File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: A Model-Based Software Solution for Simultaneous Multiple Kernels on GPUs

TitleA Model-Based Software Solution for Simultaneous Multiple Kernels on GPUs
Authors
KeywordsMultitasking
Program processors
Supercomputers
Computing resource
concurrent kernel execution
Issue Date2020
PublisherAssociation for Computing Machinery, Inc. The Journal's web site is located at http://taco.acm.org
Citation
ACM Transactions on Architecture and Code Optimization, 2020, v. 17 n. 1, p. article no. 7 How to Cite?
AbstractAs a critical computing resource in multiuser systems such as supercomputers, data centers, and cloud services, a GPU contains multiple compute units (CUs). GPU Multitasking is an intuitive solution to underutilization in GPGPU computing. Recently proposed solutions of multitasking GPUs can be classified into two categories: (1) spatially partitioned sharing (SPS), which coexecutes different kernels on disjointed sets of compute units (CU), and (2) simultaneous multikernel (SMK), which runs multiple kernels simultaneously within a CU. Compared to SPS, SMK can improve resource utilization even further due to the interleaving of instructions from kernels with low dynamic resource contentions. However, it is hard to implement SMK on current GPU architecture, because (1) techniques for applying SMK on top of GPU hardware scheduling policy are scarce and (2) finding an efficient SMK scheme is difficult due to the complex interferences of concurrently executed kernels. In this article, we propose a lightweight and effective performance model to evaluate the complex interferences of SMK. Based on the probability of independent events, our performance model is built from a totally new angle and contains limited parameters. Then, we propose a metric, symbiotic factor, which can evaluate an SMK scheme so that kernels with complementary resource utilization can corun within a CU. Also, we analyze the advantages and disadvantages of kernel slicing and kernel stretching techniques and integrate them to apply SMK on GPUs instead of simulators. We validate our model on 18 benchmarks. Compared to the optimized hardware-based concurrent kernel execution whose kernel launching order brings fast execution time, the results of corunning kernel pairs show 11%, 18%, and 12% speedup on AMD R9 290X, RX 480, and Vega 64, respectively, on average. Compared to the Warped-Slicer, the results show 29%, 18%, and 51% speedup on AMD R9 290X, RX 480, and Vega 64, respectively, on average.
Persistent Identifierhttp://hdl.handle.net/10722/283311
ISSN
2019 Impact Factor: 1.309
2015 SCImago Journal Rankings: 0.490

 

DC FieldValueLanguage
dc.contributor.authorWU, H-
dc.contributor.authorLiu, WZ-
dc.contributor.authorLIN, H-
dc.contributor.authorWang, CL-
dc.date.accessioned2020-06-22T02:54:52Z-
dc.date.available2020-06-22T02:54:52Z-
dc.date.issued2020-
dc.identifier.citationACM Transactions on Architecture and Code Optimization, 2020, v. 17 n. 1, p. article no. 7-
dc.identifier.issn1544-3566-
dc.identifier.urihttp://hdl.handle.net/10722/283311-
dc.description.abstractAs a critical computing resource in multiuser systems such as supercomputers, data centers, and cloud services, a GPU contains multiple compute units (CUs). GPU Multitasking is an intuitive solution to underutilization in GPGPU computing. Recently proposed solutions of multitasking GPUs can be classified into two categories: (1) spatially partitioned sharing (SPS), which coexecutes different kernels on disjointed sets of compute units (CU), and (2) simultaneous multikernel (SMK), which runs multiple kernels simultaneously within a CU. Compared to SPS, SMK can improve resource utilization even further due to the interleaving of instructions from kernels with low dynamic resource contentions. However, it is hard to implement SMK on current GPU architecture, because (1) techniques for applying SMK on top of GPU hardware scheduling policy are scarce and (2) finding an efficient SMK scheme is difficult due to the complex interferences of concurrently executed kernels. In this article, we propose a lightweight and effective performance model to evaluate the complex interferences of SMK. Based on the probability of independent events, our performance model is built from a totally new angle and contains limited parameters. Then, we propose a metric, symbiotic factor, which can evaluate an SMK scheme so that kernels with complementary resource utilization can corun within a CU. Also, we analyze the advantages and disadvantages of kernel slicing and kernel stretching techniques and integrate them to apply SMK on GPUs instead of simulators. We validate our model on 18 benchmarks. Compared to the optimized hardware-based concurrent kernel execution whose kernel launching order brings fast execution time, the results of corunning kernel pairs show 11%, 18%, and 12% speedup on AMD R9 290X, RX 480, and Vega 64, respectively, on average. Compared to the Warped-Slicer, the results show 29%, 18%, and 51% speedup on AMD R9 290X, RX 480, and Vega 64, respectively, on average.-
dc.languageeng-
dc.publisherAssociation for Computing Machinery, Inc. The Journal's web site is located at http://taco.acm.org-
dc.relation.ispartofACM Transactions on Architecture and Code Optimization-
dc.rightsACM Transactions on Architecture and Code Optimization. Copyright © Association for Computing Machinery, Inc.-
dc.rights©ACM, YYYY. This is the author's version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in PUBLICATION, {VOL#, ISS#, (DATE)} http://doi.acm.org/10.1145/nnnnnn.nnnnnn-
dc.subjectMultitasking-
dc.subjectProgram processors-
dc.subjectSupercomputers-
dc.subjectComputing resource-
dc.subjectconcurrent kernel execution-
dc.titleA Model-Based Software Solution for Simultaneous Multiple Kernels on GPUs-
dc.typeArticle-
dc.identifier.emailWang, CL: clwang@cs.hku.hk-
dc.identifier.authorityWang, CL=rp00183-
dc.description.naturelink_to_subscribed_fulltext-
dc.identifier.doi10.1145/3377138-
dc.identifier.scopuseid_2-s2.0-85081572012-
dc.identifier.hkuros310353-
dc.identifier.volume17-
dc.identifier.issue1-
dc.identifier.spagearticle no. 7-
dc.identifier.epagearticle no. 7-
dc.publisher.placeUnited States-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats