This gets confused by e.g. SPARSE_BIT on the TRANSFER_BIT, leading to
situations where "more specialized" is ambiguous and the logic breaks
down. So to fix it, only compare the subset we care about.
blit() implies scaling, copy() is the equivalent command to use when the
formats are compatible (same pixel size) and the rects have the same
dimensions.
This allows RAs with support for non-opaque FBO formats to use a more
appropriate FBO format for the output tex, possibly enabling a more
efficient blit operation.
This requires distinguishing between real formats (which can be used to
create textures) and fake formats (e.g. ra_gl's FBO hack).
On AMD devices, we only get one graphics pipe but several compute pipes
which can (in theory) run independently. As such, we should prefer
compute shaders over fragment shaders in scenarios where we expect them
to be better for parallelism.
This is amusingly trivial to do, and actually improves performance even
in a single-queue scenario.
Instead of using a single primary queue, we generate multiple
vk_cmdpools and pick the right one dynamically based on the intent.
This has a number of immediate benefits:
1. We can use async texture uploads
2. We can use the DMA engine for buffer updates
3. We can benefit from async compute on AMD GPUs
Unfortunately, the major downside is that due to the lack of QF
ownership tracking, we need to use CONCURRENT sharing for all resources
(buffers *and* images!). In theory, we could try figuring out a way to
get rid of the concurrent sharing for buffers (which is only needed for
compute shader UBOs), but even so, the concurrent sharing mode doesn't
really seem to have a significant impact over here (nvidia). It's
possible that other platforms may disagree.
Our deadlock-avoidance strategy is stupidly simple: Just flush the
command every time we need to switch queues, and make sure all
submission and callbacks happen in FIFO order. This required lifting the
cmds_pending and cmds_queued out from vk_cmdpool to mpvk_ctx, and some
functions died/got moved as a result, but that's a relatively minor
change.
On my hardware this is a fairly significant performance boost, mainly
due to async transfers. (Nvidia doesn't expose separate compute queues
anyway). On AMD, this should be a performance boost as well due to async
compute.
This is especially interesting for vulkan since it allows completely
skipping the layout transition as part of the renderpass. Unfortunately,
that also means it needs to be put into renderpass_params, as opposed to
renderpass_run_params (unlike #4777).
Closes#4777.
This uses the new vk_signal mechanism to order all access to textures.
This has several advantageS:
1. It allows real synchronization of image access across multiple frames
when using multiple queues for parallelism.
2. It allows using events instead of pipeline barriers, which is a
finer-grained synchronization primitive that allows for more
efficient layout transitions over longer durations.
This commit also restructures some of the implicit transition code for
renderpasses to be more flexible and correct. (Note: this technically
drops the ability to transition the image out of undefined layout when
not blending, but that was a bug anyway and needs to be done properly)
vo_gpu: vulkan: remove no-longer-true optimization
The change to the output_tex format makes this no longer true, and it
actually seems to hurt performance now as well. So just don't do it
anymore. I also realized it hurts performance when drawing an OSD, so
it's probably not a good idea anyway.
This combines VkSemaphores and VkEvents into a common umbrella
abstraction which can resolve to either.
We aggressively try to prefer VkEvents over VkSemaphores whenever the
conditions are met (1. we can unsignal the semaphore, i.e. it comes from
the same frame; and 2. it comes from the same queue).
Instead of being submitted immediately, commands are appended into an
internal submission queue, and the actual submission is done once per
frame (at the same time as queue cycling). Again, the benefits are not
immediately obvious because nothing benefits from this yet, but it will
make more sense for an upcoming vk_signal mechanism.
This also cleans up the way the ra_vk submission interacts with the
synchronization/callbacks from the ra_vk_ctx. Although currently, the
way the dependency is signalled is a bit hacky: normally it would be
associated with the ra_tex itself and waited on in the appropriate stage
implicitly. But that code is just temporary, so I'm keeping it in there
for a better commit order.
Instead of associating a single VkSemaphore with every command buffer
and allowing the user to ad-hoc wait on it during submission, make the
raw semaphores-to-signal array work like the raw semaphores-to-wait-on
array. Doesn't really provide a clear benefit yet, but it's required for
upcoming modifications.
1. No more static arrays (deps / callbacks / queues / cmds)
2. Allows safely recording multiple commands at the same time
3. Uses resources optimally by never over-allocating commands
ra_d3d11 uses the SPIR-V compiler to translate GLSL to SPIR-V, which is
then translated to HLSL. This means it always exposes the same GLSL
version that the SPIR-V compiler supports (4.50 for shaderc/glslang.)
Despite claiming to support GLSL 4.50, some features that are tied to
the GLSL version in OpenGL are not supported by ra_d3d11 when targeting
legacy Direct3D feature levels.
This includes two features that mpv relies on:
- Reading from gl_FragCoord in the fragment shader (requires FL 10_0)
- textureGather from any texture component (requires FL 11_0)
These features have been exposed as new RA caps.
Backported from @haasn's change to libplacebo, except in the current RA,
there's nothing to indicate an ra_format can be bound as a storage
image, so there's no way to force all of these formats to have a
glsl_format. Instead, the layout qualifier will be removed if
glsl_format is NULL.
This is needed for the upcoming ra_d3d11 backend. In Direct3D 11, while
loading float values from unorm images often works as expected, it's
technically undefined behaviour, and in Windows 10, it will cause the
debug layer to spam the log with error messages. Also, apparently in
GLSL, the format name must match the image's format exactly (but in
Direct3D, it just has to have the same component type.)
Backported from @haasn's change to libplacebo. More flexible than the
previous "shared || non-shared" distinction. The extra flexibility is
needed for Direct3D 11, but it also doesn't hurt code-wise.
This was confusing at best. Change it to output the actual choices.
