Presentation is loading. Please wait.

Presentation is loading. Please wait.

Key to Scalable Parallelism - Regularity and Locality

Similar presentations


Presentation on theme: "Key to Scalable Parallelism - Regularity and Locality"— Presentation transcript:

1 Key to Scalable Parallelism - Regularity and Locality
GPU Computing Forum

2 Eight Algorithm Optimizations Techniques (so far)
Scatter to Gather transformation Privatization Work granularity coarsening Data tiling/reuse Data layout and traversal ordering Input data binning Input compaction Input extraction and regularization Currently a graduate-level practical algorithm course GPU Computing Forum

3 “Orthogonal” to Traditional Parallel Algorithms for Teaching
Tiling Privatization Regularization Compaction Binning Data Layout Granularity Coarsening Scatter to Gather MRI- Gridding CutCP Histo Stencil LBM BFS DMM MRI-Q SpMV SAD Tpacf FFT GPU Computing Forum


Download ppt "Key to Scalable Parallelism - Regularity and Locality"

Similar presentations


Ads by Google