senaksm/assignment

Getting Started

Table of Contents

1. About The Project
2. Getting Started
   • Prerequisites
   • Installation

About The Project

This a playground project for paralel programing with cuda.

Request ->

    Calculate a race with 100 participants in parallel -> 
    Speeds of participants are given randomly between 1 and 5 ->
    run calculations once in a second ->
    print all participant positions when there is a winner ->
    after race ended give participant order

Decisions ->

    - Every second all participants speed is added to position
    - After a participant reaches finish line in order keep ordering correct 
      MAX_SPEED added to its position
    - There is 2 flags to check if race is still going and to check if there is 
      anyone that already finished it( to reduce calculation count on cpu side)
    - after race ended participant positions ordered and final results are 
      calculated.

(back to top)

Getting Started

Prerequisites

Local

• CUDA Toolkit
• GCC

Via Colab

• Any Browser

Setup

With google colab

• Referance

- Open https://colab.research.google.com
- Runtime -> Change Runtime Type -> Change to 'GPU' -> Save
- Run  !nvcc --version
- Run  !pip install git+https://github.com/andreinechaev/nvcc4jupyter.git
- Run %load_ext nvcc_plugin
- add %%cu at the beginning of code and you can paste cpp code

(back to top)