Make circle contours use 'foreach coordinates' for
its points. This works here, but not for general
conics. As with the other custom contours, avoid
emitting collapsed conics.
The code now follows gsk_rounded_rect_shrink() and with it the behavior
of the Cairo renderer and Webkit.
The old code did what the GL renderer and Cairo do, but I consider that
wrong.
I did not test Chrome.
Test attached
The source uniform may or may not point
to a glyph atlas. The optimization we do
for color nodes is only possible if it does,
so check this.
Fixes: #6094
Cairo and the GL renderer have a different idea of how to handle
transitioning of colors outside the defined range.
Consider these stops:
black 50%, white 50%
What color is at 0%?
Cairo would transition between the last and first stop, ie it'd do a
white-to-black transition and end up at rgb(0.5,0.5,0.5) at 0%.
GL would behave as it would for non-repeating gradients and use black
for the range [0%..50%] and white for [50%..100%].
The web would rescale the range so the first stop would be at 0% and
the last stop would be at 100%, so this gradient would be illegal.
Considering that it's possible for code to transition between the
different behaviors by adding explicit stops at 0%/100%, I could choose
any method.
So I chose the simplest one, which is what the GL renderer does and
which treats repeating and non-repeating gradients the same.
Tests attached.
We require folks to include gskglrenderer.h in order
to create a GL renderer. So we be careful to only
include header in gskglrenderer.h that won't trigger
ugly warnings.
See !6363
There is no decomposition going on for any contours,
and the tolerance argument is entirely unused.
Decomposition and tolerance is handled entirely
in gskpath.c by its trampoline.
Make gsk_path_builder_add_rect always
produce a clockwise rectangle. This matches
what we do for circles and rounded rects,
which also go clockwise. Note that we
still need to allow negative widths in
the contour code, to implement reverse().
Add a contour that optimizes some things for
rectangles. Also add rectangle detection to the
path parser, and add tests similar to what we
have for the other special contours.
This special contour takes advantage of its
rounded-rect-ness for speeding up bounding
boxes and winding numbers. It falls back
to the standard contour code for everything
else.
Add a private gsk_path_point_to_string that
can be called in the debugger if you want
to see the contents of a GskPathPoint and
are too lazy to cast it to GskRealPathPoint
yourself.
Only do the work for a curve the first time
we need it. This should greatly speed up
use cases where you only create a measure
to get the length of the path.
In order to compute path lengths efficiently, we need
to cache lookup tables. This commit adds API to let
contours allocate and free such measure data, as well
as API to use the data to go length -> point and
vice versa.
...and not around the center of the render node, as one could expect
given that the render node syntax for rotation, transform: rotate(90);,
happens to match the CSS syntax for the same thing, and CSS does rotate
around the center by default.
Signed-off-by: Sergey Bugaev <bugaevc@gmail.com>
We don't need to have the derivative as a curve,
it is enough for us to compute values of the
derivative at a given t, which we can also do
for conics.
Arcs were appealing, but they have a fatal flaw: we can't
split our arcs without changing the ellipse they trace.
That could be fixed by adding an extra parameter, but then
it is no longer any better than conics.
So switch back to conics, which have the advantage that they
are used elsewhere.
Add a new curve type for elliptical arcs
and use it for rounded rectangles and circles.
We use the 'E' command to represent elliptical
arcs in serialized paths.
The magical term to know about (because the GLSL compiler or the
validation layers sure as hell don't) is:
"dynamically uniform expression"
because if you don't have that when indexing a texture or buffer array,
you need to add nonuniformEXT() around the index variable.
Fixes the close icon on AMD having glitches of the previous icon visible
in some pixels.
We must be careful with single-point contours
that contain just a move. These never occur in
practice, but our randomized tests produce them
regularly.
Based on reverse engineering the color node and contrary to my
expectations, the matrix/offset is expressed in, and applied to,
unpremultiplied colors. The colors are being explicitly
unpremultiplied, transformed according to the matrix/offset, and
premultiplied back (see color_matrix.glsl). The matrix is getting
transposed.
Also, copy the same blurb to the corresponding GtkSnapshot function.
Signed-off-by: Sergey Bugaev <bugaevc@gmail.com>
The (out caller-allocates) and (out callee-allocates) annotations are
meant for structured or pointer types. Plain old data types are just
regular out parameters and don't need the annotation about who allocates
them.
See glib!2005, gjs#570
Take a rendernode as source and a GskPath and GskStroke,
and fill the area that is covered when stroking the path
with the given stroke parameters, like cairo_stroke() would.
This commit adds the basic infrastructure for paths.
The public APIs consists of GskPath, GskPathPoint and
GskPathBuilder.
