Optimizing memory bandwidth use and performance for matrix-vector multiplication in iterative methods.
David BolandGeorge A. ConstantinidesPublished in: ACM Trans. Reconfigurable Technol. Syst. (2011)
Keyphrases
- iterative methods
- floating point
- memory bandwidth
- image reconstruction from projections
- constrained optimization
- computationally expensive
- iterative algorithms
- level parallelism
- machine learning
- feature vectors
- processing power
- learning algorithm
- artificial intelligence
- database systems
- cloud computing
- memory access