Project 0: Amy Liu

README.md

@@ -1,13 +1,142 @@
-Project 0 Getting Started
-====================
+# CUDA Setup and Development Environment Setup
 
 **University of Pennsylvania, CIS 5650: GPU Programming and Architecture, Project 0**
 
-* (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)
+* Amy Liu
+  * [Personal Website](https://amyliu.dev/), [LinkedIn](https://linkedin.com/in/miyalana), [Github](https://github.com/mialana).
+* Tested on: Ubuntu 24.04.3 LTS, Intel(R) Core(TM) Ultra 9 275HX 32GiB, NVIDIA GeForce RTX 5070Ti 12227MiB
 
-### (TODO: Your README)
+## Part 1: Installation and Setup
+
+NVIDIA Drivers:
+- **Version:** 580.76.05
+- **Method:** Added Ubuntu PPA repository to local APT registry in order to install latest 580 beta driver. ([primary followed guide](https://linuxconfig.org/how-to-install-the-nvidia-drivers-on-ubuntu-22-04))
+
+CUDA:
+- **Version:** 13.0
+- **Method:** From NVIDIA `.run` file. Deselected auto-install of drivers in execution menu. ([primary followed guide](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/#runfile-installation))
+
+## Part 2.1: Development Environment Testing
+
+### Part 2.1.1: Project Build & Run
+
+#### GPU Optimus Mode
+
+On Linux Wayland, NVIDIA GPU is not the primary GPU. Build & run without modification leads to the following error:
+
+```bash
+Cuda error: Kernel failed!: CUDA-capable device(s) is/are busy or unavailable.
+```
+
+To solve this, two environment variables must be added wherever I run the executable (e.g. from command line, from all Nsight applications, etc ):
+
+```bash
+__PRIME_RENDER_OFFLOAD=1
+__GLX_VENDOR_LIBRARY_NAME=nvidia
+```
+
+Alternatively, modify the `/etc/environment` file to contain:
+
+```bash
+QT_QPA_PLATFORMTHEME="wayland;xcb"
+GBM_BACKEND=nvidia-drm
+__GLX_VENDOR_LIBRARY_NAME=nvidia
+ENABLE_VKBASALT=1
+LIBVA_DRIVER_NAME=nvidia
+WLR_NO_HARDWARE_CURSORS=1
+```
+
+#### GLFW and Wayland incompability
+
+When using GLFW on Wayland, the `glfwInit()` function does not work (program will exit at `Line 49` of `main.cpp`). To counter this, set two environment variables:
+
+```bash
+WAYLAND_DISPLAY=""
+XDG_SESSION_TYPE=x11
+```
+
+#### CMakeLists.txt Modification
+
+Due to having a very new GPU model, I changed the `CUDA_ARCHITECTURES` variable from `native` to my explicit arch, i.e. `120`.
+
+```bash
+if(CMAKE_VERSION VERSION_LESS "3.23.0")
+    set_target_properties(${CMAKE_PROJECT_NAME} PROPERTIES CUDA_ARCHITECTURES OFF)
+elseif(CMAKE_VERSION VERSION_LESS "3.24.0")
+    set_target_properties(${CMAKE_PROJECT_NAME} PROPERTIES CUDA_ARCHITECTURES all-major)
+else()
+    set_target_properties(${CMAKE_PROJECT_NAME} PROPERTIES CUDA_ARCHITECTURES 120)
+endif()
+```
+
+#### `run.sh`
+
+Wrote a simple run script to facilitate 1. build + run as well as 2. run without build from command line. Simplifies environment variable requirements mentioned above.
+
+Located at `cuda-gl-check/run.sh`.
+
+### Part 2.1.2: CUDA Project UI Modification
+
+![img](images/modified_project_2.1.2.png)
+
+### Part 2.1.3: Nsight Debugger
+
+Final `launch.json` configuration:
+
+```json
+{
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "CUDA C++: Launch",
+            "type": "cuda-gdb",
+            "request": "launch",
+            "program": "${workspaceFolder}/build/bin/cuda-gl-check",
+            "environment": [
+                { "name": "__NV_PRIME_RENDER_OFFLOAD", "value": "1" },
+                { "name": "__GLX_VENDOR_LIBRARY_NAME", "value": "nvidia" }
+            ]
+        }
+    ]
+}
+```
+
+### Part 2.1.4: Nsight Systems
+
+Configuration:\
+![img](images/nsight_systems_config.png)
+
+Timeline Window:
+![img](images/nsight_systems_timeline.png)
+
+Analysis Summary Window:
+![img](images/nsight_systems_analysis_summary.png)
+
+
+### Part 2.1.5: Nsight Compute
+
+Configuration:
+![img](images/nsight_compute_config.png)
+
+Summary Window:
+![img](images/nsight_compute_summary.png)
+
+Details Window:
+![img](images/nsight_compute_details.png)
+
+
+## Part 2.2: WebGL
+
+Works on both Zen (Firefox-based) and Arc (Chromium) browsers.
+![img](images/webgl_report.png)
+
+## Part 2.3: WebGPU
+
+Does **not** work on Linux Wayland.
+
+![img](images/webgpu_report_wayland.png)
+
+Works on Linux Xorg. This requires logging out and toggling a startup option.
+
+![img](images/webgpu_report_xorg.png)
 
