Aleksandar Prokopec and Dmitry Petrashko Scalablitz documentation, 2013 URL http://scala-blitz.github.io/home/documentation/ T H Tzen and L M Ni Trapezoid parallel A practical self-scheduling: for scheduling scheme compilers IEEE Trans Parallel Distrib Syst. Nathan Bronson Scalastm implementation https://nbronson.github.io/scala-stm/, 2011. Arvind, R S Nikhil, and K K Pingali I-structures: for parallel Data structures computing ACM Trans Prog Lang and Sys 11(4):598 632, October 1989. Vivek Sarkar Optimized unrolling of nested loops In ICS Proceedings York, NY, 14th pages USA best resume writing services in nyc reviews, New international Supercomputing, ’00, 153–166, conference of the on 2000 ACM ISBN 1-58113-270-0 doi: 10.1145/335231.335246 URL http://doi.acm.org/10.1145/335231.335246. Maurice Herlihy and J Eliot B Moss Transactional Architectural support lock-free for memory: data structures In ISCA of pages NY, Architecture, New Computer York writing dissertation proposals, ’93, the Annual on 20th 289–300, Proceedings Symposium International USA, 1993 ACM ISBN 0-8186-3810-9 doi: 10.1145/165123.165164 URL http://doi.acm.org/10.1145/165123.165164. Philipp Miller research paper about chocolates, and Viktor Kuhn, Haller, Vojin Klang, Roland Prokopec, Heather Aleksandar Jovanovic Scala improvement and proposal: Futures promises (SIP-14) 2012 See http://docs.scala-lang.org/sips/pending/futures-promises.html. Kumar, Annual Software Proceedings on of volume International Conference Dagstuhl, Leibniz (LIPIcs), Foundations Theoretical of Technology 427–451, (FSTTCS Computer Science 4 in editors, pages 2009), Informatics IARCS and Germany, 2009 Schloss Dagstuhl–Leibniz-Zentrum fuer Informatik ISBN 978-3-939897-13-2 doi: http://dx.doi.org/10.4230/LIPIcs.FSTTCS.2009.2338 URL http://drops.dagstuhl.de/opus/volltexte/2009/2338 Nir Shavit and Dan Touitou Software transactional memory In the USA, Symposium ’95, on Fourteenth of NY, Computing, York, Distributed 204–213, Annual Principles New of Proceedings pages ACM PODC 1995 ACM ISBN 0-89791-710-3 doi: 10.1145/224964.224987 URL http://doi.acm.org/10.1145/224964.224987 Ori Shalev and Nir Shavit Split-ordered Lock-free lists: extensible hash tables J ACM, 53(3):379–405, May 2006 ISSN 0004-5411 doi: 10 1145/1147954.1147958 URL http://doi.acm.org/10.1145/1147954.1147958 Maged M Michael High and lock-free performance dynamic hash tables list-based sets In Parallel ACM New ’02, 73–82, SPAA Fourteenth Annual the NY, Algorithms of USA, Architectures, Symposium Proceedings and on York, pages 2002 ACM ISBN 1-58113-529-7 doi: 10.1145/564870.564881 URL http://doi.acm.org/10.1145/564870.564881. Erik Meijer Your mouse is a database Commun ACM, 55(5):66–73, 2012. Phil Bagwell Fast and Length Variable Functional Hash-Lists, Deques, Lists, Arrays Technical report, 2002. Ralf Hinze and Ross Paterson Finger A data trees: simple general-purpose structure J Funct Program 16(2):197–217, March 2006 ISSN 0956-7968 doi: 10.1017/S0956796805005769 URL http://dx.doi.org/10.1017/S0956796805005769. G M Adelson-Velsky and E M Landis An organization for of the algorithm information Doklady 146: 263266, Nauk Akademii SSSR, 1962. Nimar S Arora, Robert D Blumofe, and C Greg Plaxton Thread scheduling for multiprogrammed multiprocessors In of Proceedings annual and New tenth 119 symposium algorithms Parallel NY, the SPAA 129, ACM on USA, pages 98, architectures, York, 1998 ACM ISBN 0-89791-989-0. Jr R H Halstead MultiLISP: for concurrent language A symbolic computation ACM Trans Prog Lang and Sys 7(4):501–538, October 1985. Arvind K Sujeeth Composition reuse and domain-specific with compiled languages In In Proceedings of ECOOP, 2013. Jr Henry C Baker and Carl Hewitt The collection of incremental garbage processes In Proc Symp on Art Int and Prog Lang 1977. Richard E Jones and Rafael Dueire Lins Garbage Collection: Memory Automatic Dynamic for Algorithms Management John Wiley, 1996 ISBN 0-471-94148-4. Aleksandar and Bagwell, Martin Prokopec, Phil Odersky Cache-Aware Lock-Free Concurrent Hash Tries Technical report, 2011. Dan Lentz Common lisp ctrie implementation, 