Changes between Version 18 and Version 19 of DataParallel/Replicate

Timestamp:

08/07/11 06:51:05 (22 months ago)

Author:

chak

Comment:

--

DataParallel/Replicate

@@ -22 +22 @@
         s2      = sliceP half half  xx           
       in concatP (mapP (treeLookup table) [:s1, s2:])
+      -- or better: concatP (mapP (treeLookup table) (segmentP [:half, len - half:] xx))
 }}}
 Here `table` is constant in `mapP (treeLookup table) [:s1, s2:]`; hence, the entire `table` gets duplicated on each level of the recursion, leading to space consumption that is exponential in the depth of the recursion.
@@ -90 +91 @@
 === Reduction ===
 
- TODO: rl's exmaple
+ TODO: rl's example
 
-=== Multiple levels of nesting ===
+=== Splitting and combining ===
+
+Due to lifted conditionals (or, generally, case constructs), arrays with repeated segments may be split (packed) and combined.  Arrays with repeated segments can be split (or packed) by including a repeated segment in the result exactly if one of its repetitions is included.  This can be determined by disjunctively combining all flags for one sequence of repeated segments.
+
+Arrays with repeated segments can not always be combined without duplicating the data corresponding to repeated segments (after all, a disjoint segment may be inserted into a sequence of repeated segments).  For simplicity, we may want to expand all repeated segments in `combinePA`.  (It seems that this should not lead to unexpected blow-ups as the repeated data now is part of a functions result and should be accounted for in its complexity estimate.)
+
+=== Multiple levels of nesting (unconcat and concat) ===
 
 TODO: What if we have segment descriptors on top of one with repeated segments?