  
 LONG TERM ====================  
137  ==============================  ==============================
138    LONG TERM ==================== (make Diderot more interesting/attractive from
139    ==============================  a research standpoint)
140    
141    IL support for higher-order tensor values (matrices, etc).
142        tensor construction [DONE]
143        tensor indexing [DONE]
144        tensor slicing
145        verify that hessians work correctly [DONE]
146    
147  Better handling of variables that determines the scope of a variable  Better handling of variables that determines the scope of a variable
148  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 161  Line 150 
150  scope.  Also prune out useless variables, which should include field  scope.  Also prune out useless variables, which should include field
151  variables after the translation to mid-il.  variables after the translation to mid-il.
152    
153    test/vr-kcomp2.diderot: Add support for code like
154            (F1 if x else F2)@pos
155    This will require duplication of the continuation of the conditional
156    (but we should only duplicate over the live-range of the result of the
157    conditional.
158    
159    [GLK:8] Want: non-trivial field expressions & functions.
160    scalar fields from scalar fields F and G:
161      field#0(2)[] X = (sin(F) + 1.0)/2;
162      field#0(2)[] X = F*G;
163    scalar field of vector field magnitude:
164      image(2)[2] Vimg = load(...);
165      field#0(2)[] Vlen = |Vimg ⊛ bspln3|;
166    field of normalized vectors (for LIC and vector field feature extraction)
167      field#2(2)[2] F = ...
168      field#0(2)[2] V = normalize(F);
169    scalar field of gradient magnitude (for edge detection))
170      field#2(2)[] F = Fimg ⊛ bspln3;
171      field#0(2)[] Gmag = |∇F|;
172    scalar field of squared gradient magnitude (simpler to differentiate):
173      field#2(2)[] F = Fimg ⊛ bspln3;
174      field#0(2)[] Gmsq = ∇F•∇F;
175    There is value in having these, even if the differentiation of them is
176    not supported (hence the indication of "field#0" for these above)
177    
178  co- vs contra- index distinction  co- vs contra- index distinction
179    
180  some indication of tensor symmetry  Permit field composition:
181      field#2(3)[3] warp = bspln3 ⊛ warpData;
182      field#2(3)[] F = bspln3 ⊛ img;
183      field#2(3)[] Fwarp = F ◦ warp;
184    So Fwarp(x) = F(warp(X)).  Chain rule can be used for differentation.
185    This will be instrumental for expressing non-rigid registration
186    methods (but those will require co-vs-contra index distinction)
187    
188    Allow the convolution to be specified either as a single 1D kernel
189    (as we have it now):
190      field#2(3)[] F = bspln3 ⊛ img;
191    or, as a tensor product of kernels, one for each axis, e.g.
192      field#0(3)[] F = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img;
193    This is especially important for things like time-varying data, or
194    other multi-dimensional fields where one axis of the domain is very
195    different from the rest, and hence must be treated separately when
196    it comes to convolution.  What is very unclear is how, in such cases,
197    we should notate the gradient, when we only want to differentiate with
198    respect to some subset of the axes.  One ambitious idea would be:
199      field#0(3)[] Ft = (bspln3 ⊗ bspln3 ⊗ tent) ⊛ img; // 2D time-varying field
200      field#0(2)[] F = lambda([x,y], Ft([x,y,42.0]))    // restriction to time=42.0
201      vec2 grad = ∇F([x,y]);                            // 2D gradient
202    
203    representation of tensor symmetry
204  (have to identify the group of index permutations that are symmetries)  (have to identify the group of index permutations that are symmetries)
205    
206  dot works on all tensors  dot works on all tensors

Legend:
Removed from v.1140  
changed lines
  Added in v.1165

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