Optimal Data Partitioning Shape for Matrix Multiplication on Three Fully Connected Heterogeneous Processors.
Ashley M. DeFlumereAlexey L. LastovetskyPublished in: Euro-Par Workshops (1) (2014)
Keyphrases
- data partitioning
- distributed memory
- fully connected
- matrix multiplication
- parallel execution
- highly scalable
- shared memory
- hierarchical clustering
- parallel implementation
- optimal solution
- scale free
- parallel algorithm
- neural network
- message passing
- query execution
- similarity search
- feature extraction
- higher order
- collaborative filtering
- dynamic programming
- query processing
- database systems