作者: 逃避現(xiàn)實 時間: 2025-3-21 23:32 作者: 分散 時間: 2025-3-22 01:06 作者: Musculoskeletal 時間: 2025-3-22 04:32
A new array operation,.Although the method presented here is useful for an important class of problems, it is not the best solution for all problems involving arrays. Further experience is needed to assess its advantages and shortcomings.作者: pacific 時間: 2025-3-22 12:18
https://doi.org/10.1007/978-3-8274-2782-3 empirical fashion. In this paper, we discuss the motivation for the effort, describe the basic elements of the implementation, and provide some preliminary insight distilled from our experience with an initial version of the run-time system.作者: 忘川河 時間: 2025-3-22 14:01 作者: 忘川河 時間: 2025-3-22 17:22
https://doi.org/10.1007/978-3-322-91594-8n Id. I-structures make the language non-functional, but do not raise determinacy issues. Finally, we show that even in the context of purely functional languages, I-structures are invaluable for implementing functional data abstractions.作者: Blasphemy 時間: 2025-3-22 21:19
Parallel graph reduction on a supercomputer: A status report, empirical fashion. In this paper, we discuss the motivation for the effort, describe the basic elements of the implementation, and provide some preliminary insight distilled from our experience with an initial version of the run-time system.作者: certain 時間: 2025-3-23 02:43 作者: 不遵守 時間: 2025-3-23 07:44
I-Structures: Data structures for parallel computing,n Id. I-structures make the language non-functional, but do not raise determinacy issues. Finally, we show that even in the context of purely functional languages, I-structures are invaluable for implementing functional data abstractions.作者: Aesthete 時間: 2025-3-23 10:42
https://doi.org/10.1007/978-3-662-52707-8e exploitable parallelism in a program. The run-time system supports dynamic scheduling, interprocessor communication, and storage management. A number of statistics gathered during execution are presented.作者: 上下連貫 時間: 2025-3-23 14:20
https://doi.org/10.1007/978-3-663-12240-1 typical array-assignment patterns. The syntax falls somewhere in between the purely functional and (im)purely imperative, and makes the inferencing problem fairly easy. If nothing else, the work represents an interesting intellectual excercise in the relationship between non-determinism, side-effects, and parallelism.作者: BRIEF 時間: 2025-3-23 19:39 作者: 來自于 時間: 2025-3-24 01:52
Arrays, non-determinism, side-effects, and parallelism: A functional perspective, typical array-assignment patterns. The syntax falls somewhere in between the purely functional and (im)purely imperative, and makes the inferencing problem fairly easy. If nothing else, the work represents an interesting intellectual excercise in the relationship between non-determinism, side-effects, and parallelism.作者: Tartar 時間: 2025-3-24 04:20
A flexible architectural study methodology,odel under study. The trace of each program execution is then reduced to an “architecturally neutral” precedence graph. The precedence graph can then be used repeatedly in simulations to study the effects of changes in architecture or scheduling strategy.作者: 酷熱 時間: 2025-3-24 10:27
https://doi.org/10.1007/978-3-663-04922-7odel under study. The trace of each program execution is then reduced to an “architecturally neutral” precedence graph. The precedence graph can then be used repeatedly in simulations to study the effects of changes in architecture or scheduling strategy.作者: 兒童 時間: 2025-3-24 13:40
On the correspondence of lambda style reduction and combinator style reduction,echanisms thus avoiding as many of the usual implementation details as possible. It will be shown that there is a very strong correspondence between the two reduction mechanisms. This correspondence offers a basis for a fair comparison of the complexities of both mechanisms.作者: dysphagia 時間: 2025-3-24 15:56