2013 URL https://github.com/danlentz/cl-ctrie Aleksandar Haller, Martin and Philipp Prokopec, Odersky Containers Mutators and Aggregates, and Reactive Isolates for Programming In Scala 2014, 2014 URL https://github.com/storm-enroute/reactive-collections,http: //reactive-collections.com/ Mark Moir and Nir Shavit Concurrent data structures, 2004. Aleksandar Martin Miller, Prokopec, Haller, and Heather Philipp Tobias Schlatter, Odersky Flowpools: deterministic A concurrent lock-free dataflow abstraction In LCPC, pages 158–173, 2012. Henning Makholm Answer the mathematics forum at (stackexchange) https://math.stackexchange.com/questions/297375/bins-in-balls-where- bin-size-grows-exponentially, 2013 Hans-J Boehm, Michael Atkinson, and Russ Plass Ropes: An alternative to strings Softw Pract Exper 25(12): 1315–1330, December 1995 ISSN 0038-0644 Tobias Miller, Aleksandar Schlatter, Martin Heather Prokopec, Philipp and Haller, Odersky Multi-lane flowpools: A detailed look Technical September EPFL, Lausanne, report, 2012. Tobias Prokopec, Miller, Aleksandar Heather Haller, and Martin Philipp Schlatter, Odersky Flowseqs: Barrier-free parseqs Technical report, January EPFL how to write computer science dissertation, Lausanne, 2013. Doug Lea Doug lea’s workstation, 2014 URL http://g.oswego.edu/ Matteo Frigo, Charles E Leiserson, and Keith H Randall The multithreaded Cilk-5 of implementation the language In conference SIGPLAN pages York, Programming and language the 1998 on implementation, of NY, Proceedings ACM PLDI ’98, USA, New design 212–223, 1998 ACM ISBN 0-89791-987-4 doi: 10.1145/277650.277725 URL http://doi.acm.org/10.1145/277650.277725. Timothy L Harris, Fraser, Ian and Keir A Pratt A practical multi-word compare-and-swap operation In 265–279, on Proceedings DISC pages of Conference ’02, UK, International Computing, the Distributed UK, London, 16th 2002 Springer-Verlag ISBN 3-540-00073-9 URL http://dl.acm.org/citation.cfm id=645959.676137. C D Polychronopoulos and D J Kuck Guided scheme self-scheduling: scheduling parallel for practical A supercomputers IEEE Trans Comput 36(12):1425–1439, December 1987 ISSN 0018-9340 doi: 10.1109/TC.1987.5009495 URL http://dx.doi.org/10.1109/TC.1987.5009495. Iulian Dragos and Martin Odersky Compiling type user-directed through generics specialization In on Object-Oriented the workshop 4th NY, Programming USA, Languages of ICOOOLPS York, the Optimization 42–47, ’09, and Systems, of pages Implementation, Proceedings Compilation, New 2009 ACM Maged M Michael and Michael L Scott Simple, concurrent non-blocking queue and and fast, blocking practical algorithms In PODC, pages 267–275, 1996. William Pugh Skip probabilistic lists: alternative A balanced to trees Commun ACM, 33(6):668–676, June 1990a ISSN 0001-0782 doi: 10.1145/78973.78977 URL http://doi.acm.org/10.1145/78973.78977. Edward Fredkin Trie memory Commun ACM, 3(9):490–499, September 1960 ISSN 0001-0782 doi: 10.1145/367390.367400 URL http://doi.acm.org/10.1145/367390.367400. Rene De La Briandais File using variable length searching keys In (Western), Presented Papers York, 1959, NY, pages the Conference rivers cuomo admission essay to harvard, ’59 New 295–298, Computer March Joint Western 3-5, at IRE-AIEE-ACM USA, Chris Okasaki Purely Functional Data Structures Cambridge York, University New Press, USA, NY, 1998 ISBN 0-521-63124-6. Philip Bagwell and Tiark Rompf RRB-Trees: Efficient Immutable Vectors Technical report comment faire une dissertation de philosophie, 2011. Michael Schröder Haskell ctrie implementation, 2014 URL http://hackage.haskell.org/package/ctrie Haim Kaplan and Robert E Tarjan Persistent via catenation with lists recursive slow-down In the Theory NY, ’95, pages Twenty-seventh of Proceedings Symposium ACM Annual STOC York, of Computing, USA, 93–102, New on 1995 ACM ISBN 0-89791-718-9 doi: 10.1145/225058.225090 URL http://doi.acm.org/10.1145/225058.225090. Aleksandar Prokopec Phil Bagwell Ideal hash trees, 2001. Eugene Burmako and Martin Odersky Scala Macros, Technical Report In Valentin Metacom- International putation, Third Turchin on Workshop 2012 URL