53 lines
3.3 KiB
Markdown
53 lines
3.3 KiB
Markdown
# Architecture
|
|
|
|
The renderer is two programs:
|
|
|
|
- **The adapter**, `t5deRenderEngine.dll` (32-bit, `adapter/`): a renderer plugin IMVU Classic loads in place of its
|
|
OpenGL renderer (selected by the `t5de-renderer` patch in the `t5de` repository). It implements IMVU's
|
|
`PiRenderEngine` interface (`include/t5de/imvu/`, `docs/interface.md`) and passes what the client draws to the
|
|
renderer process. It stays thin: IMVU's process is 32-bit (4 GB of address space shared with Python, Gecko and
|
|
IMVU's own code) and tied to Visual C++ 2010's ABI.
|
|
- **The renderer process**, `t5de-renderer.exe` (64-bit, `renderer/`): draws, with
|
|
[Diligent Engine](https://github.com/DiligentGraphics/DiligentCore) (Apache-2.0). Being 64-bit and separate, it can
|
|
use any modern API and library, isn't limited by the client's address space, and can crash without taking the
|
|
client down.
|
|
|
|
Why Diligent: one API over Direct3D 11 (every Windows PC, and software rendering on machines without a graphics
|
|
driver), Direct3D 12 and Vulkan (ray tracing, mesh shaders), and WebGPU (browsers, for a possible future client).
|
|
It's built with CMake and fetched by `FetchContent`, pinned to a commit (`renderer/CMakeLists.txt`); only the
|
|
Direct3D 11 backend is built for now.
|
|
|
|
## How they talk
|
|
|
|
`include/t5de/protocol/` is shared by both builds, so it uses fixed-size fields only. For each engine the client
|
|
creates (each 3D view, and again on room changes), the adapter:
|
|
|
|
1. creates a shared memory channel (`Local\t5de-renderer-<pid>-<n>`) holding IMVU's window handle and a frame
|
|
count, and an event set after each frame;
|
|
2. starts `t5de-renderer.exe --channel <name> --parent <pid>` from its own folder;
|
|
3. sets the event at each `endRendering`, and sets `closing` when the engine is destroyed.
|
|
|
|
The renderer process draws into a child window it creates inside the client's window: window handles are valid in
|
|
every process, and Windows allows child windows across processes (Chrome's GPU process works this way). The child
|
|
is disabled so mouse input goes to the client's window under it, and the adapter gives the client's window
|
|
`WS_CLIPCHILDREN` so the client's own drawing stays out of it. The process exits when the engine is destroyed or the
|
|
client exits.
|
|
|
|
## Status
|
|
|
|
A spike: the adapter still has IMVU's Direct3D 9 engine draw the scene (the pass-through), and the renderer process
|
|
only draws a box in the 3D view's corner, pulsing with the client's frames. Checked in the client (Direct3D 11 in
|
|
software, on a machine without a graphics driver): the box draws over the scene from the 64-bit process, keeps in
|
|
step with the client's frames, clicks on it reach the scene, and each renderer process exits with its engine (room
|
|
changes) and with the client. The client keeps two engines alive in a room, so two processes (about 15 MB each);
|
|
one process serving all of a client's engines would be leaner.
|
|
|
|
Next: the protocol carrying each frame's calls and the resources (buffers, textures), and the renderer process
|
|
drawing the scene itself, with IMVU's Direct3D engine left as the fallback.
|
|
|
|
## Building
|
|
|
|
Two presets, one per architecture: `cmake --workflow --preset adapter` (32-bit, with the tests) and
|
|
`cmake --workflow --preset renderer` (64-bit; the first configure fetches and builds Diligent). Both programs go in
|
|
the client's folder.
|