Show simple item record

dc.contributor.authorHandleman, Aaron
dc.contributor.authorSinger, Kyle
dc.contributor.authorSchardl, Tao B.
dc.contributor.authorLee, I-Ting Angelina
dc.date.accessioned2025-08-13T15:43:28Z
dc.date.available2025-08-13T15:43:28Z
dc.date.issued2025-07-16
dc.identifier.isbn979-8-4007-1258-6
dc.identifier.urihttps://hdl.handle.net/1721.1/162362
dc.descriptionSPAA ’25, July 28-August 1, 2025, Portland, OR, USAen_US
dc.description.abstractIn a randomized work-stealing scheduler, parallel speedup depends on the spawn overhead, which workers pay to allow tasks to execute in parallel, and the steal overhead, which thieves pay to start executing new work. The importance of minimizing the spawn overhead in a randomized work-stealing scheduler is first formalized by Frigo et al., coined as the work-first principle [15], which states that one should minimize spawn overhead even at the expense of a larger steal overhead. Since then, many strategies have been proposed to reduce the spawn overhead, which is dominated by maintaining a per-worker double-ended queue, or deque, to keep track of available parallel work. In pursuit of zero spawn overhead, this work considers a strategy that eliminates the use of deques entirely, obviating the need for a worker to perform explicit bookkeeping or set up a deque to enable parallelism. To that end, we propose DLite, a compiler and runtime ABI (Application Binary Interface) that incurs near-zero spawn overhead, empirically measured to be about 6% compared to a regular function invocation. DLite pushes the tradeoffs advocated by the work-first principle to the extreme, which decreases the spawn overhead to almost nil, at the expense of a high steal cost. Specifically, DLite employs a backtracking strategy: When a steal attempt occurs, the victim provides its current stack and base pointers to the thief, and the thief then reconstructs the necessary state to realize the parallel execution. We have implemented Cilk-DLite, which extends the OpenCilk platform [33] to implement DLite. When the application has ample parallelism, Cilk-DLite exhibits similar scalability to OpenCilk with much lower spawn overhead. When the application lacks parallelism, the high steal cost in Cilk-DLite can impede scalability due to slower work distribution. We also implemented variants of Cilk-DLite that make different design choices to evaluate the tradeoffs between spawn overhead and steal cost.en_US
dc.publisherACM|37th ACM Symposium on Parallelism in Algorithms and Architecturesen_US
dc.relation.isversionofhttps://doi.org/10.1145/3694906.3743349en_US
dc.rightsCreative Commons Attributionen_US
dc.rights.urihttps://creativecommons.org/licenses/by/4.0/en_US
dc.sourceAssociation for Computing Machineryen_US
dc.titleTowards Zero Spawn Overhead: Work Stealing Without Dequesen_US
dc.typeArticleen_US
dc.identifier.citationAaron Handleman, Kyle Singer, Tao B. Schardl, and I-Ting Angelina Lee. 2025. Towards Zero Spawn Overhead: Work Stealing Without Deques. In Proceedings of the 37th ACM Symposium on Parallelism in Algorithms and Architectures (SPAA '25). Association for Computing Machinery, New York, NY, USA, 75–88.en_US
dc.contributor.departmentMassachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratoryen_US
dc.identifier.mitlicensePUBLISHER_POLICY
dc.eprint.versionFinal published versionen_US
dc.type.urihttp://purl.org/eprint/type/ConferencePaperen_US
eprint.statushttp://purl.org/eprint/status/NonPeerRevieweden_US
dc.date.updated2025-08-01T07:55:46Z
dc.language.rfc3066en
dc.rights.holderThe author(s)
dspace.date.submission2025-08-01T07:55:46Z
mit.licensePUBLISHER_CC
mit.metadata.statusAuthority Work and Publication Information Neededen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record