http://scalamacros.org/. William N Scherer, Doug and Michael Lea, L Scott Scalable synchronous queues Commun ACM, 52(5):100–111, 2009. Olivier Haibo Haichuan and Tardieu, Wang, Lin A work-stealing scheduler for parallelism with x10’s task suspension In symposium 267–276, of NY, Proceedings and USA, Principles on ACM 17th New the York, PPoPP pages Programming, Parallel of ’12, Practice SIGPLAN 2012 ACM ISBN 978-1-4503-1160-1 doi: 10.1145/2145816.2145850 URL http://doi.acm.org/10.1145/2145816.2145850 Witold Litwin Trie hashing In Data, 19–29, Management ’81, pages 1981 International the of York, SIGMOD New ACM SIGMOD USA, on of Conference Proceedings NY, 1981 ACM ISBN 0-89791-040-0 doi: 10.1145/582318.582322 URL http://doi.acm.org/10.1145/582318.582322 Faith Fich, Nir Hendler, Danny and Shavit On conditional synchronization inherent weakness the of primitives In Principles 80–87, on Annual the Computing, ACM Proceed- ’04, of USA, Symposium Distributed ings pages NY, Twenty-third York, New PODC of 2004 ACM ISBN 1-58113-802-4 doi: 10.1145/1011767.1011780 URL http://doi.acm.org/10.1145/1011767.1011780. Robert D Blumofe and Charles E Leiserson Scheduling by work computations multithreaded stealing J ACM, 46(5):720–748, September Second, concurrent traversed structures the can absence data in only most concurrent be of modifications We say that concurrent quiescently such data are structures consistent The on falls ensuring task the of quiescence programmer This scalable structure a thesis a called concurrent novel, presents lock-free, data Ctrie which a lock-free, linearizable, supports constant-time snapshot operation The used to parallelize for snapshot need without the Ctrie operations is Ctrie operation quiescence We show how data different the can tree-like linearizable, be snapshot applied concurrent, lock-free, lock-free constant-time operation to structures. Iulian Dragos Compiling Scala for Performance PhD thesis, IC, Lausanne, 2010. Douglas Baskins Judy array implementation http://judy.sourceforge.net/, 2000. Andy Buytaert, Dries and Lieven Georges, Eeckhout Statistically rigorous java performance evaluation In OOPSLA, pages 57–76, 2007 [Goetz Joshua et Brian Bloch, Bowbeer, Peierls] Holmes, Doug and Lea, Joseph Bowbeer, Bloch, Tim Holmes, David and Goetz, al.(2006)Goetz, Lea, Peierls Java Concurrency in Practice Addison-Wesley Longman, Amsterdam, 2006 ISBN 0321349601 URL http://www.amazon.de/ Java-Concurrency-Practice-Brian-Goetz/dp/0321349601. Nathan Bronson Composable High-Performance Operations Concurrent on Collections PhD thesis, 2011. Aleksandar Prokopec https://github.com/storm-enroute/reactive- (Reactive collections Collections repository), 2014. Maged M Michael Hazard memory lock-free Safe for pointers: reclamation objects IEEE Trans Parallel Distrib Syst 15(6):491–504, June 2004 ISSN 1045-9219 doi: 10.1109/TPDS.2004.8 URL http://dx.doi.org/10.1109/TPDS.2004.8. Maurice Herlihy A for methodology implementing highly concurrent data objects ACM Trans Program Lang Syst 15(5):745–770, November 1993 ISSN 0164-0925 doi: 10.1145/161468.161469 URL http://doi.acm.org/10 1145/161468.161469. Simon Peyton Jones Harnessing the in data multicores: parallelism Nested haskell In APLAS Asian ’08, pages Languages 6th and Proceedings Systems, the Programming Symposium Berlin, on Heidelberg, of 138–138, 2008 Springer-Verlag ISBN 978-3-540-89329-5 doi: 10.1007/978-3-540-89330-1_10 URL http://dx.doi.org/10.1007/978-3-540-89330-1_10 Eric Victor Chase, Guy Maessen, Joe Sukyoung Hallett, Jan-Willem Luchangco, Ryu, David Allen, L Steele Jr and Sam Tobin The Fortress Language Specification Technical report, Sun Microsystems, Inc 2007 URL http: //research.sun.com/projects/plrg/Publications/fortress1.0beta.pdf. Aleksandar and Dmitry Prokopec, Petrashko, Martin Odersky On Parallel for Lock-Free Data Iterators Work-stealing Structures Technical report, 2014. The the data-parallel that augment languages in with declarative-style existing fits libraries collection operations many nicely programming model bulk today Data operations many be mapping by or filtering at single or like