Content deleted Content added
No edit summary |
|||
Line 27:
=== {{anchor|STRIDED}} Strided ===
[[Strided|Stride (computing)]] or simple 2D,3D access patterns (e.g. stepping through [[multi-dimensional array]]s) are similarly easy to predict, and are found in implementations of [[linear algebra]] algorithms and [[image processing]]. [[Loop tiling]] is an effective approach.
<ref>{{cite web|title=optimizing for tiling and data locality|url=http://www.cs.utexas.edu/users/mckinley/papers/par-mem-ics-92.pdf}}paper covers loop tiling and implication for parallel code</ref> Some systems with [[DMA]] provided a strided mode for transferring data between subtile of larger 2D arrays and scratchpad memory.<ref>{{cite web|title=Optimal 2D Data Partitioning for DMA Transfers on MPSoCs|url=http://www-verimag.imag.fr/~maler/Papers/dma2dim.pdf}}</ref>
=== {{anchor|LINEAR}} Linear ===
| |||