AccelerateRT

Explore acceleration structures (BVHs) for ray tracing.
Final project for Duke University CS634: Geometric Algorithms.

---

Project Setup

First make sure you have Julia language installed.
Then install project dependencies locally from command line:

>> julia
               _
   _       _ _(_)_     |  Documentation: https://docs.julialang.org
  (_)     | (_) (_)    |
   _ _   _| |_  __ _   |  Type "?" for help, "]?" for Pkg help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 1.7.0 (2021-11-30)
 _/ |\__'_|_|_|\__'_|  |  Official https://julialang.org/ release
|__/                   |

julia> ]
(@v1.7) pkg> activate .
(AccelerateRT) pkg> instantiate
julia> exit()

---

BVH Construction

Commands

>> julia .\construct.jl -h
usage: construct.jl [--algMiddle] [--algMedian] [--algSAH] [--algSAHM]
                    [--save] [--skip] [--show] [-h] model

positional arguments:
  model        obj model path

optional arguments:
  --algMiddle  construct simple BVH with middle criteria
  --algMedian  construct simple BVH with media criteria
  --algSAH     construct BVH with SAH
  --algSAHM    construct BVH with modifed SAH (new algorithm)
  --save       save constructed structure
  --skip       skip existing structure
  --show       print structure
  -h, --help   show this help message and exit

For example, following command constructs simple BVH with median criteria for bunny model.
Saving the computed structure at the end into structures folder.

>> julia .\construct.jl .\models\bunny\bunny.obj --algMedian --save

---

Visualization

Commands

>> julia .\visualize.jl -h
usage: visualize.jl [--bvh BVH] [-h] model

positional arguments:
  model       obj model path

optional arguments:
  --bvh BVH   constructed BVH path
  -h, --help  show this help message and exit

For example, following command visualizes teapot model in an interactive environment.

>> julia .\visualize.jl .\models\teapot\teapot.obj

The following command visualizes bunny model and its precomputed BVH structure.

>> julia .\visualize.jl .\models\bunny\bunny.obj --bvh .\structures\__models_bunny_bunny_obj.median.jld2

---

Benchmark

Commands

>> julia .\benchmark.jl -h
usage: benchmark.jl [--seed SEED] [--samples SAMPLES]
                    [--resolution RESOLUTION] [--skip] [--time] [-h]

optional arguments:
  --seed SEED           random seed (type: Int64, default: 0)
  --samples SAMPLES     number of random samples (type: Int64,     
                        default: 100)
  --resolution RESOLUTION
                        screen resolution (in format WxH) (default:
                        "50x50")
  --skip                whether to skip existing benchmark caches  
  --time                display time after all tasks done
  -h, --help            show this help message and exit

For example, the following executes benchmark with 4 threads available for each worker, 200 random samples for each configuration, and random seed set to 123, and display time info after all tasks done:

>> julia -t 4 .\benchmark.jl --samples 200 --seed 123 --time

By default, benchmark.jl distributes workers to all cores available. Modify ParallelCores variable to change this behavior: 0 for all cores, 1 for single-process, else customized number of workers.

More info can be found in benchmark_notebook.jl notebook, or by opening benchmark_notebook.jl.pdf file.
To execute it:

>> julia
julia> import Pluto
julia> Pluto.run()

And open benchmark_notebook.jl in browser.
To see the code only, open benchmark_notebook.jl in a text editor.