Mike Rainey’s research page

Welcome, I’m a Research Scientist at Carnegie Mellon University. Previously, I was a researcher at Indiana University, and before that, INRIA-Paris in the ERC Deepsea / Gallium group. Before that, I did my PhD with John Reppy at the University of Chicago, and a postdoc with Umut Acar at the Max Planck Institute for Software Systems in Kaiserslautern, Germany.

Contact me by email at me@mike-rainey.site.

Career

Publications

This list is organized by research topic. A number of papers appear under multiple topics, as appropriate. For a list without duplicates, see the list of references at the bottom of this page.

Granularity control

Making parallel programs more robust in the face of parallel-specific overheads: (Rainey et al. 2021; Umut A. Acar et al. 2019; Umut A. Acar et al. 2018; Umut A. Acar, Charguéraud, and Rainey 2016; Umut A. Acar, Charguéraud, and Rainey 2015; Umut A. Acar, Charguéraud, and Rainey 2011; Bergstrom et al. 2010; Bergstrom et al. 2012; Rainey 2010)

Scheduling parallel computations

Design and implementation of algorithms to map computations generated by parallel programs onto multicore machines: (Umut A. Acar, Charguéraud, and Rainey 2015; Umut A. Acar, Charguéraud, and Rainey 2013; Bergstrom et al. 2010; Bergstrom et al. 2012; Fluet et al. 2010; Fluet et al. 2008; Rainey 2010; Fluet, Rainey, and Reppy 2008)

Design of parallel programming languages and APIs

Programming languages to raise the level of abstraction of parallel programs: (Umut A. Acar et al. 2016; Umut A. Acar, Charguéraud, and Rainey 2012; Fluet et al. 2007)

Data parallelism

Work-efficient algorithm for fast parallel depth-first search of directed graphs: (Umut A. Acar, Charguéraud, and Rainey 2015)

Compiler optimization to control the layout of parallel-friendly data structures: (Bergstrom et al. 2013)

Data-parallel library supporting fusion optimizations: (Westrick et al. 2022)

Language support for safe and efficient serialization of data

A type-safe calculus for writing fast traversal and construction of serialized representation of trees: (Koparkar et al. 2021; Vollmer et al. 2019)

Algorithms and data structures

Efficient algorithms and data structures that are amenable to parallel programming: (Charguéraud and Rainey 2017; Umut A. Acar, Charguéraud, and Rainey 2015; Umut A. Acar, Charguéraud, and Rainey 2014; Wise et al. 2005)

Concurrent data structures: (Umut A. Acar, Ben-David, and Rainey 2017)

Compiler construction

Engineering the SML/NJ compiler to handle advanced features of foreign-function calls. (Blume, Rainey, and Reppy 2008)

Techniques for analyzing the performance of parallel programs

A technique to help understand the causes of poor speedups: (Umut A. Acar, Charguéraud, and Rainey 2017)

Talks

The best multicore refactoring you’ve never heard of

FHPNC Workshop, September 2023; (Rainey 2023)
short paper; video; slides; reference source code

Task Parallel Assembly Language for Uncompromising Parallelism

PLDI, June 2021; (Rainey et al. 2021)
video

Provably and Practically Efficient Granularity Control

PPoPP, February 2019; (Umut A. Acar et al. 2019)
video; slides

Heartbeat Scheduling: Provable Efficiency for Nested Parallelism

PLDI, June 2018; (Umut A. Acar et al. 2018)
video; slides

Efficient representation of Large Dynamic Sequences in ML

ML Workshop, September, 2017; (Charguéraud and Rainey 2017)
video
Supercomputing, November, 2016; (Umut A. Acar, Charguéraud, and Rainey 2015)
video; slides

Scheduling Parallel Programs by Work Stealing with Private Deques

PPoPP, February 2013; (Umut A. Acar, Charguéraud, and Rainey 2013)
slides

Higher-level Implicit Parallelism with PASL

LAME, July 2013; (Umut A. Acar, Charguéraud, and Rainey 2012)
slides

Fork-join Model and Work Stealing

MPI-SWS weekly seminar, June 2011
slides

Software

My github profile.

Task Parallel Assembly Language for Uncompromising Parallelism

Our PLDI’21 paper extends prior work on Heartbeat Scheduling: it addresses the challenge of delivering a suitable heart beat mechanism via hardware interrupts, and it addresses barriers to sequential performance by proposing an assembly language equipped with support for task parallelism. Here, we provide a long version of the paper, complete with an appendix, and supplemental materials, including the artifact evaluation and an executable dynamic semantics in PLT Redex.

