On the parallel I/O optimality of linear algebra kernels: near-optimal LU factorization.
Grzegorz KwasniewskiTal Ben-NunAlexandros Nikolaos ZiogasTimo SchneiderMaciej BestaTorsten HoeflerPublished in: PPoPP (2021)
Keyphrases
- linear algebra
- singular value decomposition
- sparse matrices
- tensor factorization
- condition number
- feature space
- optimal solution
- low rank
- computer architecture
- image processing
- least squares
- principal component analysis
- parallel computers
- dimensionality reduction
- query processing
- texture features
- matrix factorization
- information systems