CIS565-Fall-2018 · IshanRanade · Oct 16, 2018 · Oct 17, 2018 · Oct 17, 2018 · Oct 17, 2018
diff --git a/README.md b/README.md
@@ -1,18 +1,58 @@
 CUDA Rasterizer
 ===============
 
-[CLICK ME FOR INSTRUCTION OF THIS PROJECT](./INSTRUCTION.md)
-
 **University of Pennsylvania, CIS 565: GPU Programming and Architecture, Project 4**
 
-* (TODO) YOUR NAME HERE
-	* (TODO) [LinkedIn](), [personal website](), [twitter](), etc.
-* Tested on: (TODO) Windows 22, i7-2222 @ 2.22GHz 22GB, GTX 222 222MB (Moore 2222 Lab)
+* Ishan Ranade
+* Tested on personal computer: Gigabyte Aero 14, Windows 10, i7-7700HQ, GTX 1060
+
+## Demo
+
+![](renders/demo.gif)
+
+## Introduction
+
+I implemented a GPU based rasterization pipeline with a number of added on features.  For a brief overview the rasterization pipeline consists of a number of steps including vertex transformation of a set of user defined vertices, primitive assembly to arrange these vertices into shapes, a fragment depth test in which the closest fragment is saved, fragment transformation to determine the final color of a pixel, and finally rendering to the screen.  Specific extra features that I chose to implement were super sampled antialiasing, UV texture mapping with bilinear texture filtering and perspective correct texture coordinates, correct color interpolation between points on a primitive, and backface culling.
+
+## Features
+
+![](renders/truck.JPG)
+
+![](renders/texture1.JPG)
+
+- UV texture mapping with bilinear texture filtering and perspective correct texture coordinates
+
+<br/>
+
+![](renders/cow.JPG)
+
+- Correct color interpolation between points on a primitive
+
+<br/>
+
+![](renders/aa2.JPG)
+
+- Super sampled anti aliasing
+
+<br/>
+
+![](renders/noaa.JPG)
+
+- No anti aliasing
+
+<br/>
+
+## Performance Analysis
+
+The rasterization step of the pipeline seemed to take the longest time, and this is because of the heavy computation in calculating which pixels overlap which triangles.  I implemented backface culling to try to reduce the computation time, and the graphs can be seen below comparing runtimes with and without this optimization.
+
+Antialiasing also increased the runtime as the size of the PBO quadrupled in size so the computations became 4 times as expensive.  This step was necessary though because without it you can easily see jagged edges along diagonal lines as can be seen in the above images.
+
+![](renders/graph1.JPG)
+
+![](renders/graph2.JPG)
 
-### (TODO: Your README)
 
-*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.
 
 
 ### Credits

diff --git a/graphs.pptx b/graphs.pptx
diff --git a/renders/aa.JPG b/renders/aa.JPG
diff --git a/renders/aa2.JPG b/renders/aa2.JPG
diff --git a/renders/cow.JPG b/renders/cow.JPG
diff --git a/renders/demo.gif b/renders/demo.gif
diff --git a/renders/graph1.JPG b/renders/graph1.JPG
diff --git a/renders/graph2.JPG b/renders/graph2.JPG
diff --git a/renders/noaa.JPG b/renders/noaa.JPG
diff --git a/renders/texture1.JPG b/renders/texture1.JPG
diff --git a/renders/truck.JPG b/renders/truck.JPG
diff --git a/shaders/README.md b/shaders/README.md
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
@@ -6,5 +6,5 @@ set(SOURCE_FILES
 
 cuda_add_library(src
     ${SOURCE_FILES}
-    OPTIONS -arch=sm_20
+    OPTIONS -arch=sm_60
     )