Push descriptors are in theory slightly faster, but come with
limitations for which we have to check.
Either way, they're not difficult to implement, so even though
no one should be using peasant-tier descriptors, do it anyway.
This permits:
- The use of Vulkan filtering on many more devices
- Better debugging due to lack of descriptor buffer support in layers
Much of the changes here are due to a requirement that updates to
descriptors must happen between the command buffer being waited on,
and the pipeline not being bound.
We routinely did it the other way around, by updating only after
we bind the pipeline.
More recent kernel versions allow for users to extract a sync_file
handle from a DMA-BUF, which can then be imported into Vulkan as a
binary semaphore.
This finally allows for synchronization between Vulkan and DMA-BUF
images, such as those from screen capture software, or VAAPI,
avoiding any corruption artifacts.
This is done fully asynchronously, where we use the kernel's
given binary semaphores as a dependency to increment the image's
usual VkSemaphores we allocate. The old imported binary semaphores
are cleaned up after execution as usual.
In the future, hwcontext_drm should receive support for explicitly
synchronized images as well, which would make the synchronization
more robust and portable.
The issue is that we ask for storage images by default if
available, but because that is gated by the format supporting
storage images, and the check for the format supporting storage
images is gated by the usage, this resulted in a catch-22.
Vulkan encoding was designed in a very... consolidated way.
You had to know the exact codec and profile that the image was going to
eventually be encoded as at... image creation time. Unfortunately, as good
as our code is, glimpsing into the exact future isn't what its capable of.
video_maintenance1 removed that requirement, which only then made encoding
images practically possible.
The issue is that enabling features requires that the device
extension is supported. The extensions bitfield was set later,
so it was always 0, leading to no features being added.
The validation layer option only supported GPU-assisted validation.
This is mutually exclusive with shader debug printfs, so we need to
differentiate between the two.
This also fixes issues with user-given layers, and leaks in case of
errors.
Fixes:
vkCreateDevice(): pCreateInfo->pNext<VkPhysicalDeviceOpticalFlowFeaturesNV> includes a
pointer to a VkPhysicalDeviceOpticalFlowFeaturesNV, but when creating VkDevice, the
parent extension (VK_NV_optical_flow) was not included in ppEnabledExtensionNames.
The Vulkan spec states: Each pNext member of any structure (including this one) in
the pNext chain must be either NULL or a pointer to a valid struct for extending
VkDeviceCreateInfo.
This commit was long overdue. The old transfer dubiously tried to
merge as much code as possible, and had very little in the way
of optimizations, apart from basic host-mapping.
The new code uses buffer pools for any temporary bufflers, and
handles falling back to buffer-based uploads if host-mapping fails.
Roundtrip performance difference:
ffmpeg -init_hw_device "vulkan=vk:0,debug=0,disable_multiplane=1" -f lavfi \
-i color=red:s=3840x2160 -vf hwupload,hwdownload,format=yuv420p -f null -
7900XTX:
Before: 224fps
After: 502fps
Ada, with proprietary drivers:
Before: 29fps
After: 54fps
Alder Lake:
Before: 85fps
After: 108fps
With the host-mapping codepath disabled:
Before: 32fps
After: 51fps
The issue with the old mechanism is that we had to introduce new
API each time we needed a new queue family, and all the queue families
were functionally fixed to a given purpose.
Nvidia's GPUs are able to handle video encoding and compute on the
same queue, which results in a speedup when pre-processing is required.
Also, this enables us to expose optical flow queues for frame interpolation.
Otherwise nothing is written into the destination when a write mapping
is requested.
For example, a vulkan frame mapped from a drm frame (which is wrapped as
a vaapi frame in the example) is used as the output of scale_vulkan
filter, it always gets a green screen without this patch.
ffmpeg -init_hw_device vaapi=va -init_hw_device vulkan=vulkan@va
-filter_hw_device vulkan -f lavfi -i testsrc=size=352x288,format=nv12
-vf
"hwupload,scale_vulkan,hwmap=derive_device=vaapi:reverse=1,format=vaapi,hwdownload,format=nv12"
-f nut - | ffplay -
Signed-off-by: Haihao Xiang <haihao.xiang@intel.com>
There are lots of files that don't need it: The number of object
files that actually need it went down from 2011 to 884 here.
Keep it for external users in order to not cause breakages.
Also improve the other headers a bit while just at it.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Otherwise the derived device and the source device might have different
PCI ID in a multiple-device system.
Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Haihao Xiang <haihao.xiang@intel.com>
This is possible because the lifetime of these structures coincide.
It has the advantage of allowing to remove AVHWFramesInternal
from the public header; given that AVHWFramesInternal.priv is no more,
most accesses to AVHWFramesInternal are no more; indeed, the only
field accessed of it outside of hwcontext.c is the internal frame pool,
making this commit very simple.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is possible because the lifetime of both coincide.
Besides reducing the number of allocations this also simplifies
access to VulkanFramesPriv as one no longer has to
go through AVHWFramesInternal.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is possible because the lifetime of both coincide.
Besides reducing the number of allocations this also simplifies
access to VulkanDevicePriv as one no longer has to
go through AVHWDeviceInternal.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Without ff_vk_exec_discard_deps which is called by ff_vk_exec_wait,
the reference count of hwframe context cannot reach zero due to
circular reference created by ff_vk_exec_add_dep_frame.
Fix#10873
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
win32 typically doesn't have unistd.h, so always including it will break
MSVC builds. The usage of those POSIX functions are already guarded by
_WIN32, so use that to guard unistd.h include as well.
VK_KHR_PORTABILITY_ENUMERATION_EXTENSION_NAME is required on macOS,
and VK_INSTANCE_CREATE_ENUMERATE_PORTABILITY_BIT_KHR flag should
be set.
Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
When users zero-init'd the struct, or left it as-is, the encode
queue family matched the graphics queue family, which led it to be
incorrectly logged as being used for encode.
This just improves the logging so this isn't printed anymore.
Fixes multiplane support on Nvidia.
Also, remove the ENCODE usage, even if the driver signals it as supported.
Currently, it's not used, and when it is used, it'll be gated behind
two extension checks.
The alignment in vulkan_unmap_from_drm() (formerly the clone
of vulkan_frame_free()) is nicer than the in vulkan_frame_free(),
let's preserve it.
Reviewed-by: Lynne <dev@lynne.ee>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Lynne
Haihao Xiang
Andreas Rheinhardt
Zhao Zhili
Benjamin Cheng
Diederik de Haas via ffmpeg-devel
Víctor Manuel Jáquez Leal