Topic
Transactional memory
About: Transactional memory is a research topic. Over the lifetime, 2365 publications have been published within this topic receiving 60818 citations.
Papers published on a yearly basis
Papers
More filters
•
30 Jan 2001TL;DR: In this paper, the authors extend the programming model of OTS by providing a unique model that offers both flexibility and ease of use, separating the transactional behavior of CORBA method from the IDL interface.
Abstract: The present invention extends the programming model of OTS by providing a unique model that offers both flexibility and ease of use. This model separates the transactional behavior of CORBA method from the IDL interface. The transactional behavior of the CORBA method is specified in a deployment descriptor file. Each method is associated with a transactional policy. The server reads the policies of the methods during deployment time and makes decisions of making the method transactional based on the policy. Changing the transactional policy of a method is as easy as modifying the deployment descriptor and redeploying the server. If either of the two usage models that OTS offers is used, making a method transactional means that the IDL interface has to change, causing all the software components in the system to re-compile. With the present invention, such a change can be accomplished without need for such recompilation.
9 citations
••
15 May 2014TL;DR: A TM system that executes transactions without ever causing any aborts is presented, which uses a set of t-var lists, one for each transactional variable.
Abstract: We present a TM system that executes transactions without ever causing any aborts. The system uses a set of t-var lists, one for each transactional variable. The instructions of each transaction are placed in the appropriate t-var lists based on which t-variable each of them accesses. A set of worker threads are responsible to execute these instructions. Because of the way instructions are inserted in and removed from the lists, by the way the worker threads work, and by the fact that all the instructions of a transaction are placed in the appropriate t-var lists before doing so for the instructions of any subsequent transaction, it follows that no conflict will ever occur. Parallelism is fine-grained since it is achieved at the level of transactional instructions instead of transactions themselves (i.e., the instructions of a transaction may be executed concurrently).
9 citations
01 Jan 2006
TL;DR: ATLAS is an FPGA-based system that primarily serves as a rapid software development platform for the authors' transactional memory model, TCC, and is poised to support research exploring the impact of transactionalMemory on architecture, operating systems, compilers and programming models.
Abstract: At WARFP 2005, we proposed ATLAS as a scalable implementation for transactional parallel systems [5] The impetus for the development of ATLAS is to address the significant hurdles that software simulators face in multiprocessor architectural research In particular, ATLAS is an FPGA-based system that primarily serves as a rapid software development platform for our transactional memory model, TCC [4] Leveraging commodity hardware and software tools, ATLAS is poised to support research exploring the impact of transactional memory on architecture, operating systems, compilers and programming models
9 citations
••
TL;DR: This paper shows how transactional semantics into Smalltalk is introduced by using the reflective facilities of the language, based on method annotations, incremental parse tree transformations and an optimistic commit protocol.
9 citations
••
14 Apr 2008TL;DR: This work implemented the SHIM communication model in the Haskell functional language, which supports asynchronous communication and transactional memory, and describes two implementations of the model in Haskell, demonstrating the ease of writing such a library.
Abstract: The advent of multicore processors requires mainstream concurrent programming languages with high level concurrency constructs and effective debugging techniques. Unfortunately, many concurrent programming languages are non-deterministic and allow data races. We present a deterministic concurrent communication library for an existing multi-threaded language. We implemented the SHIM communication model in the Haskell functional language, which supports asynchronous communication and transactional memory. The SHIM model uses multi-way rendezvous to guarantee determinism. We describe two implementations of the model in Haskell, demonstrating the ease of writing such a library. We illustrate our library with examples and experimentally compare two implementations. We also compare our new model with equivalent sequential programs and parallel versions using Haskell's existing concurrency mechanisms.
9 citations