MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER

Fluid simulation requires a high number of particles with stable, high framerate in order to make it look realistic and seamless. This final project talks about the implementation of fluid simulation on Graphics Processing Unit (GPU) with compute shader in order to increase the performance of the si...

Full description

Saved in:
Bibliographic Details
Main Author: JONATHAN (NIM : 13513044), CLIFF
Format: Final Project
Language:Indonesia
Online Access:https://digilib.itb.ac.id/gdl/view/21537
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Institut Teknologi Bandung
Language: Indonesia
id id-itb.:21537
spelling id-itb.:215372017-10-09T10:28:06ZMATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER JONATHAN (NIM : 13513044), CLIFF Indonesia Final Project INSTITUT TEKNOLOGI BANDUNG https://digilib.itb.ac.id/gdl/view/21537 Fluid simulation requires a high number of particles with stable, high framerate in order to make it look realistic and seamless. This final project talks about the implementation of fluid simulation on Graphics Processing Unit (GPU) with compute shader in order to increase the performance of the simulation. <br /> <br /> <br /> The simulation uses Material Point Method (MPM) so the simulated fluid material can be easily adjusted to need. The implementation on GPU was done by parallelizing the fluid simulation code and compute shader will execute the parallelized simulation code. The parallelization was done by dividing the simulation into six phases that are going to be executed in serial the whole simulation, but each of them will be run in parallel. <br /> <br /> <br /> The implementation of MPM fluid simulation is done with Cinder library with C++ language. The simulation code is written on six different compute shaders based on the phases that are going to be run in parallel. The test result of the fluid simulation on GPU with compute shader shows significance improvement in the performance of the fluid simulation. The test also shows that the optimal number of work groups differs between GPUs. <br /> <br /> <br /> MPM fluid simulation succeeded to be implemented on GPU with compute shader by dividing the simulation to six phases. The result of the implementation also proved the significant improvement in performance. The optimal number of work groups is different for every GPUs, so consideration is needed when choosing the number of work groups. text
institution Institut Teknologi Bandung
building Institut Teknologi Bandung Library
continent Asia
country Indonesia
Indonesia
content_provider Institut Teknologi Bandung
collection Digital ITB
language Indonesia
description Fluid simulation requires a high number of particles with stable, high framerate in order to make it look realistic and seamless. This final project talks about the implementation of fluid simulation on Graphics Processing Unit (GPU) with compute shader in order to increase the performance of the simulation. <br /> <br /> <br /> The simulation uses Material Point Method (MPM) so the simulated fluid material can be easily adjusted to need. The implementation on GPU was done by parallelizing the fluid simulation code and compute shader will execute the parallelized simulation code. The parallelization was done by dividing the simulation into six phases that are going to be executed in serial the whole simulation, but each of them will be run in parallel. <br /> <br /> <br /> The implementation of MPM fluid simulation is done with Cinder library with C++ language. The simulation code is written on six different compute shaders based on the phases that are going to be run in parallel. The test result of the fluid simulation on GPU with compute shader shows significance improvement in the performance of the fluid simulation. The test also shows that the optimal number of work groups differs between GPUs. <br /> <br /> <br /> MPM fluid simulation succeeded to be implemented on GPU with compute shader by dividing the simulation to six phases. The result of the implementation also proved the significant improvement in performance. The optimal number of work groups is different for every GPUs, so consideration is needed when choosing the number of work groups.
format Final Project
author JONATHAN (NIM : 13513044), CLIFF
spellingShingle JONATHAN (NIM : 13513044), CLIFF
MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER
author_facet JONATHAN (NIM : 13513044), CLIFF
author_sort JONATHAN (NIM : 13513044), CLIFF
title MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER
title_short MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER
title_full MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER
title_fullStr MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER
title_full_unstemmed MATERIAL POINT METHOD BASED FLUID SIMULATION ON GRAPHICS PROCESSING UNIT WITH COMPUTE SHADER
title_sort material point method based fluid simulation on graphics processing unit with compute shader
url https://digilib.itb.ac.id/gdl/view/21537
_version_ 1822019531222024192