Heartbeat scheduling

Our PLDI’18 paper presents a proof and an experimental evaluation of our Heartbeat Scheduling technique (Umut A. Acar et al. 2018). This project page hosts the Coq source code of our proofs and the C++ prototype and benchmarking scripts we used for our experimental evaluation. This is the place to go if you are interested to check the proofs or to repeat our experiments.

A work-efficient algorithm for parallel unordered depth-first search

This project features a C++ implementation of the fast DFS-like graph-traversal algorithm from our SC’15 paper (Umut A. Acar, Charguéraud, and Rainey 2015).

Chunked sequence

This project features a C++ template library which implements ordered, in-memory containers that are based on a new B-tree-like data structure.

Parallel Algorithm Scheduling Library (PASL)

PASL is a C++ library that provides algorithms for mapping computations generated by programs with implicit threading to multicore machines.

Manticore

Manticore is a parallel programming language aimed at general-purpose applications that run on multicore processors.

Standard ML of New Jersey

I worked on the back end of the compiler. My main projects covered code generation for the x86_64 and support for foreign-function calls.

Service

References

Get the bibtex file used to generate these references.

Acar, Umut A., Vitaly Aksenov, Arthur Charguéraud, and Mike Rainey. 2019. “Provably and Practically Efficient Granularity Control.” In Proceedings of the 24th Symposium on Principles and Practice of Parallel Programming, 214–28. PPoPP ’19. New York, NY, USA: ACM. http://mike-rainey.site/papers/oracle-ppop19-long.pdf.
Acar, Umut A, Naama Ben-David, and Mike Rainey. 2017. “Contention in Structured Concurrency: Provably Efficient Dynamic NonZero Indicators for Nested Parallel Computation.” In 22nd ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming. PPoPP ’17. ACM. http://mike-rainey.site/papers/dynsnzi.pdf.
Acar, Umut A., Arthur Charguéraud, and Mike Rainey. 2017. Parallel Work Inflation, Memory Effects, and Their Empirical Analysis. CoRR. Vol. abs/1709.03767. http://arxiv.org/abs/1709.03767.
Acar, Umut A, Arthur Charguéraud, Adrien Guatto, Mike Rainey, and Filip Sieczkowski. 2018. “Heartbeat Scheduling: Provable Efficiency for Nested Parallelism.” In 39th ACM SIGPLAN Conference on Programming Language Design and Implementation. PLDI ’18. ACM. http://mike-rainey.site/papers/heartbeat.pdf.
Acar, Umut A, Arthur Charguéraud, and Mike Rainey. 2011. “Oracle Scheduling: Controlling Granularity in Implicitly Parallel Languages.” In Proceedings of the 2011 ACM International Conference on Object Oriented Programming Systems Languages and Applications, 46:499–518. OOPSLA ’11 10. ACM. http://mike-rainey.site/papers/oracle_scheduling.pdf.
———. 2012. “Efficient Primitives for Creating and Scheduling Parallel Computations.” DAMP ’12. http://mike-rainey.site/papers/damp2012_primitives.pdf.
———. 2013. “Scheduling Parallel Programs by Work Stealing with Private Deques.” In 18th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, 48:219–28. PPoPP ’13 8. ACM. http://mike-rainey.site/papers/full.pdf.
———. 2014. “Theory and Practice of Chunked Sequences.” In The 22nd Annual European Symposium on Algorithms, 25–36. ESA ’14. Springer. http://mike-rainey.site/papers/chunked_seq.pdf.
———. 2015. “A Work-Efficient Algorithm for Parallel Unordered Depth-First Search.” In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, 67:1–12. SC ’15. ACM. http://mike-rainey.site/papers/pdfs_sc15.pdf.
———. 2016. “Oracle-Guided Scheduling for Controlling Granularity in Implicitly Parallel Languages.” In Journal of Functional Programming. Cambridge University Press. http://mike-rainey.site/papers/jfp-oracle-guided.pdf.
