CIS565-Fall-2016 · krupkad · Oct 20, 2016 · Oct 21, 2016 · Oct 21, 2016 · Oct 29, 2016
diff --git a/.cproject b/.cproject
@@ -21,7 +21,7 @@
 					<folderInfo id="com.nvidia.cuda.ide.toolchain.on_path.1673801701.1152097236" name="/" resourcePath="">
 						<toolChain id="com.nvidia.cuda.ide.toolchain.on_path.1386955599" name="CUDA Toolkit on the path" superClass="com.nvidia.cuda.ide.toolchain.on_path">
 							<targetPlatform archList="all" binaryParser="com.nvidia.cuda.ide.elf;com.nvidia.cuda.ide.macho;com.nvidia.cuda.ide.cubin" id="com.nvidia.cuda.ide.targetPlatform.853269530" isAbstract="false" name="Debug Platform" osList="linux,macosx" superClass="com.nvidia.cuda.ide.targetPlatform"/>
-							<builder id="com.nvidia.cuda.ide.builder.1635905474" incrementalBuildTarget="${ConfigName}" keepEnvironmentInBuildfile="false" managedBuildOn="false" name="CUDA Toolkit 7.5 Builder" superClass="com.nvidia.cuda.ide.builder"/>
+							<builder buildPath="${workspace_loc:/Project4-CUDA-Rasterizer}/build/" id="com.nvidia.cuda.ide.builder.1635905474" incrementalBuildTarget="${ConfigName}" keepEnvironmentInBuildfile="false" managedBuildOn="false" name="CUDA Toolkit 7.5 Builder" superClass="com.nvidia.cuda.ide.builder"/>
 							<tool id="nvcc.compiler.base.854769442" name="NVCC Compiler" superClass="nvcc.compiler.base">
 								<option id="nvcc.compiler.pic.1870261830" name="Position Independent Code (-fPIC)" superClass="nvcc.compiler.pic"/>
 								<inputType id="nvcc.compiler.input.cu.943232565" superClass="nvcc.compiler.input.cu"/>
@@ -186,9 +186,14 @@
 	</storageModule>
 	<storageModule moduleId="org.eclipse.cdt.core.LanguageSettingsProviders"/>
 	<storageModule moduleId="refreshScope" versionNumber="2">
+		<configuration configurationName="DebugFast"/>
+		<configuration configurationName="DebugNDEBUG"/>
 		<configuration configurationName="Default">
 			<resource resourceType="PROJECT" workspacePath="/Project4-CUDA-Rasterizer"/>
 		</configuration>
+		<configuration configurationName="Debug"/>
+		<configuration configurationName="Release"/>
 	</storageModule>
 	<storageModule moduleId="org.eclipse.cdt.internal.ui.text.commentOwnerProjectMappings"/>
+	<storageModule moduleId="org.eclipse.cdt.make.core.buildtargets"/>
 </cproject>
diff --git a/README.md b/README.md
@@ -1,20 +1,46 @@
-CUDA Rasterizer
-===============
-
-[CLICK ME FOR INSTRUCTION OF THIS PROJECT](./INSTRUCTION.md)
+CUDA Rasterization Pipeline
+======================
 
 **University of Pennsylvania, CIS 565: GPU Programming and Architecture, Project 4**
 
-* (TODO) YOUR NAME HERE
-* Tested on: (TODO) Windows 22, i7-2222 @ 2.22GHz 22GB, GTX 222 222MB (Moore 2222 Lab)
+* Daniel Krupka
+* Tested on: Debian testing (stretch), Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz 8GB, GTX 850M
+
+
+# About
+This is a CUDA rasterization pipeline made for UPenn's CIS565. It features
+* GLTF model loading
+* Perspective correct diffuse PNG textures
+* Screen space ambient occlusion (SSAO)
+The pipeline utilizes deferred shading - the `depthPass` kernel performs depth testing and
+assembles the G-buffer prior to shading and SSAO.
+
+# Screenshots
+Video of the rasterizer in action can be found [here](https://youtu.be/_Y-9eAgICrI).
+![Normals](renders/duck-normals.png "Duck Normals")
+![Lambert](renders/duck-lambert.png "Duck Lambert")
+
 
-### (TODO: Your README)
+# Textures
+I implemented perspective-correct UV texturing for diffuse coloring, using CUDA's texture memory functionality.
+![DuckTex](renders/duck-texture.png "Duck Texture")
+![Truck](renders/truck-nossao.png "truck texture")
 
-*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.
+# SSAO
+I also implemented SSAO, both using shared memory and without.
+![Truck SSAO only](renders/truck-ssao-only.png "Truck SSAO only")
+![Truck SSAO](renders/truck-ssao.png "Truck with ssao")
 
+# Performance
+![notex nosm](renders/plt_loc_notex.png "Duck Texture")
+![tex nosm](renders/plt_loc_tex.png "Duck Texture")
+![tex sm](renders/plt_sm_tex.png "Duck Texture")
+The textureless pipeline was substantially faster for the models with textures (duck, milk truck), particulary when omitting the truck's
+fairly large texture, due to having far fewer memory accesses. However, the small textureless models did experience a small slowdown, possibly due to having
+fewer calculations to mask latency, as barycentric texture coordinate interpolation was skipped. This could be fairly easily extended to include textures for
+bump mapping, specular shading, etc.
 
-### Credits
+For SSAO, shared memory did make a small improvement in all cases, but particularly the cow and duck, which both possessed large overhangs. The `ssaoBlur` kernel
+could have used shared memory as well, but the handling for cross-block contributions was necessary somewhat complicated. Depth testing could also benefit, but
+would require a tiling scheme, also introducing more complication.
 
