Transactional data structure libraries

Spiegelman, Alexander; Golan-Gueta, Guy; Keidar, Idit

doi:10.1145/2908080.2908112

Cited by 34 publications

(34 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast, our benchmark's long transactions result in high abort rates in the absence of nesting. Nesting the log writes (green squares) improves throughput by an additional factor of up to 6, which is in line with the improvement of TDSL over TL2 reported in [47], and also reduces the abort rate by a factor of 2. The packet map is not contended in this experiment, and so transactions with nested insertion to the map behave similarly to flat ones (in terms of both throughput and abort rate).…”

Section: Resultssupporting

confidence: 83%

“…For flat transactions (red diamonds), TDSL's throughput is consistently double that of TL2 (purple octagons), as can be observed in Figure 5, which zooms in on these two curves in the same experiment. We note that the TDSL work [47] reported better performance improvements over TL2, but they ran shorter transactions that did not write to a contended log at the end, where TDSL's abort rate remained low. In contrast, our benchmark's long transactions result in high abort rates in the absence of nesting.…”

Section: Resultsmentioning

confidence: 90%

“…To mitigate this tradeoff, Spiegelman et al [47] have proposed transactional data structure libraries (TDSL). In a nutshell, the idea is to trade generality for performance.…”

Section: Transactional Librariesmentioning

confidence: 99%

“…A TDSL can mange aborts on a semantic level, e.g., two concurrent transactions can simultaneously change two different locations in the same list without aborting. While the original TDSL library [47] was written in C++, we implement our version in Java. We offer more background on TDSL in Section 2.…”

Section: Transactional Librariesmentioning

confidence: 99%

“…Our algorithm builds on ideas used in TL2 [6], which is a generic STM framework, and in TDSL [47], which suggests forgoing generality for increased efficiency. We briefly overview their modus operandi as background for our work.…”

Section: A Walk Down Transactional Data Structure Lanementioning

confidence: 99%

See 4 more Smart Citations

Nesting and composition in transactional data structure libraries

Assa

Meir²,

Golan-Gueta

et al. 2020

Proceedings of the 25th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

Self Cite

View full text Add to dashboard Cite

Transactional data structure libraries (TDSL) combine the ease-of-programming of transactions with the high performance and scalability of custom-tailored concurrent data structures. They can be very efficient thanks to their ability to exploit data structure semantics in order to reduce overhead, aborts, and wasted work compared to generalpurpose software transactional memory. However, TDSLs were not previously used for complex use-cases involving long transactions and a variety of data structures.In this paper, we boost the performance and usability of a TDSL, allowing it to support complex applications. A key idea is nesting. Nested transactions create checkpoints within a longer transaction, so as to limit the scope of abort, without changing the semantics of the original transaction. We build a Java TDSL with built-in support for nesting in a number of data structures. We conduct a case study of a complex network intrusion detection system that invests a significant amount of work to process each packet. Our study shows that our library outperforms TL2 twofold without nesting, and by up to 16x when nesting is used. Finally, we discuss crosslibrary nesting, namely dynamic composition of transactions from multiple libraries.

show abstract

Section: Resultssupporting

confidence: 83%

Section: Resultsmentioning

confidence: 90%

“…To mitigate this tradeoff, Spiegelman et al [47] have proposed transactional data structure libraries (TDSL). In a nutshell, the idea is to trade generality for performance.…”

