Using Butterfly-Patterned Partial Sums to Optimize GPU Memory Accesses for Drawing from Discrete Distributions.
Guy L. Steele Jr.Jean-Baptiste TristanPublished in: CoRR (2015)
Keyphrases
- random variables
- efficient algorithms to compute
- real time
- probability distribution
- memory requirements
- memory usage
- intel xeon
- difference equations
- memory space
- neural network
- memory access
- parallel computing
- data transfer
- continuous variables
- discrete geometry
- computational power
- graphics processors
- memory bandwidth
- access patterns
- computing power
- multi threaded