Eliminates vk_sets_mutex. Performance on average may be lower until
the descriptor mutexes are replaced and Vulkan writes are buffered
to reduce thunk calls.
This should silence warnings about some branches non returning any value
without requiring additional "return 0" statement or similar.
Also, in theory this might enable to compiler to optimize the program
a little bit more, though that's unlikely to have any measurable effect.
The existing implementation using virtual descriptor heaps, where Vk
descriptor sets are created for the bindings in the root descriptor tables,
is inefficient when multiple command lists are used with large descriptor
heaps. It also cannot support updating a descriptor set after it is bound.
This patch creates Vk sets for each D3D12 heap. Because D3D12 heaps
can contain CBV, SRV and UAV descriptors in the same heap, multiple Vk sets
are needed for each heap, however the total number of populated descriptors
is never more than (heap size + UAV counter count).
A new 'virtual_heaps' config option is introduced to make the old
implementation available when needed. It's not always possible to determine
if this is necessary when the device is created.
Up to nine Vk descriptor sets may be used. It's theoretically possible to
reduce this to eight by placing immutable samplers in the push descriptor
set layout, but contradictions in earlier versions of the Vulkan spec made
driver support inconsistent. The documentation was corrected in version
1.2.203.
This patch also adds support for UAV counter descriptor arrays. It's not
practical to add this in a separate patch due to complications with
combining the old UAV counter implementation with the new descriptor heap
implementation.
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=47713
Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=47154
Signed-off-by: Conor McCarthy <cmccarthy@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
As pointed out by Giovanni Mascellani, modular subtraction doesn't produce a
total order; in particular, it's not transitive. The nature of the types being
compared here makes it unlikely this will be an issue in practice, but it's
both fragile and setting a poor example.
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
By analogy with InterlockedIncrement. It avoids the need for a
configure check on Windows platforms.
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Use ULONG instead of unsigned long for non-Mingw msvcrt builds.
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
We tag far fewer symbols this way.
Signed-off-by: Zebediah Figura <zfigura@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
That is, tests in C, for features of the HLSL preprocessor which are difficult
to test using the shader_runner framework.
Signed-off-by: Zebediah Figura <zfigura@codeweavers.com>
Signed-off-by: Matteo Bruni <mbruni@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
We maintain separate arrays for enqueued fences and fences owned by the
fence worker thread.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
winnt.h from Wine defines ARRAY_SIZE() when __WINESRC__ is defined.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
vkd3d_log2i() is imported from wined3d.
Signed-off-by: Józef Kucia <jkucia@codeweavers.com>
Signed-off-by: Henri Verbeet <hverbeet@codeweavers.com>
Signed-off-by: Alexandre Julliard <julliard@winehq.org>
The Vulkan spec states:
* "If the query does not become available in a finite amount of time
(e.g. due to not issuing a query since the last reset), a
VK_ERROR_DEVICE_LOST error may occur."
* "If queries will never finish (e.g. due to being reset but not
issued), then vkGetQueryPoolResults may not return in finite time."
UAV counter descriptors are stored in a separate VkDescriptorSet.
A VkPipelineLayout created for compute pipeline is compatible for all
other descriptor sets with the VkPipelineLayout created for the root
signature. This ensures that only UAV counter bindings in the last
descriptor set may be disturbed when switching between pipelines with
compatible root signatures.
The implementation of vkd3d_popcount() is based on
https://graphics.stanford.edu/~seander/bithacks.html