-Include screenshots, analysis, etc. (Remember, this is public, so don't put
-anything here that you don't want to share with the world.)

cuda-gl-check/.gitignore

@@ -4,6 +4,8 @@
 *.xcodeproj
 build
 
+compute
+
 # Created by https://www.gitignore.io/api/linux,osx,sublimetext,windows,jetbrains,vim,emacs,cmake,c++,cuda,visualstudio,webstorm,eclipse,xcode
 
 ### Linux ###

cuda-gl-check/.vscode/launch.json

@@ -0,0 +1,15 @@
+{
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "CUDA C++: Launch",
+            "type": "cuda-gdb",
+            "request": "launch",
+            "program": "${workspaceFolder}/build/bin/cuda-gl-check",
+            "environment": [
+                { "name": "__NV_PRIME_RENDER_OFFLOAD", "value": "1" },
+                { "name": "__GLX_VENDOR_LIBRARY_NAME", "value": "nvidia" }
+            ]
+        }
+    ]
+}

cuda-gl-check/.vscode/settings.json

@@ -0,0 +1,6 @@
+{
+    "files.associations": {
+        "cpp": "cuda-cpp",
+        "hpp": "cuda-cpp"
+    }
+}

cuda-gl-check/CMakeLists.txt

@@ -71,7 +71,7 @@ if(CMAKE_VERSION VERSION_LESS "3.23.0")
 elseif(CMAKE_VERSION VERSION_LESS "3.24.0")
     set_target_properties(${CMAKE_PROJECT_NAME} PROPERTIES CUDA_ARCHITECTURES all-major)
 else()
-    set_target_properties(${CMAKE_PROJECT_NAME} PROPERTIES CUDA_ARCHITECTURES native)
+    set_target_properties(${CMAKE_PROJECT_NAME} PROPERTIES CUDA_ARCHITECTURES 120)
 endif()
 target_compile_options(${CMAKE_PROJECT_NAME} PRIVATE "$<$<AND:$<CONFIG:Debug,RelWithDebInfo>,$<COMPILE_LANGUAGE:CUDA>>:-G>")
 set_property(DIRECTORY ${CMAKE_CURRENT_SOURCE_DIR} PROPERTY VS_STARTUP_PROJECT ${CMAKE_PROJECT_NAME})

cuda-gl-check/run_linux.sh

@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+
+CURRENT_DIR=$(pwd)
+SCRIPT_DIR=$( cd -- "$( dirname -- "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )
+
+while true; do
+    read -p "Do you want to re-build? (y/n): " yn
+    case $yn in
+        [Yy]* )
+        cd $SCRIPT_DIR; # cd to where script is located
+
+        trash build;
+        mkdir build && cd build;
+        cmake ..;
+        make -j$(nproc --all);
+
+        cd $CURRENT_DIR; # cd to stored original directory 
+        break;;
+        [Nn]* ) 
+        break;;
+        * ) echo "Invalid input. Please answer 'y' or 'n'.";;
+    esac
+done
+
+XDG_SESSION_TYPE=x11 $SCRIPT_DIR/build/bin/cuda-gl-check

cuda-gl-check/src/main.cpp

@@ -11,7 +11,7 @@
  */
 int main(int argc, char* argv[]) {
     // TODO: Change this line to use your name!
-    m_yourName = "TODO: YOUR NAME HERE";
+    m_yourName = "Amy Liu";
 
     if (init(argc, argv)) {
         mainLoop();

cuda-introduction/.gitignore

@@ -0,0 +1,3 @@
+build
+output
+log.txt

cuda-introduction/.vscode/c_cpp_properties.json

@@ -0,0 +1,17 @@
+{
+    "configurations": [
+        {
+            "name": "Linux",
+            "includePath": [
+                "${workspaceFolder}/source/**"
+            ],
+            "defines": [],
+            "compilerPath": "/usr/local/cuda/bin/nvcc",
+            "cStandard": "gnu17",
+            "cppStandard": "gnu++17",
+            "intelliSenseMode": "linux-gcc-x64",
+            "configurationProvider": "ms-vscode.makefile-tools"
+        }
+    ],
+    "version": 4
+}

