CSE561/SCE412 Project2 - Game of Life with Pthread and CUDA

Repository for project2 of CSE561/SCE412 Advanced Computer Architecture @ Ajou University (Spring 2021)

Overview

The goal of this project is to implement the Game of Life using pthread and CUDA. Print the right results and compare the execution time of each implementation by running several sample inputs. Follow the guidelines and refer to the detailed information from the given handout.

Lab Environment

• Ubuntu 18.04
• gcc/g++ version 6.5.0
• CUDA 10.0
• 12 Physical cores (24 hardware threads)
• NVIDIA TITAN X (Pascal)

Build and Run

• We provide Makefile to compile both .cu and .cpp files.
• To compile and run:

$ make

$ ./glife sample_inputs/<file_name> <display> <nprocs> <# of generation> <width> <height>

• If display = 1, the program will dump the index output. If you run on GPU, you need to set nprocs = 0. nprocs = 1 will run by a single thread, and nprocs > 1 will run by multiple threads.
• You need to specify enough number of width and height to avoid segmentation fault errors.
• If you use your own NVIDIA GPU, you need to add compatibility number on Makefile. For example, if you use NVIDIA GeForce RTX 3090 which of the compatibility number is 8.6, you have to add GENCODE_SM86 = -gencode=arch=compute_86,code=\"sm_86,compute_86\" on the first line.

TAs

• Taeklim Kim (limkim4233 at ajou dot ac dot kr)
• Wonkyo Choe (heysid at ajou dot ac dot kr)

If you have any questions, please post on Ajou BB and TAs will give you an answer.