GskPath is a data structure for paths that consists
of contours, which in turn might contain Bézier curves.
The Bezier data structure is inspired by Skia, with separate
arrays for points and operations. One advantage of this
arrangement is that start and end points are shared
between adjacent curves.
A GskPathPoint represents a point on a path, which can
be queried for various properties.
GskPathBuilder is an auxiliary builder object for paths.
graphene_rect_t is not well-suited for this purpose,
since you end up with floating-point precision problems
at the upper bound (x + width, y + height).
Instead of scale and whatnot, pass:
1. The image size
2. The viewport to map to that image size
and compute everything else from there.
In particular, we set the Vulkan viewport to the image dimensions
instead of the viewport size.
All of this makes things a lot simpler while keeping the required
functionality.
We need them for mask-only textures.
For tiffs, we convert the formats to RGBA (the idea that tiff can save
everything needs to be buried I guess) as tiffs can't do alpha-only.
Add a bunch of inline functions for graphene_rectangle_t.
We use those quite extensively in tight loops so making them as fast as
possible via inlining has massive benefits.
The current render-heavy benchmark I am playing (th paris-30k in node-editor)
went from 49fps to 85fps on my AMD.
When a GdkMemoryFormat is not supported natively and there's
postprocessing required, add a way to mark a VulkanImage as such via the
new postprocess flags.
Also allow texting such iamges only with new_for_upload() and detect
when that is the case and then run a postprocessing step that converts
that image to a suitable format.
This is done with a new "convert" shader/op.
This now supports all formats natively, no conversions happen on the CPU
anymore (unless the GPU is old).
Add an explicit begin() and an end() op. For now, this looks like
overkill, but it allows doing renderpasses with custom ops that are not
meant to render a rendernode.
Examples for this are pre/postprocessing passes or 2-pass blur.
The API was using regions because it always had. But all the code ever
did was get the extents of the region.
So simplify everything by using rectangles everywhere.
These days, we can query it with gsk_vulkan_render_get_context().
Makes quite a few functions require one less argument.
And it also makes the GskVulkanRenderPass empty. Gotta figure out what
to do with it.
Instead, build-depnd on glslc to build them.
glslc is available in all important distros for a while:
Fedora >= 28
Ubuntu >= 23.04
Debian >= 12
Arch
Opensuse >= 15.2
msys2
are the ones I checked.
So we can depend on it and avoid having to deal with keeping spirv files
up-to-date in all commits.
It's also 700kB of data, and not updating it helps.
We now store all the relevant state of the image inside the VulkanImage
struct, so we can delay barriers for as long as possible.
Whenever we want to use an image, we call the new
gsk_vulkan_image_transition() and it will add a barrier to the desired
state if one is necessary.
... and all the remaining functions still using it.
It's all unused and has been replaced by upload and download ops.
With this change, all GPU operations now go via GskVulkanOp.command()
and no more side channels exist.
This op queues a download of an image. The image will only be available
once the commands finished executing, so it requires waiting for the
render to finish, which makes the API a bit awkward.
Included is also a download_png_op() useful for debugging.
The render pass ops were not updating the image's layout to the final
layout when a render pass ends.
Fix that.
Also make the layouts explicit arguments to the render pass op.
Split out the function that uploads using a buffer, so that it can be
used with an area to only update parts of the image.
That feature is not used yet, but will be in future commits.
If a command takes too long to execute, Vulkan drivers will think they
are inflooping and abort what they were doing.
For the simple color shader with smallish nodes, this happens around
10M instances, as tested with the output of
./tests/rendernode-create-tests 10000000 colors.node
So just limit it to way lower, so that we barely never hit it, ut still
pick a big number so this optimization stays noticable.
For small regions, the optimization doesn't matter that much, so we
don't need to do lots of work on the CPU.
In particular, this should catch icons and their backgrounds (32x32),
but I was generous in selecting the number.
Gets my discrete AMD on widget-factory back to the 1900fps it had before
this optimization while making the driver clock the GPU's shader at
1.7GHz instead of the 2.1GHz it used before.
Using clear avoids the shader engine (see last commit), so if we can get
pixels out of it, we should.
So we detect the overlap with the rounded corners of the clip region and
emit shaders for those, but then use Clear() for the rest.
With this in place, widget-factory on my integrated Intel TigerLake gets
a 60% performance boost.
The op emits a vkCmdClearAttachments() with a given color. That can be
used with color nodes that are pixel-aligned and opaque to significantly
speed up rendering when the window background is a solid color.
However, currently this fails a bit outside of fullscreen when rounded
clip rectangles are in use to draw rounded corners.
