CIS565-Fall-2021 · Scoutydren · Oct 2, 2021 · Oct 2, 2021 · Oct 3, 2021 · Oct 3, 2021
diff --git a/README.md b/README.md
@@ -1,13 +1,110 @@
 CUDA Path Tracer
 ================
 
-**University of Pennsylvania, CIS 565: GPU Programming and Architecture, Project 3**
+**University of Pennsylvania, CIS 565: GPU Programming and Architecture, Project 2**
 
-* (TODO) YOUR NAME HERE
-* Tested on: (TODO) Windows 22, i7-2222 @ 2.22GHz 22GB, GTX 222 222MB (Moore 2222 Lab)
+* **Name:**  Beini Gu
 
-### (TODO: Your README)
+  * [LinkedIn](https://www.linkedin.com/in/rgu/), [personal website](https://www.seas.upenn.edu/~gubeini/), [twitter](https://twitter.com/scoutydren), etc.
+* Tested on: Windows 10, AMD Ryzen 7 3700X 8-Core Processor 3.59 GHz 16GB, GTX 2060
 
-*DO NOT* leave the README to the last minute! It is a crucial part of the
-project, and we will not be able to grade you without a good README.
+## Project Description
+
+This is a CUDA-based path tracer with global illumination able to render scenes utilizing the parallel computing power of GPU. 
+
+![bunny](img/main.png)
+
+## Table of Contents
+
+[Features Overview](#overview-features)   
+[Miscellaneous](#misc)   
+[Performance Analysis](#performance-analysis)   
+[Bloopers](#bloopers)
+
+[Reference](#Reference)
+
+<a name="overview-features"/> 
+
+## Features Overview
+
+#### Obj Mesh Loading
+
+I use the [tinyObj](https://github.com/tinyobjloader/tinyobjloader) library for parsing and loading the custom triangulated meshes in obj format. Using the VBO data to get the position of the triangle vertices, calculate the normals and calculate their intersections with rays. 
+
+To optimize performance and reduce number of ray-triangle intersection tests, I use a bounding box for the mesh which first checks if the ray hits anywhere inside the bounding box volume in order to reduce unnecessary rays which intersects somewhere outside the bounding box. 
+
+
+| Bunny                   | Cow                 |
+| ----------------------- | ------------------- |
+| ![bunny](img/bunny.png) | ![cow](img/cow.png) |
+
+#### Refraction
+
+I implement refractive materials using Schlick's approximation and Snell's Law. This is how it looks
+
+* Glass caustics on the floor (IOR = 1.52)
+
+![no_dof](img/refraction.png)
+
+#### Depth of Field
+
+I implement a togglable depth of field effect which simulates a realistic camera with focal distance and lens radius. 
+
+| Without Depth of Field    | With Depth of Field (Focal distance = 10.0) | With Depth of Field (Focal Distance = 5.0) |
+| ------------------------- | ------------------------------------------- | ------------------------------------------ |
+| ![no_dof](img/no_dof.png) | ![dof_10](img/dof_10.png)                   | ![dof_10](img/dof_5.png)                   |
+
+#### Anti-aliasing
+
+| Without Anti-Aliasing                    | With Anti-Aliasing                         |
+| ---------------------------------------- | ------------------------------------------ |
+| ![no_dof](img/no_anti_aliasing_demo.png) | ![no_dof](img/with_anti_aliasing_demo.png) |
+
+#### Stratified and Halton Sampling
+
+I implemented two other hemisphere sampling methods: stratified and Halton sampling. Under larger numbers of iterations, there won't be a lot of difference. 
+
+| Random Sampling (iters = 20) | Stratified Sampling (iters = 20) | Halton Sampling (iters = 20) |
+| ---------------------------- | -------------------------------- | ---------------------------- |
+| ![no_dof](img/random.png)    | ![no_dof](img/stratified.png)    | ![no_dof](img/halton.png)    |
+
+Halton sampling is quasi-random number generation, it leaves a pattern on the render when sample number is small. (The above image of Halton Sampling uses 1000 as the sample number, while stratified sampling uses 100 as the sample number. )
+
+| Halton (Sequence length = 100) | Halton (Sequence length = 500) | Halton (Sequence length = 1000) |
+| ------------------------------ | ------------------------------ | ------------------------------- |
+| ![no_dof](img/halton_100.png)  | ![no_dof](img/halton_500.png)  | ![no_dof](img/halton_1000.png)  |
+
+<a name="misc"/> 
+
+## Miscellaneous
+
+I implemented material sort using thrust to improve the performance of path tracer. In addition, there is a togglable option for caching the first intersections (does not work together with anti-aliasing). 
+
+<a name="performance-analysis"/> 
+## Performance Analysis
+
+#### Mesh Culling
+
+Applying mesh culling gives a performance advantage compared to no bounding volume checks. 
+
+The following performance is tested rendering for the profile main scene. It's a 1.6 times performance increase. 
+
+![boundcheck1](img/boundcheck_comp.png)
+
+For the mesh only scene of the bunny render, the performance is also measured. We can see the advantage is less significant. Since in the previous render, the rabbit is smaller in the entire render, using a bounding box to bound that smaller region gives much more advantage. 
+
+![Snipaste_2021-10-06_13-04-02](img/boundcheck2.png)
+
+<a name="blooper"/> 
+
+## Bloopers
+
+| Normal Error & Distance calculation error                    | Refraction calculation error                                 | Randomness error                                             |
+| ------------------------------------------------------------ | ------------------------------------------------------------ | ------------------------------------------------------------ |
+| ![Snipaste_2021-10-06_13-04-02](img/Snipaste_2021-10-06_13-04-02.png) | ![Snipaste_2021-10-03_18-55-19](img/Snipaste_2021-10-03_18-55-19.png) | ![Snipaste_2021-10-08_21-11-44](img/Snipaste_2021-10-08_21-11-44.png) |
+
+## Reference
+
+* PBRT book [Physically Based Rendering: From Theory to Practice](https://www.pbr-book.org/3ed-2018/contents)
+* en.wikipedia.org/wiki/Halton_sequence
 
