اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدConcurrency and Scalability versus Fragmentation and Compaction with Compact-fit
205
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
We study, formally and experimentally, the trade-off in temporal and spatial overhead when managing contiguous blocks of memory using the explicit, dynamic and real-time heap management system Compact-fit (CF). The key property of CF is that temporal and spatial overhead can be bounded, related, and predicted in constant time through the notion of partial and incremental compaction. Partial compaction determines the maximally tolerated degree of memory fragmentation. Incremental compaction of objects, introduced here, determines the maximal amount of memory involved in any, logically atomic, portion of a compaction operation. We explore CFs potential application space on (1) multiprocessor and multicore systems as well as on (2) memory-constrained uniprocessor systems. For (1), we argue that little or no compaction is likely to avoid the worst case in temporal as well as spatial overhead but also observe that scalability only improves by a constant factor. Scalability can be further improved significantly by reducing overall data sharing through separate instances of Compact-fit. For (2), we observe that incremental compaction can effectively trade-off throughput and memory fragmentation for lower latency.
قيم البحث
اقرأ أيضاً
This volume contains the proceedings of ICE 2013, the 6th Interaction and Concurrency Experience workshop, which was held in Florence, Italy on the 6th of June 2013 as a satellite event of DisCoTec 2013. The ICE procedure for paper selection allows P
C members to interact, anonymously, with authors. During the review phase, each submitted paper is published on a Wiki and associated with a discussion forum whose access is restricted to the authors and to all the PC members not declaring a conflict of interests. The PC members post comments and questions that the authors reply to. Each paper was reviewed by three PC members, and altogether 6 papers were accepted for publication. We were proud to host two invited talks, Davide Sangiorgi and Filippo Bonchi, whose abstracts are included in this volume together with the regular papers. The workshop also featured a brief announcement of an already published paper.
This volume contains the proceedings of ICE20, the 13th Interaction and Concurrency Experience, which was held online on the 19th of June 2020, as a satellite event of DisCoTec20. The ICE workshop series features a distinguishing review and selection
procedure, allowing PC members to interact anonymously with authors. As in the past 12 editions, this interaction considerably improved the accuracy of the feedback from the reviewers and the quality of accepted papers, and offered the basis for lively discussion during the workshop. The 2020 edition of ICE included double blind reviewing of original research papers, in order to increase fairness and avoid bias in reviewing. Each paper was reviewed by three PC members, and altogether 5 papers were accepted for publication - plus 5 oral presentations which are not part of this volume. We were proud to host 2 invited talks, by Cinzia Di Giusto and Karoliina Lehtinen. The abstracts of these talks are included in this volume together with the regular papers. The fin
Co-operative and pre-emptive scheduling are usually considered to be complementary models of threading. In the case of virtual machines, we show that they can be unified using a single concept, the bounded execution of a thread of control, essentiall
y providing a first-class representation of a computation as it is reduced. Furthermore this technique can be used to surface the thread scheduler of a language into the language itself, allowing programs to provide their own schedulers without any additional support in the virtual machine, and allowing the same virtual machine to support different thread models simultaneously and without re-compilation.
The process of social expansion in Europe can be better understood with various concepts related to complexity science. Findings of exploratory research show a typical process of social expansion in Europe within the period 1495-1945, in which wars h
ave been instrumental. Furthermore, this research enables the identification of vulnerabilities, and conditions for success in a process of social expansion.
We demonstrate that general-purpose memory allocation involving many threads on many cores can be done with high performance, multicore scalability, and low memory consumption. For this purpose, we have designed and implemented scalloc, a concurrent
allocator that generally performs and scales in our experiments better than other allocators while using less memory, and is still competitive otherwise. The main ideas behind the design of scalloc are: uniform treatment of small and big objects through so-called virtual spans, efficiently and effectively reclaiming free memory through fast and scalable global data structures, and constant-time (modulo synchronization) allocation and deallocation operations that trade off memory reuse and spatial locality without being subject to false sharing.
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
