Matches in SemOpenAlex for { <https://semopenalex.org/work/W2617031569> ?p ?o ?g. }
Showing items 1 to 71 of
71
with 100 items per page.
- W2617031569 abstract "Convolution is a fundamental operation in many applications, such as computer vision, natural language processing, image processing, etc. Recent successes of convolutional neural networks in various deep learning applications put even higher demand on fast convolution. The high computation throughput and memory bandwidth of graphics processing units (GPUs) make GPUs a natural choice for accelerating convolution operations. However, maximally exploiting the available memory bandwidth of GPUs for convolution is a challenging task. This paper introduces a general model to address the mismatch between the memory bank width of GPUs and computation data width of threads. Based on this model, we develop two convolution kernels, one for the general case and the other for a special case with one input channel. By carefully optimizing memory access patterns and computation patterns, we design a communication-optimized kernel for the special case and a communication-reduced kernel for the general case. Experimental data based on implementations on Kepler GPUs show that our kernels achieve 5.16× and 35.5% average performance improvement over the latest cuDNN library, for the special case and the general case, respectively." @default.
- W2617031569 created "2017-06-05" @default.
- W2617031569 creator A5052870302 @default.
- W2617031569 creator A5064527448 @default.
- W2617031569 creator A5087244134 @default.
- W2617031569 date "2017-06-18" @default.
- W2617031569 modified "2023-10-16" @default.
- W2617031569 title "Optimizing Memory Efficiency for Convolution Kernels on Kepler GPUs" @default.
- W2617031569 cites W2099021415 @default.
- W2617031569 cites W2102605133 @default.
- W2617031569 cites W2147800946 @default.
- W2617031569 cites W2166524747 @default.
- W2617031569 cites W2172654076 @default.
- W2617031569 cites W2178700889 @default.
- W2617031569 cites W2530879419 @default.
- W2617031569 doi "https://doi.org/10.1145/3061639.3062297" @default.
- W2617031569 hasPublicationYear "2017" @default.
- W2617031569 type Work @default.
- W2617031569 sameAs 2617031569 @default.
- W2617031569 citedByCount "9" @default.
- W2617031569 countsByYear W26170315692018 @default.
- W2617031569 countsByYear W26170315692019 @default.
- W2617031569 countsByYear W26170315692020 @default.
- W2617031569 countsByYear W26170315692022 @default.
- W2617031569 crossrefType "proceedings-article" @default.
- W2617031569 hasAuthorship W2617031569A5052870302 @default.
- W2617031569 hasAuthorship W2617031569A5064527448 @default.
- W2617031569 hasAuthorship W2617031569A5087244134 @default.
- W2617031569 hasBestOaLocation W26170315692 @default.
- W2617031569 hasConcept C118615104 @default.
- W2617031569 hasConcept C150846664 @default.
- W2617031569 hasConcept C154945302 @default.
- W2617031569 hasConcept C173608175 @default.
- W2617031569 hasConcept C207963374 @default.
- W2617031569 hasConcept C31972630 @default.
- W2617031569 hasConcept C33923547 @default.
- W2617031569 hasConcept C41008148 @default.
- W2617031569 hasConcept C45347329 @default.
- W2617031569 hasConcept C459310 @default.
- W2617031569 hasConcept C50644808 @default.
- W2617031569 hasConcept C74193536 @default.
- W2617031569 hasConceptScore W2617031569C118615104 @default.
- W2617031569 hasConceptScore W2617031569C150846664 @default.
- W2617031569 hasConceptScore W2617031569C154945302 @default.
- W2617031569 hasConceptScore W2617031569C173608175 @default.
- W2617031569 hasConceptScore W2617031569C207963374 @default.
- W2617031569 hasConceptScore W2617031569C31972630 @default.
- W2617031569 hasConceptScore W2617031569C33923547 @default.
- W2617031569 hasConceptScore W2617031569C41008148 @default.
- W2617031569 hasConceptScore W2617031569C45347329 @default.
- W2617031569 hasConceptScore W2617031569C459310 @default.
- W2617031569 hasConceptScore W2617031569C50644808 @default.
- W2617031569 hasConceptScore W2617031569C74193536 @default.
- W2617031569 hasLocation W26170315691 @default.
- W2617031569 hasLocation W26170315692 @default.
- W2617031569 hasOpenAccess W2617031569 @default.
- W2617031569 hasPrimaryLocation W26170315691 @default.
- W2617031569 hasRelatedWork W1580730938 @default.
- W2617031569 hasRelatedWork W1789336918 @default.
- W2617031569 hasRelatedWork W1970773070 @default.
- W2617031569 hasRelatedWork W2073045545 @default.
- W2617031569 hasRelatedWork W2624440775 @default.
- W2617031569 hasRelatedWork W2963367891 @default.
- W2617031569 hasRelatedWork W2965967938 @default.
- W2617031569 hasRelatedWork W3022446491 @default.
- W2617031569 hasRelatedWork W4250047567 @default.
- W2617031569 hasRelatedWork W4293772185 @default.
- W2617031569 isParatext "false" @default.
- W2617031569 isRetracted "false" @default.
- W2617031569 magId "2617031569" @default.
- W2617031569 workType "article" @default.