diff --git a/img/Snipaste_2021-10-03_17-29-07.png b/img/Snipaste_2021-10-03_17-29-07.png
diff --git a/img/Snipaste_2021-10-03_18-55-19.png b/img/Snipaste_2021-10-03_18-55-19.png
diff --git a/img/Snipaste_2021-10-04_13-40-51.png b/img/Snipaste_2021-10-04_13-40-51.png
diff --git a/img/Snipaste_2021-10-04_13-52-23.png b/img/Snipaste_2021-10-04_13-52-23.png
diff --git a/img/Snipaste_2021-10-04_14-01-59.png b/img/Snipaste_2021-10-04_14-01-59.png
diff --git a/img/Snipaste_2021-10-06_13-00-16.png b/img/Snipaste_2021-10-06_13-00-16.png
diff --git a/img/Snipaste_2021-10-06_13-04-02.png b/img/Snipaste_2021-10-06_13-04-02.png
diff --git a/img/Snipaste_2021-10-06_14-34-43.png b/img/Snipaste_2021-10-06_14-34-43.png
diff --git a/img/Snipaste_2021-10-08_02-46-48.png b/img/Snipaste_2021-10-08_02-46-48.png
diff --git a/img/Snipaste_2021-10-08_21-11-44.png b/img/Snipaste_2021-10-08_21-11-44.png
diff --git a/img/blooper.png b/img/blooper.png
diff --git a/img/blooper2.png b/img/blooper2.png
diff --git a/img/blooper3.png b/img/blooper3.png
diff --git a/img/boundcheck2.png b/img/boundcheck2.png
diff --git a/img/boundcheck_comp.png b/img/boundcheck_comp.png
diff --git a/img/bunny.png b/img/bunny.png
diff --git a/img/cow.png b/img/cow.png
diff --git a/img/dof_10.png b/img/dof_10.png
diff --git a/img/dof_5.png b/img/dof_5.png
diff --git a/img/halton.png b/img/halton.png
diff --git a/img/halton_100.png b/img/halton_100.png
diff --git a/img/halton_1000.png b/img/halton_1000.png
diff --git a/img/halton_500.png b/img/halton_500.png
diff --git a/img/main.png b/img/main.png
diff --git a/img/no_anti_aliasing_demo.png b/img/no_anti_aliasing_demo.png
diff --git a/img/no_antialiasing.png b/img/no_antialiasing.png
diff --git a/img/no_dof.png b/img/no_dof.png
diff --git a/img/random.png b/img/random.png
diff --git a/img/refraction.png b/img/refraction.png
diff --git a/img/stratified.png b/img/stratified.png
diff --git a/img/with_anti_aliasing_demo.png b/img/with_anti_aliasing_demo.png
diff --git a/img/with_antialiasing.png b/img/with_antialiasing.png
diff --git a/scenes/cornell.txt b/scenes/cornell.txt
@@ -48,6 +48,26 @@ REFR        0
 REFRIOR     0
 EMITTANCE   0
 
+//glass
+MATERIAL 5
+RGB         .98 .98 .98
+SPECEX      0
+SPECRGB     .98 .98 .98
+REFL        1
+REFR        1
+REFRIOR     1.52
+EMITTANCE   0
+
+// Specular Purple
+MATERIAL 6
+RGB        0.463, 0.133, 0.647
+SPECEX      0
+SPECRGB    0.463, 0.133, 0.647
+REFL        1
+REFR        0
+REFRIOR     0
+EMITTANCE   0
+
 // Camera
 CAMERA
 RES         800 800
@@ -66,7 +86,7 @@ cube
 material 0
 TRANS       0 10 0
 ROTAT       0 0 0
-SCALE       3 .3 3
+SCALE       5 .5 5
 
