- Sep 13, 2012
-
-
Andreas Klöckner authored
-
- Sep 11, 2012
-
-
Andreas Klöckner authored
-
- Sep 02, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Aug 31, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
- This (plus a few more fixes) restore race detection. - Also get rid of sequential_inames--realize that by a new tag, force-sequential. - Also move loop_nest_map to the scheduler file, because it tweaks its results to suit scheduler internals.
-
- Aug 30, 2012
-
-
Andreas Klöckner authored
-
- Aug 27, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Aug 26, 2012
-
-
Andreas Klöckner authored
-
- Aug 25, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Aug 24, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Aug 23, 2012
-
-
Andreas Klöckner authored
-
- Aug 13, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Aug 12, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Aug 09, 2012
-
-
Andreas Klöckner authored
-
- Jul 15, 2012
-
-
Andreas Klöckner authored
-
- May 20, 2012
-
-
Andreas Klöckner authored
-
- Apr 19, 2012
-
-
Andreas Klöckner authored
- Remove dtype state from reduction. - Introduce preamble_generators. - Introduce function_manglers, to replace result_dtype_getters - Introduce defines.
-
- Apr 18, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Mar 26, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
- Mar 03, 2012
-
-
Andreas Klöckner authored
- Rename apply_subst to expand_subst. - Track if precomputes in other substitution rules are uniformly in-footprint, and if so, perform substitution to storage reference even within substitution rules. - Remove 'formal argument' sweep spec from precompute(). This is purely counterintuitive, because it finds an intermediate set of inames that then get swept out and thereby may pick up much more than the formal argument referenced. - Fix the 'in-footprint' criterion to be the storage-to-sweep map with the sweep dimensions projected out.
-
- Mar 02, 2012
-
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-
Andreas Klöckner authored
-