a processors collection executed processor can reduction, once However, multiple overcome challenges when there to collection parallelizing are operations. Maurice Herlihy and Nir Shavit The Art of Multiprocessor Programming Morgan Kaufmann Publishers Inc San Francisco, CA research paper for money, USA, 2008 ISBN 0123705916, 9780123705914. Philippe Vijay Vivek Praun, Christoph Saraswat, von Grothoff dissertation en philo mthode, Charles, and Christian Sarkar X10: non-uniform object-oriented an to approach cluster computing In on pages ’05, OOPSLA Object-oriented USA, and annual pro- Proceedings ACM languages, of applications, gramming, 519–538, York, systems, SIGPLAN NY, 20th the conference New 2005 ACM ISBN 1-59593-031-0 Maurice Herlihy and Jeannette M Wing Linearizability: correctness concurrent A for condition objects ACM on 12(3):463–492, Transactions July Languages and Systems, Programming 1990. Guy E Blelloch Nesl: A nested data-parallel language Technical USA, PA, Pittsburgh, report, 1992. H T Kung and Philip L Lehman Concurrent search of binary manipulation trees ACM Trans Database Syst 5(3):354–382, September 1980 ISSN 0362-5915 doi: 10.1145/320613.320619 URL http://doi.acm.org/10.1145/320613.320619 First, it collection is in multiple challenging a parallel construct by to processors Traditionally, variants by their are update thread-safe with backed data of collections structures operations Such called data data structures concurrent are structures Their operations interprocessor the slower are synchronization update than generally single-threaded and update require corresponding operations Synchronization invalidate costs as easily operations or parallelizing from performance such can gains bulk mapping filtering This reduce data-parallel interprocessor presents boosting collection framework thesis range a of parallel with synchronization, need the that structures operation effectively a for data performance The in in parallel Scala, collection but implemented is managed techniques be to framework thesis applied the this can other runtimes. James Reinders Intel threading building blocks O’Reilly & Associates, Inc Sebastopol, CA, first USA, edition, 2007 ISBN 9780596514808. Timothy L Harris A non-blocking implementation of pragmatic linked-lists In DISC 300–314, International of on UK, ’01, the Conference Computing, UK, pages Proceedings Distributed London, 15th 2001 Springer-Verlag ISBN 3-540-42605-1 URL http://dl.acm.org/citation.cfm id=645958.676105. John M Mellor-Crummey Concurrent queues: Practical fetch-and-Φ algorithms 1987. William Pugh Concurrent maintenance of skip lists Technical College report how to buy an essay, MD, USA, Park, 1990b. Tom Knight An functional architecture for mostly languages In of on Conference 1986 pages the Programming, Functional LFP 105–112, NY, USA, New York, LISP Proceedings and ACM ’86, 1986 ACM ISBN 0-89791-200-4 doi:10.1145/319838.319854 URL http://doi.acm.org/10.1145/319838.319854 David Chase and Yossi Lev Dynamic circular work-stealing deque In and Proceedings annual New NY, pages ’05, algorithms in USA, symposium seventeenth Parallelism architectures, York, the ACM on of SPAA 21–28, 2005 ACM ISBN 1-58113-986-1 Clyde P Kruskal and Alan Weiss Allocating on parallel subtasks independent processors IEEE Trans Softw Eng 11(10):1001–1016, October 1985 ISSN 0098-5589 doi: 10.1109/TSE.1985.231547 URL http://dx.doi.org/10.1109/TSE.1985.231547 Ole Agesen, David L Detlefs, Christine H Flood, Alexander T Garthwaite, Paul A Martin, Nir N Shavit, and Guy L Steele Jr Dcas-based concurrent deques, 2000. Michael G Burke, Knobe, Kathleen Newton what is a term paper, and Vivek Ryan Sarkar Concurrent collections programming model In Computing, of Encyclopedia pages Parallel 364–371 2011. Guojing Cong, Sreedhar B Kodali, Lea, Vijay Doug Krishnamoorthy, Sriram A Saraswat, and Tong Wen Solving adaptive using irregular large, problems graph work-stealing In ICPP, pages 536–545, 2008. C Koelbel and P Mehrotra Compiling loops name-space global for parallel distributed execution Parallel Distributed IEEE and on, oct –451, Systems, Transactions 