This function is only used for filtering and generic compute.
The issue is that a view inherits the usage flags from the image
by default, and the spec says the view format must be compatible
with the usage. VkImageViewUsageCreateInfo allows us to filter out
the indeded uses of the imageview.
Pffff.
When using **integer** images inside shaders, it turns out
that conversion doesn't automatically happen, but we need to
explicitly use the imageviews to get the image exposed as
a suitable representation for the shader.
Finally enables bitexact image representations.
The original either reported 8 or 16-bit conversion from the
original, rather than being able to return the actual original.
This makes it usable in a situation where preserving exactness
is required.
The flexible array member struct can have padding added by
the compiler which was not taken into account properly, which
could lead to a heap buffer overflow.
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.
The old query code never worked properly, and did some hideous
heuristics to read the status bit, and work that into a return
code.
This is all best left to callers to do, which simplifies
our code a lot.
This also fixes minor validation errors regarding calling queries
which are not in their active state.
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
