Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287365004> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4287365004 abstract "There are several factorizations of multi-dimensional tensors into lower-dimensional components, known as `tensor networks'. We consider the popular `tensor-train' (TT) format and ask: How efficiently can we compute a low-rank approximation from a full tensor on current multi-core CPUs? Compared to sparse and dense linear algebra, kernel libraries for multi-linear algebra are rare and typically not as well optimized. Linear algebra libraries like BLAS and LAPACK may provide the required operations in principle, but often at the cost of additional data movements for rearranging memory layouts. Furthermore, these libraries are typically optimized for the compute-bound case (e.g. square matrix operations) whereas low-rank tensor decompositions lead to memory bandwidth limited operations. We propose a `tensor-train singular value decomposition' (TT-SVD) algorithm based on two building blocks: a `Q-less tall-skinny QR' factorization, and a fused tall-skinny matrix-matrix multiplication and reshape operation. We analyze the performance of the resulting TT-SVD algorithm using the Roofline performance model. In addition, we present performance results for different algorithmic variants for shared-memory as well as distributed-memory architectures. Our experiments show that commonly used TT-SVD implementations suffer severe performance penalties. We conclude that a dedicated library for tensor factorization kernels would benefit the community: Computing a low-rank approximation can be as cheap as reading the data twice from main memory. As a consequence, an implementation that achieves realistic performance will move the limit at which one has to resort to randomized methods that only process part of the data." @default.
- W4287365004 created "2022-07-25" @default.
- W4287365004 creator A5010377968 @default.
- W4287365004 creator A5048878210 @default.
- W4287365004 creator A5070137241 @default.
- W4287365004 date "2021-01-29" @default.
- W4287365004 modified "2023-09-28" @default.
- W4287365004 title "Performance of the low-rank tensor-train SVD (TT-SVD) for large dense tensors on modern multi-core CPUs" @default.
- W4287365004 doi "https://doi.org/10.48550/arxiv.2102.00104" @default.
- W4287365004 hasPublicationYear "2021" @default.
- W4287365004 type Work @default.
- W4287365004 citedByCount "0" @default.
- W4287365004 crossrefType "posted-content" @default.
- W4287365004 hasAuthorship W4287365004A5010377968 @default.
- W4287365004 hasAuthorship W4287365004A5048878210 @default.
- W4287365004 hasAuthorship W4287365004A5070137241 @default.
- W4287365004 hasBestOaLocation W42873650041 @default.
- W4287365004 hasConcept C106487976 @default.
- W4287365004 hasConcept C11413529 @default.
- W4287365004 hasConcept C114614502 @default.
- W4287365004 hasConcept C118615104 @default.
- W4287365004 hasConcept C121332964 @default.
- W4287365004 hasConcept C139352143 @default.
- W4287365004 hasConcept C155281189 @default.
- W4287365004 hasConcept C158693339 @default.
- W4287365004 hasConcept C159985019 @default.
- W4287365004 hasConcept C164226766 @default.
- W4287365004 hasConcept C17349429 @default.
- W4287365004 hasConcept C173608175 @default.
- W4287365004 hasConcept C187834632 @default.
- W4287365004 hasConcept C188045654 @default.
- W4287365004 hasConcept C192562407 @default.
- W4287365004 hasConcept C202444582 @default.
- W4287365004 hasConcept C22789450 @default.
- W4287365004 hasConcept C2524010 @default.
- W4287365004 hasConcept C33923547 @default.
- W4287365004 hasConcept C41008148 @default.
- W4287365004 hasConcept C42355184 @default.
- W4287365004 hasConcept C62520636 @default.
- W4287365004 hasConcept C74193536 @default.
- W4287365004 hasConcept C84114770 @default.
- W4287365004 hasConceptScore W4287365004C106487976 @default.
- W4287365004 hasConceptScore W4287365004C11413529 @default.
- W4287365004 hasConceptScore W4287365004C114614502 @default.
- W4287365004 hasConceptScore W4287365004C118615104 @default.
- W4287365004 hasConceptScore W4287365004C121332964 @default.
- W4287365004 hasConceptScore W4287365004C139352143 @default.
- W4287365004 hasConceptScore W4287365004C155281189 @default.
- W4287365004 hasConceptScore W4287365004C158693339 @default.
- W4287365004 hasConceptScore W4287365004C159985019 @default.
- W4287365004 hasConceptScore W4287365004C164226766 @default.
- W4287365004 hasConceptScore W4287365004C17349429 @default.
- W4287365004 hasConceptScore W4287365004C173608175 @default.
- W4287365004 hasConceptScore W4287365004C187834632 @default.
- W4287365004 hasConceptScore W4287365004C188045654 @default.
- W4287365004 hasConceptScore W4287365004C192562407 @default.
- W4287365004 hasConceptScore W4287365004C202444582 @default.
- W4287365004 hasConceptScore W4287365004C22789450 @default.
- W4287365004 hasConceptScore W4287365004C2524010 @default.
- W4287365004 hasConceptScore W4287365004C33923547 @default.
- W4287365004 hasConceptScore W4287365004C41008148 @default.
- W4287365004 hasConceptScore W4287365004C42355184 @default.
- W4287365004 hasConceptScore W4287365004C62520636 @default.
- W4287365004 hasConceptScore W4287365004C74193536 @default.
- W4287365004 hasConceptScore W4287365004C84114770 @default.
- W4287365004 hasLocation W42873650041 @default.
- W4287365004 hasOpenAccess W4287365004 @default.
- W4287365004 hasPrimaryLocation W42873650041 @default.
- W4287365004 hasRelatedWork W121369467 @default.
- W4287365004 hasRelatedWork W2903666957 @default.
- W4287365004 hasRelatedWork W2963484322 @default.
- W4287365004 hasRelatedWork W3046452920 @default.
- W4287365004 hasRelatedWork W3118611730 @default.
- W4287365004 hasRelatedWork W3127237338 @default.
- W4287365004 hasRelatedWork W3210041882 @default.
- W4287365004 hasRelatedWork W4221140916 @default.
- W4287365004 hasRelatedWork W4287365004 @default.
- W4287365004 hasRelatedWork W210173153 @default.
- W4287365004 isParatext "false" @default.
- W4287365004 isRetracted "false" @default.
- W4287365004 workType "article" @default.