Head order reduction: A graph reduction scheme for the operational lambda calculus,rm of a lambda expression corresponds closely to K. J. Greene‘s lazy normal form of his LNF-calculus, that is, the head order normal form will have the same variable (respectively, constant) as head and the same number of arguments as its normal form. In the context of future implementations, the he作者: Retrieval 時間: 2025-3-24 19:42 作者: 新陳代謝 時間: 2025-3-25 00:38
Concurrent term rewriting as a model of computation,anced massively concurrent architectures. At the highest level of abstraction, this model views computation as replacing selected subterms by others, at multiple sites concurrently. In this view, concurrent term rewriting provides a standard of correctness, and the choice between using trees or grap作者: Ornithologist 時間: 2025-3-25 04:40
Alfalfa: Distributed graph reduction on a hypercube multiprocessor,of both graph reduction and stack oriented execution. Alfalfa consists of two major components, a compiler and a run-time system. The source language, Alfl, contains no constructs that allow the programmer to specify parallelism or synchronization and thus it is the task of the compiler to detect th作者: Encoding 時間: 2025-3-25 09:56
Parallel graph reduction on a supercomputer: A status report,presently augmented by the functional language compiler of the ALFALFA system. Admittedly, parallel graph reduction is hardly a novel idea. The interesting notion is the provision of a parallel execution environment sufficiently powerful to support development of large prototypical scientific and sy作者: NAG 時間: 2025-3-25 14:45
Target code generation from G-machine code,ode generation from G-machine code is performed in a compiler for Lazy ML. We briefly review the G-machine, and the compiler phases preceding the target code generation. Two different target code generation methods are described, as well as essential parts of the run-time system. The first target co作者: altruism 時間: 2025-3-25 18:58 作者: LIMN 時間: 2025-3-25 21:23 作者: Cosmopolitan 時間: 2025-3-26 00:47 作者: 幾何學(xué)家 時間: 2025-3-26 08:12 作者: Orchiectomy 時間: 2025-3-26 09:30
Parallel garbage collection for graph machines,ve, because only the resident data items are scanned, irrespective of the amount of real activity. Parallel mark-sweep algorithms are also unacceptable because they require very fine-grained synchronization with the mutator, therefore slowing down the computational process..We have presented a novel作者: impale 時間: 2025-3-26 14:54 作者: ENACT 時間: 2025-3-26 18:59
A flexible architectural study methodology,f example programs are generated by the emulator. The execution method of the emulator exercises all possible parallelism available in the execution model under study. The trace of each program execution is then reduced to an “architecturally neutral” precedence graph. The precedence graph can then 作者: nephritis 時間: 2025-3-27 00:16 作者: crucial 時間: 2025-3-27 03:17 作者: 絕種 時間: 2025-3-27 07:08 作者: Externalize 時間: 2025-3-27 09:49 作者: PLAYS 時間: 2025-3-27 13:55 作者: BROOK 時間: 2025-3-27 20:03 作者: Amorous 時間: 2025-3-27 22:21
https://doi.org/10.1007/978-3-8274-2253-8anced massively concurrent architectures. At the highest level of abstraction, this model views computation as replacing selected subterms by others, at multiple sites concurrently. In this view, concurrent term rewriting provides a standard of correctness, and the choice between using trees or grap作者: 來這真柔軟 時間: 2025-3-28 04:07 作者: 殖民地 時間: 2025-3-28 08:06 作者: pulse-pressure 時間: 2025-3-28 12:04 作者: 欲望 時間: 2025-3-28 18:23
https://doi.org/10.1007/978-3-642-57123-7 the language and by architectural considerations..Perhaps the most important feature of the machine is the load balancing mechanism, that includes the decentralized computation and distribution of an instantaneous load information, allowing tasks to be equally distributed among processors. Furtherm作者: 完成才能戰(zhàn)勝 時間: 2025-3-28 20:05 作者: Neuralgia 時間: 2025-3-29 00:12 作者: 充滿裝飾 時間: 2025-3-29 06:54
Mikrosystemtechnik auf Siliziumuirements, or reduce the total amount of work performed..We show by a series of example that the reduction partial order can be easily controlled by using a combination of call-by-value, call-by-name, and call-by-speculation (an eager form of call-by-need).作者: 大炮 時間: 2025-3-29 09:07