cuda-introduction/.vscode/extension.json

@@ -0,0 +1,7 @@
+{
+    "recommendations": [
+        "nvidia.nsight-vscode-edition",
+        "ms-vscode.cpptools",
+        "ms-vscode.makefile-tools"
+    ]
+}

cuda-introduction/.vscode/launch.json

@@ -0,0 +1,13 @@
+{
+    "version": "0.2.0",
+    "configurations": [
+        {
+            "name": "CUDA C++: Launch",
+            "preLaunchTask": "CMake: build",
+            "type": "cuda-gdb",
+            "request": "launch",
+            "program": "${command:cmake.launchTargetPath}",
+            "logFile": "${workspaceFolder}/log.txt",
+        }
+    ]
+}

cuda-introduction/.vscode/settings.json

@@ -0,0 +1,63 @@
+{
+    "files.associations": {
+        "cpp": "cuda-cpp",
+        "hpp": "cuda-cpp",
+        "array": "cpp",
+        "atomic": "cpp",
+        "bit": "cpp",
+        "cctype": "cpp",
+        "charconv": "cpp",
+        "chrono": "cpp",
+        "clocale": "cpp",
+        "cmath": "cpp",
+        "compare": "cpp",
+        "concepts": "cpp",
+        "cstdarg": "cpp",
+        "cstddef": "cpp",
+        "cstdint": "cpp",
+        "cstdio": "cpp",
+        "cstdlib": "cpp",
+        "cstring": "cpp",
+        "ctime": "cpp",
+        "cwchar": "cpp",
+        "cwctype": "cpp",
+        "deque": "cpp",
+        "string": "cpp",
+        "unordered_map": "cpp",
+        "vector": "cpp",
+        "exception": "cpp",
+        "algorithm": "cpp",
+        "functional": "cpp",
+        "iterator": "cpp",
+        "memory": "cpp",
+        "memory_resource": "cpp",
+        "numeric": "cpp",
+        "optional": "cpp",
+        "random": "cpp",
+        "ratio": "cpp",
+        "string_view": "cpp",
+        "system_error": "cpp",
+        "tuple": "cpp",
+        "type_traits": "cpp",
+        "utility": "cpp",
+        "format": "cpp",
+        "initializer_list": "cpp",
+        "iomanip": "cpp",
+        "iosfwd": "cpp",
+        "iostream": "cpp",
+        "istream": "cpp",
+        "limits": "cpp",
+        "new": "cpp",
+        "numbers": "cpp",
+        "ostream": "cpp",
+        "queue": "cpp",
+        "ranges": "cpp",
+        "span": "cpp",
+        "sstream": "cpp",
+        "stdexcept": "cpp",
+        "streambuf": "cpp",
+        "text_encoding": "cpp",
+        "typeinfo": "cpp",
+        "variant": "cpp"
+    }
+}

cuda-introduction/.vscode/tasks.json

@@ -0,0 +1,16 @@
+{
+    "version": "2.0.0",
+    "tasks": [
+        {
+            "type": "cmake",
+            "label": "CMake: build",
+            "command": "build",
+            "targets": [
+                "${command:cmake.buildTargetName}"
+            ],
+            "group": "build",
+            "problemMatcher": ["$nvcc"],
+            "detail": "CMake template build task"
+        },
+    ]
+}

cuda-introduction/source/common.cu

@@ -9,7 +9,7 @@ unsigned divup(unsigned size, unsigned div)
 {
     // TODO: implement a 1 line function to return the divup operation.
     // Note: You only need to use addition, subtraction, and division operations.
-    return 0;
+    return (size + div - 1) / div;
 }
 
 void clearHostAndDeviceArray(float *res, float *dev_res, unsigned size, const int value)
@@ -26,6 +26,11 @@ bool compareReferenceAndResult(const float *ref, const float *res, unsigned size
     bool passed = true;
     for (unsigned i = 0; i < size; i++)
     {
+#if 0
+        // print thread info regardless of fail or not
+        std::cout << "ID: " << i << " \t Res: " << res[i] << " \t Ref: " << ref[i] << std::endl;
+#endif
+
         // LOOK: Check if floating point values are equal within an epsilon as returns can vary slightly between CPU and GPU
         if (std::fabs(res[i] - ref[i]) > epsilon)
         {