Acar, Umut A, Arthur Charguéraud, Mike Rainey, and Filip Sieczkowski. 2016. “Dag-Calculus: A Calculus for Parallel Computation.” In The 26th ACM SIGPLAN International Conference on Functional Programming. ICFP ’16. ACM. http://mike-rainey.site/papers/dag-calculus.pdf.
Bergstrom, Lars, Matthew Fluet, Mike Rainey, John Reppy, Stephen Rosen, and Adam Shaw. 2013. “Data-Only Flattening for Nested Data Parallelism.” In 18th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, 48:81–92. PPoPP ’13 8. ACM. http://mike-rainey.site/papers/ppopp13-flat.pdf.
Bergstrom, Lars, Matthew Fluet, Mike Rainey, John Reppy, and Adam Shaw. 2012. “Lazy Tree Splitting.” In Journal of Functional Programming, 22:382–438. 4-5. Cambridge University Press. http://mike-rainey.site/papers/jfp-lts-submitted.pdf.
Bergstrom, Lars, Mike Rainey, John Reppy, Adam Shaw, and Matthew Fluet. 2010. “Lazy Tree Splitting.” In The 20th ACM SIGPLAN International Conference on Functional Programming, 45:93–104. ICFP ’10 9. ACM. http://mike-rainey.site/papers/icfp10-lts.pdf.
Blume, Matthias, Michael Rainey, and John Reppy. 2008. “Calling Variadic Functions from a Strongly-Typed Language.” ML ’08. ACM. http://mike-rainey.site/papers/ml-varargs.pdf.
Charguéraud, Arthur, and Mike Rainey. 2017. Efficient Representations for Large Dynamic Sequences in ML.” ML Family Workshop. https://hal.inria.fr/hal-01669407/file/chunkseq_ml.pdf.
Fluet, Matthew, Mike Rainey, and John Reppy. 2008. “A Scheduling Framework for General-Purpose Parallel Languages.” In The 13th ACM SIGPLAN International Conference on Functional Programming, 43:241–52. ICFP ’08 9. ACM. http://mike-rainey.site/papers/icfp08-sched.pdf.
Fluet, Matthew, Mike Rainey, John Reppy, and Adam Shaw. 2008. “Implicitly-Threaded Parallelism in Manticore.” In The 13th ACM SIGPLAN International Conference on Functional Programming, 43:119–30. ICFP ’08 9. ACM. http://mike-rainey.site/papers/icfp08-implicit.pdf.
———. 2010. “Implicitly Threaded Parallelism in Manticore.” In Journal of Functional Programming, 20:537–76. 5-6. Cambridge University Press.
Fluet, Matthew, Mike Rainey, John Reppy, Adam Shaw, and Yingqi Xiao. 2007. “Manticore: A Heterogeneous Parallel Language.” ACM.
Koparkar, Chaitanya, Mike Rainey, Michael Vollmer, Milind Kulkarni, and Ryan R. Newton. 2021. “Efficient Tree-Traversals: Reconciling Parallelism and Dense Data Representations.” In Proc. ACM Program. Lang. Vol. 5. ICFP. New York, NY, USA: Association for Computing Machinery. doi:10.1145/3473596.
Rainey, Mike. 2010. “Effective Scheduling Techniques for High-Level Parallel Programming Languages.” PhD thesis, University of Chicago. http://mike-rainey.site/papers/rainey-phd.pdf.
———. 2023. The Best Multicore Refactoring You’ve Never Heard of. http://mike-rainey.site/papers/pardefunc.pdf.
Rainey, Mike, Kyle Hale, Ryan R. Newton, Nikos Hardavellas, Simone Campanoni, Peter Dinda, and Umut A. Acar. 2021. “Task Parallel Assembly Language for Uncompromising Parallelism.” In Proceedings of the 42nd ACM SIGPLAN Conference on Programming Language Design and Implementation. PLDI ’21. New York, NY, USA: ACM. http://mike-rainey.site/papers/tpal-long.pdf.
Vollmer, Michael, Chaitanya Koparkar, Mike Rainey, Laith Sakka, Milind Kulkarni, and Ryan R. Newton. 2019. “LoCal: A Language for Programs Operating on Serialized Data.” In 40th ACM SIGPLAN Conference on Programming Language Design and Implementation. PLDI ’19. ACM. http://mike-rainey.site/papers/lo-cal19.pdf.
Westrick, Sam, Mike Rainey, Daniel Anderson, and Guy E. Blelloch. 2022. “Parallel Block-Delayed Sequences.” In Proceedings of the 27th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming, 61–75. PPoPP ’22. New York, NY, USA: Association for Computing Machinery. doi:10.1145/3503221.3508434.
Wise, David S., Craig Citro, Joshua Hursey, Fang Liu, and Michael Rainey. 2005. “A Paradigm for Parallel Matrix Algorithms: Scalable Cholesky.” In In Euro-Par 2005 – Parallel Processing. http://dx.doi.org/10.1007/11549468_76.