How do the authors improve the performance of the protocol?

The authors also propose a way of improving the performance of the protocol used to maintain sequential consistency by using speculations.

What is the way to avoid race conditions on shared locations?

Shared memory works best when programmers have the ability to use complementary synchronization techniques to avoid race conditions on shared locations.

What are the advantages of using speculations for shared memory?

The speculative total-order communication and the use of speculations for shared memory accesses and locks allow us to provide the user with a sequential consistency model.

What is the proposed direction for the use of new architectural designs?

The latter direction proposes the use of new architectural designs to allow hardware and software collaboration to support a “monitor and recover” programming paradigm [15].

How many threads can run the matrix in parallel?

Even when the authors split the matrix in four on their dual-processor machines, the running time of both parallel and shared memory version is at least twice the running time of one thread.

Open AccessProceedings Article10.1109/CCGRID.2003.1199405

Kernel level speculative DSM

C. Tapus, +2 more

- 12 May 2003

- pp 487-494

TL;DR: This paper describes a speculative total order communication protocol, a speculative sequential consistency model, and a speculative distributed locking mechanism that are supported by a mathematical model showing the advantages of speculative execution over traditional execution.

Abstract: Interprocess communication (IPC) is ubiquitous in today's computing world. One of the simplest mechanisms for IPC is shared memory. We present a system that enhances the System V IPC API to support distributed shared memory (DSM) by using speculations. Speculations provide performance improvements by enabling rollback of overly optimistic speculative executions. This paper describes a speculative total order communication protocol, a speculative sequential consistency model, and a speculative distributed locking mechanism. All these are supported by a mathematical model showing the advantages of speculative execution over traditional execution. Our DSM system is part of the Mojave system, which consists of a compiler and the extensions of the operating system designed to support speculations and process migration. The goal of our system is to provide a simpler programming paradigm for designers of distributed systems.

AI Agents for this Paper

Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps

Most frequently asked questions

1. How many machines are connected to the network?

The cluster consists of 16 machines, each with 700MHz dual processors, running Red Hat 8.0 and the 2.4.18 Linux kernel, and connected through a 100Mbit network.

2. What is the function that keeps track of the previous block?

The Mojave system keeps track of the previous block, and in the event that the speculation is aborted, restores the pointer table entry to point to the previous block.

3. How many threads are used to migrate the binary?

For the specific case of the matrix multiplication program, the binary that the authors migrate has about 200k, so the overhead due to migration is merely 0.7 seconds.

4. What is the meaning of the term "Total order protocol"?

In a total order protocol, when a message M with ID (t1;m1) is received, the message is held in a queue until all messagesM 0 with smaller ID have been delivered to the application.

Figure 2. Algorithm for total-order communication using speculations

Figure 3. Effective computation (in sec)

Citations

Proceedings Article•10.1109/RTCSA.2012.12

Distributed IPC Using Virtual Device Driver in Monolithic Kernel

Susmit Bagchi

- 19 Aug 2012

TL;DR: The proposed architecture and implementation of kernel-level distributed IPC mechanism using device driver framework offering transparency and high performance offer 12.73% ~ 75.84% enhanced performance as compared to other distributed I PC models depending upon data length.

...read moreread less

•Dissertation•10.7907/RBAD-VH06.

High-Confidence, Modular Compiler Development in a Formal Environment

Nathaniel Asoka Gray

- 01 Jan 2005

TL;DR: A methodology for realistic compiler development in an existing formal methods framework where program transformations and analyses are implemented as term rewrites and inference rules, and automated proof search techniques are used to drive the compilation process.

...read moreread less

•Dissertation

Speculative execution by using software transactional memory

Emanuel Amaral Couto

- 01 Jan 2009

TL;DR: Speculative Execution as discussed by the authors ) is an example of speculative execution, where the author proposes speculative execution, and the execution is executed.and the executioner is killed.

...read moreread less

•Dissertation

Speculative execution across layers

Peter M. Chen, +1 more

- 01 Jan 2012

TL;DR: This dissertation shows how visible speculations across layers can be applied to application development, network protocols, and dynamic application analysis by developing an algorithm for parallelizing data race detection by expressing speculative parallelism and handling it in the algorithm itself.

...read moreread less

Distributed Synchronization withSharedSemaphore Sets

Cristian Tapu

- 01 Jan 2005

TL;DR: This paper presents a kernel level distributed interprocess communication library (KDIPC) with support for distributed shared memory and distributed semaphores, and implementation is done at the Linux kernel level to reduce the head induced by the strict consistency model.

...read moreread less

References

•Journal Article•10.1145/75104.75105

Memory coherence in shared virtual memory systems

Kai Li, +1 more

- 01 Nov 1989

- ACM Transactions on Computer Systems

TL;DR: Both theoretical and practical results show that the memory coherence problem can indeed be solved efficiently on a loosely coupled multiprocessor.

...read moreread less

1.3K

•Proceedings Article•10.1145/121132.121159

Implementation and performance of Munin

John B. Carter, +2 more

- 01 Sep 1991

TL;DR: This work evaluates the implementation of Munin and describes the execution of two Munin programs that achieve performance within ten percent of message passing implementations of the same programs.

...read moreread less

772

•Proceedings Article•10.1145/10590.10610

Memory coherence in shared virtual memory systems

Kai Li, +1 more

- 01 Nov 1986

TL;DR: Both theoretical and practical results show that the memory coherence problem can indeed be solved efficiently on a loosely-coupled multiprocessor.

...read moreread less

623

•Dissertation

Lazy release consistency for distributed shared memory

P. Keleher

- 03 Oct 1996

TL;DR: A new class of protocols that has lower communication requirements than previous DSM protocols, and can consequently achieve higher performance, is presented.

...read moreread less

154

•Book

Enhancing software reliability with speculative threads

Jeffrey Oplinger, +1 more

- 01 Jan 2004

TL;DR: This paper advocates the use of a monitor-and-recover programming paradigm to enhance the reliability of software, and proposes an architectural design that allows software and hardware to cooperate in making this paradigm more efficient and easier to program.

...read moreread less

137

Kernel level speculative DSM

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. How many machines are connected to the network?

2. What is the function that keeps track of the previous block?

3. How many threads are used to migrate the binary?

4. What is the meaning of the term "Total order protocol"?

5. How do the authors improve the performance of the protocol?

6. What is the way to avoid race conditions on shared locations?

7. What are the advantages of using speculations for shared memory?

8. What is the proposed direction for the use of new architectural designs?

9. How many threads can run the matrix in parallel?

Figures

Citations

Distributed IPC Using Virtual Device Driver in Monolithic Kernel

High-Confidence, Modular Compiler Development in a Formal Environment

Speculative execution by using software transactional memory

Speculative execution across layers

Distributed Synchronization withSharedSemaphore Sets

References

Memory coherence in shared virtual memory systems

Implementation and performance of Munin

Memory coherence in shared virtual memory systems

Lazy release consistency for distributed shared memory

Enhancing software reliability with speculative threads

Related Papers (5)

Transaction Processing: Concepts and Techniques

Virtual time

JaSPEx: Speculative Parallelization on the Java Platform

TALM: A Hybrid Execution Model with Distributed Speculation Support

Jasmine: A shared-object multi-locking distributed shared memory system for heterogeneous computers