(Seems like in the end it's always me who has to clean up other people's
bullshit.)
Context names were not unique - but they should be, so fix it. The whole
point of the original --opengl-backend option was to side-step the
tricky auto-detection, so you know exactly what you get. The goal of
this commit is to make --gpu-context work the same way. Fix the
non-unique names by appending "vk" to the names.
Keep in mind that this was not suitable for slecting the "UI" backend
anyway, since "x11" would force GLX, whereas people on not-NVIDIA
actually want "x11egl". Users trying to use --gpu-context=x11 to force
the X11 backend would always end up with GLX, which would at least break
VAAPI hardware decoding for them. Basically the idea that this option
could select the "UI" type is completely broken - it selects an
implementation, which implies a UI. Selecting the UI type This would
require a separate mechanism. (Although in theory this separate
mechanism could be part of the --gpu-context option - in any case,
someone would have to implement it.)
To achieve help output that can actually be understood, just duplicate
the code. Most of that code is duplicated anyway, and trying to share
just the list code with the result of making the output unreadable
doesn't make too much sense. If we wanted to save code/effort, we could
just remove the help output altogether.
--gpu-api has non-unique entries, and it would be nice to group them
(e.g. list all OpenGL capable contexts with "opengl"), but C makes this
simple idea too much of a pain, so don't do it.
Also remove a stray tab from the android entry on the manpage.
This commit:
- Implements output tracking (e.g. monitor plug/unplug)
- Creates the surface during registry (no other dependencies)
- Queues the callback immediately after surface creation
- Cleaner and better event handling (functions return directly)
- Better reconfigure handling (resizes reduced to 1 during init)
- Don't unnecessarily resize (if dimensions match)
Apart from that fixes 2 potential memory leaks (mime type and window
title), 2 string ownership issues (output name and make need to be
dup'd), fixes some style issues (switches were indented) and finally
adds messages when disabling/enabling idle inhibition.
The callback setter function was removed in preparation for the commit
which will use the frame event cb because it was unnecessary.
The wayland code was written more than 4 years ago when wayland wasn't
even at version 1.0. This commit rewrites everything in a more modern way,
switches to using the new xdg v6 shell interface which solves a lot of bugs
and makes mpv tiling-friedly, adds support for drag and drop, adds support
for touchscreens, adds support for KDE's server decorations protocol,
and finally adds support for the new idle-inhibitor protocol.
It does not yet use the frame callback as a main rendering loop driver,
this will happen with a later commit.
FlagBits is just the name of the enum. The actual data type representing
a combination of these flags follows the *Flags convention. (The
relevant difference is that the latter is defined to be uint32_t instead
of left implicit)
For consistency, use *Flags everywhere instead of randomly switching
between *Flags and *FlagBits.
Also fix a wrong type name on `stageFlags`, pointed out by @atomnuker
Using renderpass layout transitions is more optimal and doesn't require
a redundant pipeline barrier.
Since our render passes are static and don't change throughout the
lifetime of a ra_renderpass, we unfortunately don't have much
flexibility here - so just hard-code SHADER_READ_ONLY_OPTIMAL as the
output format as this will be the most common case.
We also can't short-circuit the transition when we need to preserve the
framebuffer contents, since that depends on the current layout; so we
still use an explicit tex_barrier in this case. (Most optimal for this
scenario would be an input attachment anyway)
In addition to the built-in nvidia compiler, we now also support a
backend based on libshaderc. shaderc is sort of like glslang except it
has a C API and is available as a dynamic library.
The generated SPIR-V is now cached alongside the VkPipeline in the
cached_program. We use a special cache header to ensure validity of this
cache before passing it blindly to the vulkan implementation, since
passing invalid SPIR-V can cause all sorts of nasty things. It's also
designed to self-invalidate if the compiler gets better, by offering a
catch-all `int compiler_version` that implementations can use as a cache
invalidation marker.
This time based on ra/vo_gpu. 2017 is the year of the vulkan desktop!
Current problems / limitations / improvement opportunities:
1. The swapchain/flipping code violates the vulkan spec, by assuming
that the presentation queue will be bounded (in cases where rendering
is significantly faster than vsync). But apparently, there's simply
no better way to do this right now, to the point where even the
stupid cube.c examples from LunarG etc. do it wrong.
(cf. https://github.com/KhronosGroup/Vulkan-Docs/issues/370)
2. The memory allocator could be improved. (This is a universal
constant)
3. Could explore using push descriptors instead of descriptor sets,
especially since we expect to switch descriptors semi-often for some
passes (like interpolation). Probably won't make a difference, but
the synchronization overhead might be a factor. Who knows.
4. Parallelism across frames / async transfer is not well-defined, we
either need to use a better semaphore / command buffer strategy or a
resource pooling layer to safely handle cross-frame parallelism.
(That said, I gave resource pooling a try and was not happy with the
result at all - so I'm still exploring the semaphore strategy)
5. We aggressively use pipeline barriers where events would offer a much
more fine-grained synchronization mechanism. As a result of this, we
might be suffering from GPU bubbles due to too-short dependencies on
objects. (That said, I'm also exploring the use of semaphores as a an
ordering tactic which would allow cross-frame time slicing in theory)
Some minor changes to the vo_gpu and infrastructure, but nothing
consequential.
NOTE: For safety, all use of asynchronous commands / multiple command
pools is currently disabled completely. There are some left-over relics
of this in the code (e.g. the distinction between dev_poll and
pool_poll), but that is kept in place mostly because this will be
re-extended in the future (vulkan rev 2).
The queue count is also currently capped to 1, because of the lack of
cross-frame semaphores means we need the implicit synchronization from
the same-queue semantics to guarantee a correct result.
Niklas Haas
James Ross-Gowan
wm4
Rostislav Pehlivanov