SpeedForcealpha

Run faster, launch in SpeedForce!

SpeedForce is a framework for distributed parallel computing. It is developed by using C++ programming language. It's simple, but powerful. It has dozens of instructions set which is very user-friendly. SpeedForce supports a variety of operating systems, like Linux (Ubuntu, Fedora, CentOS etc.), Windows Subsystem for Linux, Android and macOS. It also supports multiple users and multiple tasks.

Speedforce includes three modules: process management, information center, and task management. The process management is mainly used to start or kill processes. It's basically executed in the background. The information center provides the message notification of task submission, file transfer, task execution etc. The task management involves all management related to executable programs, including task execution, information feedback, and task monitoring. In addition, SpeedForce has a few more functions as well. For instance, initialization function could allow users to save significant time on deployment of SpeedForce. Speedforce also can easily add local users, as well as add users on other nodes remotely.

SpeedForce Package

Download

http://speedforce.run/sf.zip

Conventions

Global variables

Deployment

Set up authenticated SSH connection between master node and other nodes for password-less logins
Name the nodes
Initialization

The initialization would generate the programs of the three main modules: process management (sfroot), information center (sfinfo) and task management (sfexec). The remote initialization is performed by process management module. The sfroot need start in the background before remote initialization.

Environment variables
Add users

Instructions

Process management

Information center

Task management

Task submission

There are several types of task, including txt, file, root_file, cmd, exe. Also, there are some additional types of task for batch distributed computing, like bat0, bat1, bat2, bat3.

Batch computing

bat0: multi-nodes execute the same script run.sh; bat1: the master node will execute the script continue .sh when other multi-nodes have completed execution of the script run.sh; bat2: each node executes its own script; bat3: the master node will execute the script continue.sh when other multi-nodes have completed execution of their own scripts.

Other functions

©2020 SpeedForce. Developed by Cheng Wang. The last update time: 16:35, 2020-06-09