Instead of using the upload vfunc and going via the code in
GskVulkanImage, copy/paste the relevant code into the command() vfunc.
This is meant to achieve multiple things:
1. Get rid of GskVulkanUploader and its own command buffer and general
non-integration with operations.
2. Get rid of GskVulkanOp:upload()
3. Get the upload/download code machinery for GskVulkanImage and put it
with the actual operations.
The current code can't do direct upload/download, that will follow in a
future commit.
... instead of doing the equivalent things manually by creating a
RenderPass and calling the relevant functions.
Now all renderpass operations are indeed stored in ops.
Also reshuffle the command emission code, because we no longer need to
emit the ops for the base renderpass.
As a result we only submit a single command buffer containing all the
render passes instead of once per render pass.
We also bind vertex buffers and descriptor sets only once now at the
start instead of once per renderpass.
Use the OpClass.stage to order operations:
1. Put upload ops first
This way we can ensure they are executed first.
2. Move subpasses for offscreens in front of the pass using them.
This is a massive refactoring because it collects all the renderops
of all renderpasses into one long array in the Render object.
Lots of code in there is still flaky and needs cleanup. That will
follow in further commits.
Other than that it does work fine though.
All the ops that just execute a shader do pretty much the same stuff, so
put it all in a single function that they all call.
It's basically faking a base class for them.
Instead of recreating the same renderpass object in every frame and for
every offscreen, just reuse it.
Technically, we can save this per-renderer or even per-display (it
should really be cached by VkDevice), but we have no infrastructure for
that.
The function name gsk_vulkan_render_get_pipeline() had been used for
GskVulkanPipeline. Since those are gone now, we can use that name for
VkPipelines.
Renderpasses get recreated every frame, but we keep render objects
around. So if we keep the vertex buffer in the render object, we can
also keep it around and just reuse it.
Also, we only need one buffer for all the render passes, which is
another bonus.
The initial buffer size is chosen at 128kB. Maximized Nautilus,
gnome-text-editor with an open file and widget-factory take ~100kB when
doing a full redraw. Other apps are between 30-50kB usually.
So I chose a value that is not too big, but catches ~90% of cases.
Interning strings is slow, especially if we can instead do direct
pointer compares.
Also refactor the pipeline lookup code a bit to make use of the
refactored code.
Set it after creating all the ops and then use it for iterating.
Note that we cannot set it while creating the ops because the array may
be realloc()ed into a different memory region which would invalidate all
the pointers.
It currently has no use, but that will come later.
Also put the typedefs into headers in gsk/vulkan, they have nthing to do
outside that directory.
Remove the function to add a node from both the GskVulkanRender and the
GskVulkanRenderPass.
That means they are both now meant to draw exactly one node.
This is a rudimentary - but working - port.
Glyph uploads are still using the old machinery, a bunch of functions
still exist that probably aren't necessary anymore and each glyph emits
its own node.
This will need to be improved in further commits.
This shader is an updated version of the mask shader, but I want to use
the mask name for the mask node and that's a different functionality.
Also, add an operation for it and partially implement the mask node
using it, so we can test that this shader works.
Replacing the shader used for text rendering is the next step.
The benefit here is that we can now properly cross-fade when one of
start/end is fully clipped out by just replacing it with an opacity op
for the other.
This was not possible with the old way we did things.
Instead of creating a pipeline GObject, just ask for the VkPipeline.
And instead of having the Op handle it, just let the renderpass look
up/create the relevant pipeline while creating commands so that it can
insert vkCmdBindPipeline calls as-needed.
This reverts commit 0f184d3270.
The renderer is good enough to make use of the clip region.
Or rather: If it isn't, the renderpass should take care of that, not the
render object.
This reverts most of commit f420c143e0
again because it turns out GPUs like combined images and samplers.
But: The one thing we don't revert is allowing the C code to select any
combination of sampler and image:
gsk_vulkan_render_get_image_descriptor() now takes a 2nd argument
specifying the sampler.
This allows the same flexibility as before, we just combine things
early.
This change was inspired by
https://developer.nvidia.com/blog/vulkan-dos-donts/
Have a resource path => vkShaderModule hash table instead of doing fancy
custom objects.
A benefit is that shader modules are now shared between all renderers
and pipelines.
Instead of creating the op manually, just pass in the renderpass and
have the op created from there.
This way ops aren't really initialized anymore, they are more appended
to the queue, so instead of foo_op_init() we can just call the function
foo_op().
The new code always uses an offscreen, even for children that are
exactly fitting texture nodes.
I would have had to write more code and didn't consider it worth it,
especially because it would have required complicating the
get_as_image() function.
This was the last node using the texture pipeline.