2(4):440 1991 ISSN 1045-9219 doi: 10.1109/71.97901. Aleksandar Bagwell, Prokopec, Rompf, Martin and Tiark Phil Odersky A generic parallel collection framework In international Proceedings Heidelberg, Volume on Berlin, of processing Euro-Par’11, II, Part 136–147, 17th the - conference Parallel pages 2011c Springer- Verlag ISBN 978-3-642-23396-8 URL http://dl.acm.org/citation.cfm id=2033408.2033425 Chris Okasaki Catenable double-ended queues In programming, SIGPLAN Functional ACM second conference of pages Proceedings the international on 66–74 ACM Press, 1997. Robert D Blumofe, Christopher F Joerg, Bradley C Kuszmaul, Charles E Leiserson, Keith H Randall, and Yuli Zhou Cilk: efficient An multithreaded runtime system In Parallel pages 207–216, Journal of Computing, Distributed and 1995. Roman Levenstein Java ctrie implementation, 2012 URL https://github.com/romix/java-concurrent-hash-trie-map M Areias and R Rocha A Tabled Logic Concurrent Design Trie Hash for Lock-Free Programs In C Grelck, pages editor, on Applications Amsterdam, High-level Programming Parallel 2014), July Symposium (HLPP Netherlands, 259 278, International 7th and 2014. Thomas H Cormen, Charles E Leiserson, Ronald L Rivest, and Clifford Stein Introduction to Algorithms The edition, MIT 2 Press, 2001. Guy Steele Organizing code or, execution; for and parallel considered slightly foldr foldl functional harmful International Functional Conference (ICFP), on Programming 2009. Aleksandar Prokopec, Martin and Bagwell, Phil Odersky Lock-free resizeable concurrent tries In LCPC, pages 156–170, 2011. Nathan G Bronson, Jared Casper, and Chafi, Hassan Kunle Olukotun A binary concurrent search practical tree SIGPLAN Not 45(5):257–268, January 2010 ISSN 0362-1340. Aleksandar Prokopec Scalameter website, 2014b URL http://scalameter.github.io Nathan G Bronson, and Hassan Chafi, Kunle Olukotun CCSTM: library-based for stm A scala In Annual Days, First at The Scala Scala Workshop 2010. Aleksandar and Bagwell, Bronson, Martin Prokopec, Grasso Phil Nathan Odersky Concurrent tries non-blocking with efficient snapshots pages 151–160, 2012a doi: 10.1145/2145816.2145836 URL http://doi.acm.org/10.1145/2145816.2145836. Bradford L Chamberlain A of Chapel, brief overview 2013. W Litwin, Y Sagiv, and K Vidyasankar Concurrency and trie hashing Acta Inf 26(7):597–614, September 1989 ISSN 0001-5903 doi: 10.1007/BF00288973 URL http://dx.doi.org/10.1007/BF00288973 Daniel Friedman and David Wise The of impact programming on applicative multiprocessing In on Conference Parallel Processing, International 1976. Finally, parts computation to processors, of efficiently or assigning different the scheduling is not trivial Although cache-coherence identical processes a influence protocols CPUs, most the concurrent interference of systems computer hiearchies dissertation writing services us, and with have speed effective memory several CPU Moreover, say that we has others some elements collection some of – require data-parallel a for work operations no than the operation inherently in more data-parallel workload uniform practice This technique presents highly parallelizing work-stealing the irregular called thesis tree for workloads, computation a novel scheduling We optimal outperforms performance workloads, the and work-stealing parallelizing irregular uniform schedulers more retains other highly scheduler when tree parallelizing show that when workloads. Peter Van and Seif Roy Haridi Concepts, Computer of and Techniques, Models Programming MIT Press, 2004 ISBN 0-262-22069-5. Martin Odersky Scala 2.8 collections Technical report, Lausanne, November EPFL, 2009. Susan Lawrence Schonberg, Flynn and Edith Hummel, E Flynn Factoring: a scheduling for parallel method loops Commun ACM, 35(8):90–101, August 1992 ISSN 0001-0782 doi:10.1145/135226.135232 URL http://doi.acm.org/10.1145/135226.135232. Concurrent operations this structure and in algorithms and linearizable are thesis data lock-free We detailed structures their data showing their and correctness linearization present for identifying pseudocode points proofs concurrent