I think there was an earlier commit with a similar name but turns out
I'd only done foundational work in that commit and never actually
properly set it up. Of course once I did, there were several issues that
needed fixing to make it work correctly, but now it works as expected.
Heavily optimizes larger projects by allowing cache jobs to only copy
what has changed.
The nodes now have more control over how their accelerated shaders/sample
functions are run, as well as how items are popped off the value tables.
This allows for various optimizations that we didn't have access to before.
Addresses scheduling issues where a backend might be closed (or even
destroyed) before it processes a waveform signal from a worker. Requires
extra multithreading code.
Turned the two-step PCM transcode into one step and simplified/removed much of
the unnecessary infrastructure that supported it. This makes the code cleaner
and generally improves the code paths.
As opposed to simply leaving all values in the table, nodes can now "take"
values that they use to free up memory (e.g. "taking" input buffers if they're
used to produce an output buffer).
This is a fairly large change, expect regressions.
functions
Indexing is a lengthy process and had a high chance of getting RenderWorkers
stuck doing it rather than being responsive to cache requests. This commit
introduces a system where workers never index media, but instead signal that
media is not ready to their RenderBackends which ensure that the media gets
indexed and re-queues the affected frames when those indexes are ready.
The renderer backend can now distinguish between jobs. Previously if two jobs
of the same frame were started (which is legal if the user made a change while
frames were still being rendered), an earlier job in some situations could
finish AFTER a later job, and the backend would have no way of distinguishing
between them. This meant a frame could be erroneously set to an old value
rather than the newest. This commit introduces job identification so that old
jobs are automatically discarded.
This mostly builds on the keyframing we already set up for video, but the
audio renderer will now appropriately updated keyframe inputs per sample in
accordance with keyframe values.
This was many changes that were largely fundamentally related. They included:
- More const modifiers to enforce read only node graphs
- Support for fragment and vertex shaders from the nodes
- Support for node code loaded externally (embedded into the binary)
- Fixed issue preventing two textures from being used in a shader
- Removed several unused functions and cleaned up code
- Fixed video media node misreading its matrix input
Workers run in different threads and the backends can poll whether the worker
is currently busy or not. However the previous iteration has the worker (and an
atomic int) provide the busy state which could easily desync with the main
thread (since all workers run in different threads). By holding the busy states
in the main thread, the main thread will always be able to poll the busy state
accurately.
If the nodes are now stateless, there's nothing stopping the renderer from
rendering multiple frames at once. Earlier since the nodes held some of their
input/output data (and that data could change per frame), it was not possible
to render multiple frames at once without conflicts. Now that the node state is
held in render threads, they can do whatever they want at any time.
If Nodes only have the one output, we don't need to do so much differentiation
between them. Previous iteration used outputs as like a distinct function
within a Node (e.g. length output would return one result, buffer output would
produce a different result - each run different code to produce their results).
Now in this iteration, it's more accurate to say a Node is just one function
(which seems more appropriate for a node system anyway).
Largely conforming renderers to new NodeValue system. Code seems a lot cleaner
this way which is a nice advantage. Likely non-functional as this won't
compile just yet and still needs probably another day or two of testing to get
it back to where it was before.
I was under the impression Q_DISABLE_COPY_MOVE was a much older function than
it is. Since we historically target 5.6 and these functions are small
convenience functions, I just implemented them in the actual codebase.
Updating values rapidly would cause strange jitters as a
byproduct of the viewer trying to update from the renderer while
it was still working. Rather than the viewer trying to access the
the renderer, we now send textures in the initial update signal
to keep everything synchronized.
Major refactoring work to try sharing as much code as possible between the
video renderers and audio renderers, as well as make them as
platform-independent as possible.
itsmattkc