Giter Club home page Giter Club logo

libdsmu's Introduction

libdsmu

Distributed Shared Memory in Userspace Library for Linux Applications. Modeled after the Ivy system. Final project for 6.824.

Authors

  • Webb Horn
  • Tim Donegan
  • Julian Gonzalez
  • Ameesh Goyal

Setup

libdsmu was tested on Ubuntu 14.04 x86/64 running the 3.13.0-24-generic kernel. Older kernels on a non-x86/64 machine probably will not work because the code that detects fault types is non-portable.

Third party libraries:

$ sudo apt-get install libb64-dev

Build

$ cd src
$ make

Running the manager

$ python manager/manager.py &

Running the Matrix Multiply Benchmark

matrixmultiply uses the alternation-style sharding algorithm for nodes-to-rows mapping, while matrixmultiply2 allocates rows to nodes in contiguous chunks to reduce the effects of page thrashing. They are both launched the same way.

First, start the manager:

$ python manager/manager.py &

Next, start up instances of matrixultiply in parallel. To start three instances, launch them like this:

$ cd src
$ make
$ ./matrixmultiply2 127.0.0.1 4444 1 3 && \
  ./matrixmultiply2 127.0.0.1 4444 2 3 && \
  ./matrixmultiply2 127.0.0.1 4444 3 3:

libdsmu's People

Contributors

ameeshg avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar

libdsmu's Issues

Test programs do not exist.

We should at least write a matrix multiplication program.

We should record time required to do the multiplication on multiple machines with 1..n machines, and plot the times to determine the actual speedup.

We need this for the writeup.

Errors should be #define'd.

Right now we just randomly return negative ints, and don't handle most of them.

We should give them meaningful names and handle each one appropriately eventually.

Matrixmultiply sometime hangs

Unclear if the problem is in client or manager, but in any case, progress halts.

To reproduce:

  1. Set size of matrix to 300 in matrixmultiply.c
  2. Start manager
  3. Recompile
  4. ./matrixmultiply 127.0.0.1 4444 1 1

matrixmultiply will hang, and not always at the same page.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.