FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2908650493> ?p ?o ?g. }

• W2908650493 abstract "Abstract Single GPU scaling is unable to keep pace with the soaring demand for high throughput computing. As such executing an application on multiple GPUs connected through an off-chip interconnect will become an attractive option to explore. However, much of the current code is written for a single GPU system. Porting such a code for execution on multiple GPUs is difficulty task. In particular, it requires programmer effort to determine how data is partitioned across multiple GPU cards and then launch the appropriate thread blocks that mostly accesses the data that is local to that card. Otherwise, cross-card data movement is an expensive operation. In this work we explore hardware support to efficiently parallelize a single GPU code for execution on multiple GPUs. In particular, our approach focuses on minimizing the number of remote memory accesses across the off-chip network without burdening the programmer to perform data partitioning and workload assignment. We propose a data-location aware thread block scheduler to schedule the thread blocks on the GPU that has most of its input data. The scheduler exploits well known observation that GPU workloads tend to launch a kernel multiple times iteratively to process large volumes of data. The memory accesses of the thread block across different iterations of a kernel launch exhibit correlated behavior. Our data location aware scheduler exploits this predictability to track memory access affinity of each thread block to a specific GPU card and stores this information to make scheduling decisions for future iterations. To further reduce the number of remote accesses we propose a hybrid mechanism that enables migrating or copying the pages between the memory of multiple GPUs based on their access behavior. Hence, most of the memory accesses are to the local GPU memory. Over an architecture consisting of two GPUs, our proposed schemes are able to improve the performance by 1.55× when compared to single GPU execution across widely used Rodinia [17], Parboil [18], and Graph [23] benchmarks." @default.
• W2908650493 created "2019-01-25" @default.
• W2908650493 creator A5018033573 @default.
• W2908650493 creator A5019948031 @default.
• W2908650493 creator A5051502689 @default.
• W2908650493 date "2019-05-01" @default.
• W2908650493 modified "2023-09-24" @default.
• W2908650493 title "Efficient automatic parallelization of a single GPU program for a multiple GPU system" @default.
• W2908650493 cites W1501501999 @default.
• W2908650493 cites W1979527452 @default.
• W2908650493 cites W1987773847 @default.
• W2908650493 cites W2003840611 @default.
• W2908650493 cites W2059301531 @default.
• W2908650493 cites W2067280503 @default.
• W2908650493 cites W2080592089 @default.
• W2908650493 cites W2082000934 @default.
• W2908650493 cites W2096939574 @default.
• W2908650493 cites W2102843684 @default.
• W2908650493 cites W2128268549 @default.
• W2908650493 cites W2130561688 @default.
• W2908650493 cites W2140455011 @default.
• W2908650493 cites W2273440736 @default.
• W2908650493 cites W2415201000 @default.
• W2908650493 cites W2464177207 @default.
• W2908650493 cites W2612048219 @default.
• W2908650493 cites W2625200202 @default.
• W2908650493 cites W2761710529 @default.
• W2908650493 cites W2801700034 @default.
• W2908650493 doi "https://doi.org/10.1016/j.vlsi.2018.12.006" @default.
• W2908650493 hasPublicationYear "2019" @default.
• W2908650493 type Work @default.
• W2908650493 sameAs 2908650493 @default.
• W2908650493 citedByCount "2" @default.
• W2908650493 countsByYear W29086504932020 @default.
• W2908650493 countsByYear W29086504932021 @default.
• W2908650493 crossrefType "journal-article" @default.
• W2908650493 hasAuthorship W2908650493A5018033573 @default.
• W2908650493 hasAuthorship W2908650493A5019948031 @default.
• W2908650493 hasAuthorship W2908650493A5051502689 @default.
• W2908650493 hasConcept C106251023 @default.
• W2908650493 hasConcept C111919701 @default.
• W2908650493 hasConcept C138101251 @default.
• W2908650493 hasConcept C165696696 @default.
• W2908650493 hasConcept C173608175 @default.
• W2908650493 hasConcept C199360897 @default.
• W2908650493 hasConcept C2777904410 @default.
• W2908650493 hasConcept C2778514511 @default.
• W2908650493 hasConcept C38652104 @default.
• W2908650493 hasConcept C41008148 @default.
• W2908650493 hasConcept C47487241 @default.
• W2908650493 hasConcept C68339613 @default.
• W2908650493 hasConceptScore W2908650493C106251023 @default.
• W2908650493 hasConceptScore W2908650493C111919701 @default.
• W2908650493 hasConceptScore W2908650493C138101251 @default.
• W2908650493 hasConceptScore W2908650493C165696696 @default.
• W2908650493 hasConceptScore W2908650493C173608175 @default.
• W2908650493 hasConceptScore W2908650493C199360897 @default.
• W2908650493 hasConceptScore W2908650493C2777904410 @default.
• W2908650493 hasConceptScore W2908650493C2778514511 @default.
• W2908650493 hasConceptScore W2908650493C38652104 @default.
• W2908650493 hasConceptScore W2908650493C41008148 @default.
• W2908650493 hasConceptScore W2908650493C47487241 @default.
• W2908650493 hasConceptScore W2908650493C68339613 @default.
• W2908650493 hasLocation W29086504931 @default.
• W2908650493 hasOpenAccess W2908650493 @default.
• W2908650493 hasPrimaryLocation W29086504931 @default.
• W2908650493 hasRelatedWork W2050019858 @default.
• W2908650493 hasRelatedWork W2101320423 @default.
• W2908650493 hasRelatedWork W2126830109 @default.
• W2908650493 hasRelatedWork W2276395270 @default.
• W2908650493 hasRelatedWork W2403702930 @default.
• W2908650493 hasRelatedWork W2755395451 @default.
• W2908650493 hasRelatedWork W2785460572 @default.
• W2908650493 hasRelatedWork W2931791443 @default.
• W2908650493 hasRelatedWork W2999509507 @default.
• W2908650493 hasRelatedWork W3012356635 @default.
• W2908650493 hasRelatedWork W3035128007 @default.
• W2908650493 hasRelatedWork W3040726395 @default.
• W2908650493 hasRelatedWork W3088700869 @default.
• W2908650493 hasRelatedWork W3090742253 @default.
• W2908650493 hasRelatedWork W3090784541 @default.
• W2908650493 hasRelatedWork W3112502950 @default.
• W2908650493 hasRelatedWork W3135674429 @default.
• W2908650493 hasRelatedWork W3157733976 @default.
• W2908650493 hasRelatedWork W3158564519 @default.
• W2908650493 hasRelatedWork W3185575180 @default.
• W2908650493 isParatext "false" @default.
• W2908650493 isRetracted "false" @default.
• W2908650493 magId "2908650493" @default.
• W2908650493 workType "article" @default.