Changes between Version 6 and Version 7 of AnnPerformance

Timestamp:

05/13/08 05:58:26 (15 years ago)

Author:

hartono

Comment:

--

AnnPerformance

@@ -18 +18 @@
 }}}
 
+=== Optimizations Used ===
+In this experiment, we used PLuTo (an auto-parallelization and locality optimization tool based on polyhedral modelling) as a polyhedral-based code transformator. And then, we also extended ancc with additional modules used to perform syntactical transformations. Below are the (polyhedral and syntactic) optimizations used in this experiment.
+
+Polyhedral transformations (from PLuTo):
+ * Loop tiling for L1 and L2 caches
+ * Parallelization for multicore architectures
+ * Register tiling (for rectangular iteration spaces)
+
+Syntactical transformations (from ancc modules):
+ * Register tiling (for both rectangular and no-rectangular iteration spaces)
+ * Scalar replacement (to enhance register reuse)
+ 
+It is to be noted that the register tiling approach used by PLuTo is limited to only rectangular loops. To further improve the resulting performance, we implemented our own register tiling approach as one of the ancc's transformation modules. Our register tiling approach is so general that it can handle both rectangular and non-rectangular loops.
+
 === Sequential (single core) ===
  [[Image(lu.png,nolink)]]