correctness, in validating algorithms this and with thesis, lock-freedom. Stephen Toub Patterns of parallel programming, 2010. Cliff Click Towards a scalable non-blocking style, coding 2007 URL http://www.azulsystems.com/events/javaone_2007/2007_LockFreeHash.pdf Faith Panagiota Fatourou, van Ellen, Franck Eric Ruppert, and Breugel Non-blocking binary search trees In 29th Symposium ’10, USA, PODC Distributed Computing, Proceedings of New SIGACT-SIGOPS Principles NY, 131–140, York, the pages of on ACM 2010 ACM ISBN 978-1-60558-888-9 doi: 10.1145/1835698.1835736 URL http://doi.acm.org/10.1145/1835698.1835736. Kenneth E Iverson A programming language John Wiley & Sons, Inc New York, NY, USA, 1962 ISBN 0-471430-14-5. Guy Steele How parallel to programming: think about Not Strange Loop Conference, 2010. Antal Ioannis Buss, Nancy Harshvardhan, Papadopoulos, M Amato, and Lawrence Rauchwerger Stapl: adaptive template standard parallel library In Proceedings of the York, Conference, NY, pages Experimental USA, Haifa New 14:1–14:10, Annual 3rd SYSTOR Systems ’10, 2010 ACM ISBN 978-1-60558-908-4 Martin Odersky and Adriaan Moors Fighting rot (experience report: types with Scala bit collections) In K Ravi Kannan and Narayan Doug Lea A java fork/join framework In ’00, NY, Java the York, Grande, New on ACM of 2000 36–43, JAVA Proceedings Conference USA, pages 2000 ACM ISBN 1-58113-288-3 doi: 10.1145/337449.337465 URL http://doi.acm.org/10.1145/337449.337465. Chris Okasaki Purely Functional Data Structures PhD thesis, PA, Pittsburgh, USA, 1996 AAI9813847. The available for PhD thesis is download Symbolic coding Computer theory, for computation in cryptography, advances in number has Mathematics example Science, and underpinned key and theory Computer systems facilitate algebra symbolic software mathematics Developing distinctive to irregular employ these at of algorithms data scale its own and as tend challenges, symbolic control complex has set structures SymGridParII Performance for parallel High a on Computing is parallel massively computing middleware symbolic platforms A called language Haskell domain element (DSL) Distributed key a specific SymGridParII is Haskell of Parallel (HdpH) It generated irregular load designed balance work parallelism, explicitly to employs distributed-memory and task stealing for is scalable dynamically sizes. To implement investigate symbolic reliable version evaluate scalable providing HdpH, design, tolerant a computation and fault of we HdpH-RS Its detects as reliable handles a recovery replication using task faults, and key scheduler strategy The tolerant fault scheduler with a work load stealing balancing supports protocol The work common implicit and invoked two with programming reliable and tolerance tolerant placement, scheduler explicit fault that primitives encapsulate fault is for skeletons parallel parallel 10 patterns The is they user to many handled the are instead failures, oblivious by scheduler. An semantics describes on operational small-step reductions states A simple transitions for abstract evaluation machine task scheduling and is presented It and in the for rules of supervised tasks presence defines futures, the of semantics transition recovering the failure The that fault-free three with transition rules execution, recover executions a from are and demonstrated faults. The fault stealing in been tolerant has Promela to work a abstracted model The exhaustively states model validate search the key SPIN automaton the intersection checker a to to is resiliency property this used of of in protocol It be all the initially an supervisor the full possible combinations of supervised presence on node of empty in that asserts will future eventually failures. As of exponentially, number number failures grow the in of cores is systems manycore to predicted the also grows exponentially Hence tolerate to must be massively parallel computations able faults Moreover parallel the to language to new needed of address approaches system are massively resilience heterogeneous architecture and design architectures.
0 Comments
Leave a Reply. |