Section: Transactional Librariesmentioning

confidence: 99%

Section: Transactional Librariesmentioning

confidence: 99%

Section: A Walk Down Transactional Data Structure Lanementioning

confidence: 99%

See 3 more Smart Citations

Nesting and composition in transactional data structure libraries

Assa

Meir²,

Golan-Gueta

et al. 2020

Proceedings of the 25th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming

Self Cite

View full text Add to dashboard Cite

show abstract

Dynamic Transactional Transformation

LaBorde

Lebanoff

Peterson

et al. 2020

Concurrency and Computation

View full text Add to dashboard Cite

Transactional data structures support threads executing a sequence of operations atomically. Dynamic transactions allow operands to be generated on the y and allows threads to execute code in between the operations of a transaction, in contrast to static transactions which need to know the operands in advance. A framework called Lock-free Transactional Transformation (LFTT) allows data structures to run high-performance transactions, but it only supports static transactions. We present Dynamic Transactional Transformation, an extension to LFTT to add support for dynamic transactions and wait-free progress while retaining its speed. The thread-helping scheme of LFTT presents a unique challenge to dynamic transactions. We overcome this challenge by changing the input of LFTT from a list of operations to a function, forcing helping threads to always start at the beginning of the transaction, and allowing threads to skip completed operations through the use of a list of return values. We thoroughly evaluate the performance impact of support for dynamic transactions and wait-free progress and nd that these features do not hurt the performance of LFTT for our test cases. LaBorde et alThe straightforward way to implement a transactional data structure from a sequential container is to use software transactional memory (STM) 11 . An STM instruments memory accesses by recording the locations a thread reads in a read set, and the locations it writes in a write set.If the read/write sets of different transactions overlap, only one transaction is allowed to commit, the concurrent transactions are aborted and restarted. A drawback of STM is that the runtime system that keeps track of read/write sets and detects conflicts can have a detrimental impact on performance 12 .The inherent disadvantage of STM concurrency control is that low-level memory access conflicts do not necessarily correspond to high-level semantic conflicts. Consider a set implemented as an ordered linked list, where each node has two fields, an integer value and a pointer to the next node.The initial state of the set is {0, 3, 6, 9, 10}. Thread 1 and Thread 2 intend to insert 4 and 1, respectively. Since these two operations commute, it is feasible to execute them concurrently 13 . Commutative data structure operations are those which have no dependencies on each other; reordering them yields the same abstract state of the container. Existing concurrent linked lists employing lock-free or fine-grained locking synchronizations allow concurrent execution of the two operations. Nevertheless, these operations have a read/write conflict and the STM has to abort one of them.An alternative approach called Lock-free Transactional Transformation (LFTT) 14 includes semantic conflict detection that uses information about the data structures and which operations are being executed, to prevent conflicting operations from unnecessarily causing transactions to abort.A significant advantage of using data structure transactions is that this semantic information is availabl...

show abstract

Efficient Means of Achieving Composability Using Object Based Semantics in Transactional Memory Systems

2019

View full text Add to dashboard Cite

Composing together the individual atomic methods of concurrent data-structures (cds) pose multiple design and consistency challenges. In this context composition provided by transactions in software transaction memory (STM) can be handy. However, most of the STMs offer read/write primitives to access shared cds. These read/write primitives result in unnecessary aborts. Instead, semantically rich higher-level methods of the underlying cds like lookup, insert or delete (in case of hash-table or lists) aid in ignoring unimportant lower level read/write conflicts and allow better concurrency.In this paper, we adapt transaction tree model in databases to propose OSTM which enables efficient composition in cds. We extend the traditional notion of conflicts and legality to higher level methods of cds using STMs and lay down detailed correctness proof to show that it is co-opaque. We implement OSTM with concurrent closed addressed hash-table (HT-OSTM) and list (list-OSTM) which exports the higher-level operations as transaction interface.In our experiments with varying workloads and randomly generated transaction operations, HT-OSTM shows speedup of 3 to 6 times and w.r.t aborts HT-OSTM is 3 to 7 times better than ESTM and read/write based STM, respectively. Where as, list-OSTM outperforms state of the art lock-free transactional list, NOrec STM list and boosted list by 30% to 80% across all workloads and scenarios. Further, list-OSTM incurred negligible aborts in comparison to other techniques considered in the paper. * A preliminary version of this work was accepted in AADDA 2018 as work in progress. † Author sequence follows lexical order of last names.

show abstract

Transactional data structure libraries

Cited by 34 publications

References 34 publications

Nesting and composition in transactional data structure libraries

Nesting and composition in transactional data structure libraries

Dynamic Transactional Transformation

Efficient Means of Achieving Composability Using Object Based Semantics in Transactional Memory Systems

Contact Info

Product

Resources

About