This creates non-sRGB texture views for sRGB texture formats to allow for interfacing with these views in compute shaders using imageLoad and imageStore.
Co-Authored-By: Rodrigo Locatti <reinuseslisp@airmail.cc>
Load the current tick to a local variable, moving it out of an atomic
and allowing us to compare the value without going through a pointer
each time. This should make the loop more optimizable.
Fix a tragic off-by-one condition that causes Vulkan's stream buffer to
think it's always full, using fallback memory. The OpenGL was also
affected by this bug to a lesser extent.
We are already using robustness2 features without requiring it
explicitly, causing potential crashes on drivers without the extension.
Requiring this at boot allows better diagnostics for it and formalizes
our usage on the extension.
There was still a code path that could wait on a timeline semaphore tick
that would never be signalled.
While we are at it, make use of more STL algorithms.
Games can bind a null index buffer (size=0) where all indices are
evaluated as zero. VK_EXT_robustness2 doesn't support this and all
drivers segfault when a null index buffer is passed to
vkCmdBindIndexBuffer.
Workaround this by creating a 4 byte buffer and filling it with zeroes.
If it's read out of bounds, robustness takes care of returning zeroes as
indices.
Bind extra bytes beyond the guest API's bound range.
This is due to some games like Astral Chain operating out of bounds.
Binding the whole map range would be technically correct, but games
have large maps that make this approach unaffordable for now.
Avoids waiting idle while the GPU finishes to do work, and fixes an
issue where we'd wait forever if a single command buffer (logic tick)
all the data.
Due to BindBufferRangeNV limitations and poor quality code emission from
our side, assembly shaders are currently slower than GLSL. Their build
time and feature advantages are still relevant, but they are outweighted
by their runtime performance.
Detect when a memory region has been joined several times and increase
the size of the created buffer on those instances. The buffer is assumed
to be a "stream buffer", increasing its size should stop us from
constantly recreating it and fragmenting memory.
Ports from OpenGL the optimization to skip small 3D uniform buffer
uploads. This will take advantage of the previously introduced stream
buffer.
Fixes instances where the staging buffer offset was being ignored.
This uses a ring buffer similar to OpenGL's stream buffer for small
uploads. This stops us from allocating several small buffers, reducing
memory fragmentation and cache locality.
It uses dedicated allocations when possible.
Reimplement the buffer cache using cached bindings and page level
granularity for modification tracking. This also drops the usage of
shared pointers and virtual functions from the cache.
- Bindings are cached, allowing to skip work when the game changes few
bits between draws.
- OpenGL Assembly shaders no longer copy when a region has been modified
from the GPU to emulate constant buffers, instead GL_EXT_memory_object
is used to alias sub-buffers within the same allocation.
- OpenGL Assembly shaders stream constant buffer data using
glProgramBufferParametersIuivNV, from NV_parameter_buffer_object. In
theory this should save one hash table resolve inside the driver
compared to glBufferSubData.
- A new OpenGL stream buffer is implemented based on fences for drivers
that are not Nvidia's proprietary, due to their low performance on
partial glBufferSubData calls synchronized with 3D rendering (that
some games use a lot).
- Most optimizations are shared between APIs now, allowing Vulkan to
cache more bindings than before, skipping unnecesarry work.
This commit adds the necessary infrastructure to use Vulkan object from
OpenGL. Overall, it improves performance and fixes some bugs present on
the old cache. There are still some edge cases hit by some games that
harm performance on some vendors, this are planned to be fixed in later
commits.
Workaround an issue on Nvidia where creating a Vulkan instance from an
active OpenGL thread disables threaded optimization on the driver.
This optimization is important to have good performance on Nvidia
OpenGL.
Instead of using a two step initialization to report errors, initialize
the GPU renderer and rasterizer on the constructor and report errors
through std::runtime_error.