 // Floor
 OBJECT 1
@@ -111,7 +131,48 @@ SCALE       .01 10 10
 // Sphere
 OBJECT 6
 sphere
-material 4
+material 3
 TRANS       -1 4 -1
 ROTAT       0 0 0
 SCALE       3 3 3
+
+// Sphere
+OBJECT 7
+sphere
+material 2
+TRANS       1 2 1
+ROTAT       0 0 0
+SCALE       2 2 2
+
+// Sphere
+OBJECT 8
+sphere
+material 2
+TRANS       3 2 4
+ROTAT       0 0 0
+SCALE       4 4 4
+
+
+// Sphere
+OBJECT 9
+sphere
+material 3
+TRANS       -2 2 6
+ROTAT       0 0 0
+SCALE       2 2 2
+
+//Cube
+OBJECT 10
+cube
+material 2
+TRANS 0 0 0
+ROTAT 0 0 0
+SCALE 1 4 1
+
+//Bunny
+OBJECT 11
+custom_obj
+material 6
+TRANS 0 0 3
+ROTAT 0 0 0
+SCALE 20 20 20
diff --git a/scenes/cornell_box.mtl b/scenes/cornell_box.mtl
@@ -0,0 +1,24 @@
+newmtl white
+Ka 0 0 0
+Kd 1 1 1
+Ks 0 0 0
+
+newmtl red
+Ka 0 0 0
+Kd 1 0 0
+Ks 0 0 0
+
+newmtl green
+Ka 0 0 0
+Kd 0 1 0
+Ks 0 0 0
+
+newmtl blue
+Ka 0 0 0
+Kd 0 0 1
+Ks 0 0 0
+
+newmtl light
+Ka 20 20 20
+Kd 1 1 1
+Ks 0 0 0
diff --git a/scenes/custom.txt b/scenes/custom.txt
@@ -0,0 +1,161 @@
+// Emissive material (light)
+MATERIAL 0
+RGB         1 1 1
+SPECEX      0
+SPECRGB     0 0 0
+REFL        0
+REFR        0
+REFRIOR     0
+EMITTANCE   5
+
+// Diffuse white
+MATERIAL 1
+RGB         .98 .98 .98
+SPECEX      0
+SPECRGB     0 0 0
+REFL        0
+REFR        0
+REFRIOR     0
+EMITTANCE   0
+
+// Diffuse red
+MATERIAL 2
+RGB         .85 .35 .35
+SPECEX      0
+SPECRGB     0 0 0
+REFL        0
+REFR        0
+REFRIOR     0
+EMITTANCE   0
+
+// Diffuse green
+MATERIAL 3
+RGB         .35 .85 .35
+SPECEX      0
+SPECRGB     0 0 0
+REFL        0
+REFR        0
+REFRIOR     0
+EMITTANCE   0
+
+// Specular white
+MATERIAL 4
+RGB         .98 .98 .98
+SPECEX      0
+SPECRGB     .98 .98 .98
+REFL        1
+REFR        0
+REFRIOR     0
+EMITTANCE   0
+
+//glass
+MATERIAL 5
+RGB         .98 .98 .98
+SPECEX      0
+SPECRGB     .98 .98 .98
+REFL        1
+REFR        1
+REFRIOR     1.52
+EMITTANCE   0
+
+// Specular Purple
+MATERIAL 6
+RGB        0.463, 0.133, 0.647
+SPECEX      0
+SPECRGB    0.463, 0.133, 0.647
+REFL        1
+REFR        0
+REFRIOR     0
+EMITTANCE   0
+
+// Camera
+CAMERA
+RES         800 800
+FOVY        45
+ITERATIONS  5000
+DEPTH       8
+FILE        cornell
+EYE         0.0 5 10.5
+LOOKAT      0 5 0
+UP          0 1 0
+
+
+// Ceiling light
+OBJECT 0
+cube
+material 0
+TRANS       0 10 0
+ROTAT       0 0 0
+SCALE       4 .4 4
+
+// Floor
+OBJECT 1
+cube
+material 1
+TRANS       0 0 0
+ROTAT       0 0 0
+SCALE       10 .01 10
+
+// Ceiling
+OBJECT 2
+cube
+material 1
+TRANS       0 10 0
+ROTAT       0 0 90
+SCALE       .01 10 10
+
+// Back wall
+OBJECT 3
+cube
+material 1
+TRANS       0 5 -5
+ROTAT       0 90 0
+SCALE       .01 10 10
+
+// Left wall
+OBJECT 4
+cube
+material 2
+TRANS       -5 5 0
+ROTAT       0 0 0
+SCALE       .01 10 10
+
+// Right wall
+OBJECT 5
+cube
+material 3
+TRANS       5 5 0
+ROTAT       0 0 0
+SCALE       .01 10 10
+
+// Sphere
+OBJECT 6
+sphere
+material 5
+TRANS       -1 4 -1
+ROTAT       0 0 0
+SCALE       3 3 3
+
+// Sphere
+OBJECT 7
+sphere
+material 6
+TRANS       1 2 1
+ROTAT       0 0 0
+SCALE       2 2 2
+
+// Sphere
+// OBJECT 8
+// cube
+// material 6
+// TRANS       3 1 2
+// ROTAT       0 0 0
+// SCALE       2 2 2
+
+// Custom Object
+OBJECT 7
+custom_obj
+material 2
+TRANS       2 0 3
+ROTAT       0 0 0
+SCALE       30 30 30