Matches in SemOpenAlex for { <https://semopenalex.org/work/W4294335104> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W4294335104 abstract "Sliding window convolutional networks (ConvNets) have become a popular approach to computer vision problems such as image segmentation, and object detection and localization. Here we consider the problem of inference, the application of a previously trained ConvNet, with emphasis on 3D images. Our goal is to maximize throughput, defined as average number of output voxels computed per unit time. Other things being equal, processing a larger image tends to increase throughput, because fractionally less computation is wasted on the borders of the image. It follows that an apparently slower algorithm may end up having higher throughput if it can process a larger image within the constraint of the available RAM. We introduce novel CPU and GPU primitives for convolutional and pooling layers, which are designed to minimize memory overhead. The primitives include convolution based on highly efficient pruned FFTs. Our theoretical analyses and empirical tests reveal a number of interesting findings. For some ConvNet architectures, cuDNN is outperformed by our FFT-based GPU primitives, and these in turn can be outperformed by our CPU primitives. The CPU manages to achieve higher throughput because of its fast access to more RAM. A novel primitive in which the GPU accesses host RAM can significantly increase GPU throughput. Finally, a CPU-GPU algorithm achieves the greatest throughput of all, 10x or more than other publicly available implementations of sliding window 3D ConvNets. All of our code has been made available as open source project." @default.
- W4294335104 created "2022-09-02" @default.
- W4294335104 creator A5029097412 @default.
- W4294335104 creator A5061204760 @default.
- W4294335104 creator A5080613658 @default.
- W4294335104 date "2016-06-17" @default.
- W4294335104 modified "2023-09-30" @default.
- W4294335104 title "ZNNi - Maximizing the Inference Throughput of 3D Convolutional Networks on Multi-Core CPUs and GPUs" @default.
- W4294335104 doi "https://doi.org/10.48550/arxiv.1606.05688" @default.
- W4294335104 hasPublicationYear "2016" @default.
- W4294335104 type Work @default.
- W4294335104 citedByCount "0" @default.
- W4294335104 crossrefType "posted-content" @default.
- W4294335104 hasAuthorship W4294335104A5029097412 @default.
- W4294335104 hasAuthorship W4294335104A5061204760 @default.
- W4294335104 hasAuthorship W4294335104A5080613658 @default.
- W4294335104 hasBestOaLocation W42943351041 @default.
- W4294335104 hasConcept C102392041 @default.
- W4294335104 hasConcept C111919701 @default.
- W4294335104 hasConcept C113775141 @default.
- W4294335104 hasConcept C11413529 @default.
- W4294335104 hasConcept C114614502 @default.
- W4294335104 hasConcept C154945302 @default.
- W4294335104 hasConcept C157764524 @default.
- W4294335104 hasConcept C173608175 @default.
- W4294335104 hasConcept C199360897 @default.
- W4294335104 hasConcept C2776214188 @default.
- W4294335104 hasConcept C2778751112 @default.
- W4294335104 hasConcept C2779960059 @default.
- W4294335104 hasConcept C33923547 @default.
- W4294335104 hasConcept C41008148 @default.
- W4294335104 hasConcept C43521106 @default.
- W4294335104 hasConcept C45347329 @default.
- W4294335104 hasConcept C49154492 @default.
- W4294335104 hasConcept C50644808 @default.
- W4294335104 hasConcept C555944384 @default.
- W4294335104 hasConcept C70437156 @default.
- W4294335104 hasConcept C74193536 @default.
- W4294335104 hasConcept C75172450 @default.
- W4294335104 hasConcept C76155785 @default.
- W4294335104 hasConcept C78766204 @default.
- W4294335104 hasConcept C9390403 @default.
- W4294335104 hasConceptScore W4294335104C102392041 @default.
- W4294335104 hasConceptScore W4294335104C111919701 @default.
- W4294335104 hasConceptScore W4294335104C113775141 @default.
- W4294335104 hasConceptScore W4294335104C11413529 @default.
- W4294335104 hasConceptScore W4294335104C114614502 @default.
- W4294335104 hasConceptScore W4294335104C154945302 @default.
- W4294335104 hasConceptScore W4294335104C157764524 @default.
- W4294335104 hasConceptScore W4294335104C173608175 @default.
- W4294335104 hasConceptScore W4294335104C199360897 @default.
- W4294335104 hasConceptScore W4294335104C2776214188 @default.
- W4294335104 hasConceptScore W4294335104C2778751112 @default.
- W4294335104 hasConceptScore W4294335104C2779960059 @default.
- W4294335104 hasConceptScore W4294335104C33923547 @default.
- W4294335104 hasConceptScore W4294335104C41008148 @default.
- W4294335104 hasConceptScore W4294335104C43521106 @default.
- W4294335104 hasConceptScore W4294335104C45347329 @default.
- W4294335104 hasConceptScore W4294335104C49154492 @default.
- W4294335104 hasConceptScore W4294335104C50644808 @default.
- W4294335104 hasConceptScore W4294335104C555944384 @default.
- W4294335104 hasConceptScore W4294335104C70437156 @default.
- W4294335104 hasConceptScore W4294335104C74193536 @default.
- W4294335104 hasConceptScore W4294335104C75172450 @default.
- W4294335104 hasConceptScore W4294335104C76155785 @default.
- W4294335104 hasConceptScore W4294335104C78766204 @default.
- W4294335104 hasConceptScore W4294335104C9390403 @default.
- W4294335104 hasLocation W42943351041 @default.
- W4294335104 hasOpenAccess W4294335104 @default.
- W4294335104 hasPrimaryLocation W42943351041 @default.
- W4294335104 hasRelatedWork W1595151633 @default.
- W4294335104 hasRelatedWork W1789336918 @default.
- W4294335104 hasRelatedWork W2264773568 @default.
- W4294335104 hasRelatedWork W2473113737 @default.
- W4294335104 hasRelatedWork W26297114 @default.
- W4294335104 hasRelatedWork W2963367891 @default.
- W4294335104 hasRelatedWork W3046739674 @default.
- W4294335104 hasRelatedWork W4210398058 @default.
- W4294335104 hasRelatedWork W4293772185 @default.
- W4294335104 hasRelatedWork W4294335104 @default.
- W4294335104 isParatext "false" @default.
- W4294335104 isRetracted "false" @default.
- W4294335104 workType "article" @default.