-* [tinygltfloader](https://github.com/syoyo/tinygltfloader) by [@soyoyo](https://github.com/syoyo)
-* [glTF Sample Models](https://github.com/KhronosGroup/glTF/blob/master/sampleModels/README.md)
diff --git a/cis565_rasterizer.launch b/cis565_rasterizer.launch
@@ -1,16 +1,24 @@
 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <launchConfiguration type="org.eclipse.cdt.launch.applicationLaunchType">
 <stringAttribute key="bad_container_name" value="/Project4-CUDA-Rasterize"/>
+<booleanAttribute key="nvvp.apiTraceProfiling" value="true"/>
+<booleanAttribute key="nvvp.appReplay" value="false"/>
+<booleanAttribute key="nvvp.concurrentProfiling" value="true"/>
+<booleanAttribute key="nvvp.environmentProfiling" value="false"/>
+<intAttribute key="nvvp.multiProcess" value="0"/>
+<booleanAttribute key="nvvp.startProfiling" value="true"/>
+<intAttribute key="nvvp.timeout" value="0"/>
+<booleanAttribute key="nvvp.uvmProfiling" value="true"/>
 <stringAttribute key="org.eclipse.cdt.debug.mi.core.DEBUG_NAME" value="${cuda_tk_bin:/Project4-CUDA-Rasterizer}/cuda-gdb"/>
 <stringAttribute key="org.eclipse.cdt.debug.mi.core.GDB_INIT" value=".cuda-gdbinit"/>
 <stringAttribute key="org.eclipse.cdt.debug.mi.core.protocol" value="mi"/>
 <intAttribute key="org.eclipse.cdt.launch.ATTR_BUILD_BEFORE_LAUNCH_ATTR" value="2"/>
 <stringAttribute key="org.eclipse.cdt.launch.DEBUGGER_ID" value="com.nvidia.cuda.ide.debug.cudagdb"/>
 <stringAttribute key="org.eclipse.cdt.launch.DEBUGGER_START_MODE" value="run"/>
-<stringAttribute key="org.eclipse.cdt.launch.PROGRAM_ARGUMENTS" value="objs/tri.obj"/>
+<stringAttribute key="org.eclipse.cdt.launch.PROGRAM_ARGUMENTS" value="gltfs/box/box.gltf"/>
 <stringAttribute key="org.eclipse.cdt.launch.PROGRAM_NAME" value="build/cis565_rasterizer"/>
 <stringAttribute key="org.eclipse.cdt.launch.PROJECT_ATTR" value="Project4-CUDA-Rasterizer"/>
-<stringAttribute key="org.eclipse.cdt.launch.PROJECT_BUILD_CONFIG_ID_ATTR" value="com.nvidia.cuda.ide.toolchain.on_path.1673801701.1348522543"/>
+<stringAttribute key="org.eclipse.cdt.launch.PROJECT_BUILD_CONFIG_ID_ATTR" value=""/>
 <listAttribute key="org.eclipse.debug.core.MAPPED_RESOURCE_PATHS">
 <listEntry value="/Project4-CUDA-Rasterizer"/>
 </listAttribute>

diff --git a/gltfs/flower/flower.gltf b/gltfs/flower/flower.gltf
diff --git a/renders/2-cylinder-engine-load-correct.png b/renders/2-cylinder-engine-load-correct.png
diff --git a/renders/box-shared-bug.png b/renders/box-shared-bug.png
diff --git a/renders/duck-diffuse-texture.png b/renders/duck-diffuse-texture.png
diff --git a/renders/duck-lambert.png b/renders/duck-lambert.png
diff --git a/renders/duck-normals.png b/renders/duck-normals.png
diff --git a/renders/duck-texture.png b/renders/duck-texture.png
diff --git a/renders/plt_loc_notex.png b/renders/plt_loc_notex.png
diff --git a/renders/plt_loc_tex.png b/renders/plt_loc_tex.png
diff --git a/renders/plt_sm_tex.png b/renders/plt_sm_tex.png
diff --git a/renders/triangle.png b/renders/triangle.png
diff --git a/renders/truck-no-ssao.png b/renders/truck-no-ssao.png
diff --git a/renders/truck-ssao-only.png b/renders/truck-ssao-only.png
diff --git a/renders/truck-ssao.png b/renders/truck-ssao.png
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_50 -std=c++11 -lineinfo
     )
diff --git a/src/main.cpp b/src/main.cpp
@@ -119,7 +119,7 @@ void runCuda() {
 	glm::mat4 MVP = P * MV;
 
     cudaGLMapBufferObject((void **)&dptr, pbo);
-	rasterize(dptr, MVP, MV, MV_normal);
+	rasterize(dptr, MVP, MV, MV_normal, P);
     cudaGLUnmapBufferObject(pbo);
 
     frame++;
@@ -332,7 +332,7 @@ void keyCallback(GLFWwindow *window, int key, int scancode, int action, int mods
 //----------------------------
 //----- util -----------------
 //----------------------------
-static std::string getFilePathExtension(const std::string &FileName) {
+std::string getFilePathExtension(const std::string &FileName) {
 	if (FileName.find_last_of(".") != std::string::npos)
 		return FileName.substr(FileName.find_last_of(".") + 1);
 	return "";

diff --git a/src/main.hpp b/src/main.hpp
@@ -101,4 +101,4 @@ std::string getFilePathExtension(const std::string &FileName);
 
 void mouseButtonCallback(GLFWwindow* window, int button, int action, int mods);
 void mouseMotionCallback(GLFWwindow* window, double xpos, double ypos);
-void mouseWheelCallback(GLFWwindow* window, double xoffset, double yoffset);
+void mouseWheelCallback(GLFWwindow* window, double xoffset, double yoffset);