FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4308596141> ?p ?o ?g. }

• W4308596141 endingPage "24" @default.
• W4308596141 startingPage "1" @default.
• W4308596141 abstract "A wide range of scientific and machine learning applications depend on highly optimized implementations of tensor computations. Exploiting the full capacity of a given processor architecture remains a challenging task, due to the complexity of the microarchitectural features that come into play when seeking near-peak performance. Among the state-of-the-art techniques for loop transformations for performance optimization, AutoScheduler [Zheng et al. 2020a ] tends to outperform other systems. It often yields higher performance as compared to vendor libraries, but takes a large number of runs to converge, while also involving a complex training environment. In this article, we define a structured configuration space that enables much faster convergence to high-performance code versions, using only random sampling of candidates. We focus on two-dimensional convolutions on CPUs. Compared to state-of-the-art libraries, our structured search space enables higher performance for typical tensor shapes encountered in convolution stages in deep learning pipelines. Compared to auto-tuning code generators like AutoScheduler, it prunes the search space while increasing the density of efficient implementations. We analyze the impact on convergence speed and performance distribution, on two Intel x86 processors and one ARM AArch64 processor. We match or outperform the performance of the state-of-the-art oneDNN library and TVM’s AutoScheduler, while reducing the autotuning effort by at least an order of magnitude." @default.
• W4308596141 created "2022-11-12" @default.
• W4308596141 creator A5008882200 @default.
• W4308596141 creator A5011198538 @default.
• W4308596141 creator A5015721738 @default.
• W4308596141 creator A5027517817 @default.
• W4308596141 creator A5054129043 @default.
• W4308596141 creator A5063162788 @default.
• W4308596141 creator A5081852342 @default.
• W4308596141 creator A5089282120 @default.
• W4308596141 date "2023-03-01" @default.
• W4308596141 modified "2023-10-01" @default.
• W4308596141 title "Autotuning Convolutions Is Easier Than You Think" @default.
• W4308596141 cites W1191262899 @default.
• W4308596141 cites W1562841074 @default.
• W4308596141 cites W1970141743 @default.
• W4308596141 cites W2034761517 @default.
• W4308596141 cites W2055312318 @default.
• W4308596141 cites W2077143534 @default.
• W4308596141 cites W2084379367 @default.
• W4308596141 cites W2129740858 @default.
• W4308596141 cites W2194775991 @default.
• W4308596141 cites W2570343428 @default.
• W4308596141 cites W2585460399 @default.
• W4308596141 cites W2588061952 @default.
• W4308596141 cites W2806891462 @default.
• W4308596141 cites W2983923412 @default.
• W4308596141 cites W2985039650 @default.
• W4308596141 cites W3005909967 @default.
• W4308596141 cites W3012249773 @default.
• W4308596141 cites W3156745629 @default.
• W4308596141 cites W4244254628 @default.
• W4308596141 cites W4245312332 @default.
• W4308596141 cites W4318256790 @default.
• W4308596141 doi "https://doi.org/10.1145/3570641" @default.
• W4308596141 hasPublicationYear "2023" @default.
• W4308596141 type Work @default.
• W4308596141 citedByCount "4" @default.
• W4308596141 countsByYear W43085961412023 @default.
• W4308596141 crossrefType "journal-article" @default.
• W4308596141 hasAuthorship W4308596141A5008882200 @default.
• W4308596141 hasAuthorship W4308596141A5011198538 @default.
• W4308596141 hasAuthorship W4308596141A5015721738 @default.
• W4308596141 hasAuthorship W4308596141A5027517817 @default.
• W4308596141 hasAuthorship W4308596141A5054129043 @default.
• W4308596141 hasAuthorship W4308596141A5063162788 @default.
• W4308596141 hasAuthorship W4308596141A5081852342 @default.
• W4308596141 hasAuthorship W4308596141A5089282120 @default.
• W4308596141 hasBestOaLocation W43085961411 @default.
• W4308596141 hasConcept C107598950 @default.
• W4308596141 hasConcept C113775141 @default.
• W4308596141 hasConcept C154945302 @default.
• W4308596141 hasConcept C159985019 @default.
• W4308596141 hasConcept C162324750 @default.
• W4308596141 hasConcept C166957645 @default.
• W4308596141 hasConcept C170723468 @default.
• W4308596141 hasConcept C173608175 @default.
• W4308596141 hasConcept C177264268 @default.
• W4308596141 hasConcept C192562407 @default.
• W4308596141 hasConcept C199360897 @default.
• W4308596141 hasConcept C204323151 @default.
• W4308596141 hasConcept C2776760102 @default.
• W4308596141 hasConcept C2777303404 @default.
• W4308596141 hasConcept C2777904410 @default.
• W4308596141 hasConcept C41008148 @default.
• W4308596141 hasConcept C45347329 @default.
• W4308596141 hasConcept C50522688 @default.
• W4308596141 hasConcept C50644808 @default.
• W4308596141 hasConcept C68339613 @default.
• W4308596141 hasConcept C79581498 @default.
• W4308596141 hasConcept C95457728 @default.
• W4308596141 hasConceptScore W4308596141C107598950 @default.
• W4308596141 hasConceptScore W4308596141C113775141 @default.
• W4308596141 hasConceptScore W4308596141C154945302 @default.
• W4308596141 hasConceptScore W4308596141C159985019 @default.
• W4308596141 hasConceptScore W4308596141C162324750 @default.
• W4308596141 hasConceptScore W4308596141C166957645 @default.
• W4308596141 hasConceptScore W4308596141C170723468 @default.
• W4308596141 hasConceptScore W4308596141C173608175 @default.
• W4308596141 hasConceptScore W4308596141C177264268 @default.
• W4308596141 hasConceptScore W4308596141C192562407 @default.
• W4308596141 hasConceptScore W4308596141C199360897 @default.
• W4308596141 hasConceptScore W4308596141C204323151 @default.
• W4308596141 hasConceptScore W4308596141C2776760102 @default.
• W4308596141 hasConceptScore W4308596141C2777303404 @default.
• W4308596141 hasConceptScore W4308596141C2777904410 @default.
• W4308596141 hasConceptScore W4308596141C41008148 @default.
• W4308596141 hasConceptScore W4308596141C45347329 @default.
• W4308596141 hasConceptScore W4308596141C50522688 @default.
• W4308596141 hasConceptScore W4308596141C50644808 @default.
• W4308596141 hasConceptScore W4308596141C68339613 @default.
• W4308596141 hasConceptScore W4308596141C79581498 @default.
• W4308596141 hasConceptScore W4308596141C95457728 @default.
• W4308596141 hasIssue "2" @default.
• W4308596141 hasLocation W43085961411 @default.
• W4308596141 hasLocation W43085961412 @default.
• W4308596141 hasLocation W43085961413 @default.
• W4308596141 hasOpenAccess W4308596141 @default.

• next