https://doi.org/10.1007/978-3-662-08761-9ve, because only the resident data items are scanned, irrespective of the amount of real activity. Parallel mark-sweep algorithms are also unacceptable because they require very fine-grained synchronization with the mutator, therefore slowing down the computational process..We have presented a novel作者: 針葉類的樹 時間: 2025-3-29 13:03 作者: Vertebra 時間: 2025-3-29 16:13 作者: 注意力集中 時間: 2025-3-29 22:45 作者: Indebted 時間: 2025-3-30 01:24 作者: 驚惶 時間: 2025-3-30 04:08
https://doi.org/10.1007/978-3-322-91594-8trate this through careful analysis of program examples using three common functional data-structuring approaches—lists using Cons and arrays using Update (both fine-grained operators), and arrays using make-array (a “bulk” operator). We then present I-structures as an alternative, defining precisel作者: cushion 時間: 2025-3-30 10:19
https://doi.org/10.1007/978-3-642-50740-3 identity-reduction in-the-large. These in-the-large operations represent a new concept. They take sequences of bindings and/or applications as initial form, and produce a result in-the-large. They comprise a multitude of single reductions (., beta, eta) occurring in a specific pattern..A geometric,作者: 蟄伏 時間: 2025-3-30 13:45
https://doi.org/10.1007/978-3-8274-2253-8writing its own fragment. A number of implementation and optimization issues are also discussed, including overlapping rewrites, rule ordering, compilation, and flow analysis. Concurrent E-strategies are introduced as a flexible control mechanism to optimize performance and facilitate systems progra作者: Inflammation 時間: 2025-3-30 16:43
https://doi.org/10.1007/978-3-322-84878-9te parallel machine can be built to realise this abstract architecture, and describe the design of an experimental prototype machine. This prototype is a hardware emulator of the ideal concrete architecture which is to be built in customised VLSI, and is constructed using the INMOS Transputer as the作者: 悅耳 時間: 2025-3-30 21:13
https://doi.org/10.1007/978-3-8351-9022-1duction strategies yet..In this paper three formal methods for specifying reduction strategies in TRS‘s are presented. In the first method the reduction strategy is encoded in the TRS itself. The original TRS is transformed to a so called annotation TRS in which the strategy is encoded using functio作者: 植物茂盛 時間: 2025-3-31 04:09 作者: neoplasm 時間: 2025-3-31 07:16
Performance measurement of a G-machine implementation,作者: 裝飾 時間: 2025-3-31 10:01 作者: Blemish 時間: 2025-3-31 14:35
Concurrent term rewriting as a model of computation,writing its own fragment. A number of implementation and optimization issues are also discussed, including overlapping rewrites, rule ordering, compilation, and flow analysis. Concurrent E-strategies are introduced as a flexible control mechanism to optimize performance and facilitate systems progra作者: SLAG 時間: 2025-3-31 21:03 作者: Analogy 時間: 2025-3-31 22:44
Specification of reduction strategies in term rewriting systems,duction strategies yet..In this paper three formal methods for specifying reduction strategies in TRS‘s are presented. In the first method the reduction strategy is encoded in the TRS itself. The original TRS is transformed to a so called annotation TRS in which the strategy is encoded using functio作者: Initial 時間: 2025-4-1 04:51 作者: 過去分詞 時間: 2025-4-1 06:36
https://doi.org/10.1007/978-3-642-99211-7echanisms thus avoiding as many of the usual implementation details as possible. It will be shown that there is a very strong correspondence between the two reduction mechanisms. This correspondence offers a basis for a fair comparison of the complexities of both mechanisms.作者: interrupt 時間: 2025-4-1 11:35 作者: hair-bulb 時間: 2025-4-1 15:20
Mikrosystemtechnik auf Siliziumuirements, or reduce the total amount of work performed..We show by a series of example that the reduction partial order can be easily controlled by using a combination of call-by-value, call-by-name, and call-by-speculation (an eager form of call-by-need).作者: 燦爛 時間: 2025-4-1 19:25
