Win32DirectX12-FPSCamera

Basic FPS camera starter code for compiling an Win32/DirectX12 Shader environment, just installing visual studio

Steps to compile and run:

1. Install visual studio 2019
2. Clone repo
3. Open command prompt in repo
4. run: .\Compile.bat
5. run: .\FPSCameraBasic.exe

To Debug:

1. run: .\Compile.bat
2. run: devenv .\FPSCameraBasicDebug.exe
3. When Visual Studio is running, press F11

Controls

• WSAD or ArrowKeys to move
• Space and shift to move up and down
• Mouse to look around
• Esc to pause/unpause
• F to go in and out of fullscreen
• V to turn off and on V-Sync in windowed mode

Improvements to make:

• Set an inital State to the command lists
• Alt+Tab while FullScreen should hide screen/make it non visible (minimizing screen has an animation which I don't recommend) (remember to take into account the fact that people might have multiple monitors, so clicking outside of the fullscreen window in a different monitor shouldn't minimize it, but pressing Alt+Tab while fullscreen window is in focus will minimize it) (look into fullscreen settings in here https://developer.nvidia.com/dx12-dos-and-donts and https://stackoverflow.com/questions/72156247/alt-tab-in-fullscreen-with-setfullscreenstate-directx12-does-not-minimize-window , maybe true fullscreen but a borderless window is better, have to test performance)
• try accelerators for alt-tab?
• When resizing the window super fast, there appears to be a white area on the new part of the window. (get rid of that)
• See if the dxc compiler produces higher quality shader asm than fxc
• while resizing take into account passage of time during redraws
• Use EnumAdapterByGpuPreference to get the adapter
• Take into account different DirectX12/DXGI Versions better
• While fullscreen use a better sync interval (v-sync off possible?)
• There seems to be ghosting while in fullscreen, need to fix that.
• have code entry point at WinMainCRTStartup instead for better startup speed for Release build for better start up time?
• look into shader optimization with https://github.com/jbarczak/Pyramid and https://renderdoc.org/ and https://www.guru3d.com/files-details/basemark-gpu-benchmark.html and https://devblogs.microsoft.com/pix/download/ and https://graphics.stanford.edu/~mdfisher/GPUView.html
• Should upload all the models into one preallocated GPU heap using CreatePlacedResource in the advanced model. (upload in batches)
• Distribute dll's for directx12 if they have the ability to run it, but don't have it installed
• What should the D3D_FEATURE_LEVEL_12_0 be? or D3D_FEATURE_LEVEL_11_0?
• Virtual Alloc for memory allocation (taking into account RAM and VRAM limits) and every system reports its projected memory usage over the life time of the program (have a memory counter for always around persitent memory and per scene memory, can add them up to get total projected memory usage)
• Shader Hot reloading in debug build using .cso files in debug build with a hotkey
• Take into account different DPI
• Handle Device Removal and SwapChain/Resource loss https://docs.microsoft.com/en-us/windows/win32/api/d3d12/nf-d3d12-id3d12device-getdeviceremovedreason (this but in D3D12 https://docs.microsoft.com/en-us/previous-versions/windows/apps/dn458383(v=win.10)?redirectedfrom=MSDN)
• Bundle Command List commands
• Look into https://developer.nvidia.com/blog/advanced-api-performance-memory-and-resources/ and Mesh Shaders, Sampler Feedback, Direct Storage (directx11 and directx12 features)
• Have MSAA (multi sample anti-aliasing) option from the start for the engine (make it toggeable), too hard to implement years down the line, or if not MSAA (cause MSAA might be too slow without optimizations), then do SMAA (subpixel morphological anti-aliasing) plus a TA x2 (a 2 sample temporal aliasing)
• Have a temporal upscaling as an option
• If doing raytracing look into doing RTX combined with ReStir (unless another major breakthrough occurs before then)
• Read criticisms in here http://www.mamoniem.com/behind-the-pretty-frames-elden-ring/
• use bundles to speed up command lists
• https://developer.nvidia.com/blog/the-peak-performance-analysis-method-for-optimizing-any-gpu-workload/ http://32ipi028l5q82yhj72224m8j.wpengine.netdna-cdn.com/wp-content/uploads/2017/03/GDC2017-Asynchronous-Compute-Deep-Dive.pdf https://docs.microsoft.com/en-us/windows/win32/direct3d12/user-mode-heap-synchronization look into async gpu work and optimizing pipeline throughput
• Save 1 Byte on alpha channel of color for opaque objects by not uploading an alpha for the colors