Commit Graph

4220 Commits

Author SHA1 Message Date
Philip Rebohle 735e741681
[d3d9] Add spec constant for null samplers
And rework the way spec constants are updated.
2022-07-03 13:41:06 +02:00
Philip Rebohle 186dc29384
[d3d9] Remove bound spec constants from fixed-function shaders
Texture stages are disabled anyway when textures are unbound,
so we will never hit the unbound case.
2022-07-03 13:41:06 +02:00
Philip Rebohle 8f03c3a419
[dxvk] Don't check binding mask when processing barriers 2022-07-03 13:41:06 +02:00
Philip Rebohle 2e7e9eac7a
[dxvk] Require EXT_robustness2 and the null descriptor feature 2022-07-03 13:41:03 +02:00
Blisto91 f95f541852
[util] Limit Bionic Commando to 60fps (#2685) 2022-07-02 16:51:04 +02:00
Blisto91 98dcd722ea
[util] Add workaround for Garden Warfare 2 (#2700) 2022-07-02 16:50:12 +02:00
Philip Rebohle e406484b84
[d3d9] Don't use VK_RESOLVE_MODE_AVERAGE_BIT_KHR for stencil resolves
Doesn't work, always write sample zero instead.
2022-07-02 16:48:10 +02:00
Philip Rebohle 76ba03398d
[d3d9,dxso] Fix push constant validation errors
Derp.
2022-07-01 18:06:47 +02:00
Philip Rebohle e81094533b
[dxvk] Fix gamma texture bind point 2022-06-30 20:49:43 +02:00
Philip Rebohle 1296890083
[dxvk] Remove error logging from updateResourceBindings
It's not super useful to emit log messages in the hottest code path, and
for some reason GCC compiles some stringstream initialization code into
those parts of the function that are unconditionally executed.
2022-06-28 14:35:58 +02:00
Philip Rebohle 0e38b11569
[dxvk] Rework DxvkResource lifetime tracking
Reduces the number of atomic operations required for lifetime tracking by
using a single 64-bit integer for usage tracking and reference counting.
2022-06-28 14:35:58 +02:00
Philip Rebohle 8d1d3d66e0
[dxvk] Optimize descriptor set binding further 2022-06-28 14:35:58 +02:00
Philip Rebohle db07861518
[dxvk] Remove old pipeline layout implementation 2022-06-28 14:35:58 +02:00
Philip Rebohle c67481b904
[hud] Add HUD item for descriptor stats 2022-06-28 14:35:58 +02:00
Philip Rebohle d4a3b823a2
[dxvk] Add stat counters for descriptor sets and pools 2022-06-28 14:35:58 +02:00
Philip Rebohle 758ba5a80d
[dxvk] Inline all frequently-used binding methods 2022-06-28 14:35:58 +02:00
Philip Rebohle d4d87123b4
[dxvk] Add safety mechanism to submit large descriptor pools 2022-06-28 14:35:58 +02:00
Philip Rebohle cfc06405d2
[dxvk] Recycle Vulkan descriptor pools as well
Reduces the number of expensive reallocations when large descriptor
pools get reset and repopulated.
2022-06-28 14:35:58 +02:00
Philip Rebohle eea5c9f0da
[dxvk] Rename new descriptor pool implementation to DxvkDescriptorPool 2022-06-28 14:35:58 +02:00
Philip Rebohle 6aeed40af2
[dxvk] Remove old descriptor pool implementation 2022-06-28 14:35:58 +02:00
Philip Rebohle af418dcffd
[dxvk] Fix pipeline invalidation
We need to update descriptors and other graphics state when changing
between compute and graphics. This happened to work by chance since
any real-world app binds a new set of shaders around mode switches
anyway, but it could theoretically happen that we wouldn't update
descriptor sets on the first draw after a dispatch.
2022-06-28 14:35:58 +02:00
Philip Rebohle a27448bc76
[dxvk] Bind consecutive descriptor sets in one go
Most of the time we'll be able to bind all sets in one iteration. Binding
sets is expected to be cheap in the driver, but we should avoid unnecessary
function call overhead for this frequently called function.
2022-06-28 14:35:58 +02:00
Philip Rebohle e2b7522034
[dxvk] Use persistent descriptor pool for regular descriptor sets 2022-06-28 14:35:57 +02:00
Philip Rebohle f4e9b76515
[dxvk] Add descriptor pool tracking to command list 2022-06-28 14:35:57 +02:00
Philip Rebohle 9b0b1edf74
[dxvk] Introduce persistent descriptor pool 2022-06-28 14:35:57 +02:00
Philip Rebohle 7e42939a4a
[d3d11] Call endFrame at the end of each frame 2022-06-28 14:35:57 +02:00
Philip Rebohle acf70501d2
[d3d9] Call endFrame at the end of each frame 2022-06-28 14:35:57 +02:00
Philip Rebohle 5610b3a742
[dxvk] Introduce endFrame method 2022-06-28 14:35:57 +02:00
Philip Rebohle ab0c15ea54
[dxvk] Introduce DxvkContextType 2022-06-28 14:35:57 +02:00
Philip Rebohle 8dde72da06
[dxvk] Do not allocate descriptor set space for dynamic UBOs 2022-06-28 14:35:57 +02:00
Philip Rebohle f34d1c886a
[dxvk] Only use descriptor update templates in 32-bit builds 2022-06-28 14:35:57 +02:00
Philip Rebohle 15cf130369
[dxvk] Optimize descriptor set binding further 2022-06-28 14:35:57 +02:00
Philip Rebohle f9e6d8e23a
[dxvk] Remove old resource update code 2022-06-28 14:35:57 +02:00
Philip Rebohle 219853aa9f
[dxvk] Rework dirty descriptor state tracking 2022-06-28 14:35:57 +02:00
Philip Rebohle db85de8c91
[dxvk] Add method to query shader stages that can access a buffer 2022-06-28 14:35:45 +02:00
Philip Rebohle 893183a7cc
[dxvk] Remove old pipeline layout from pipeline objects 2022-06-28 14:34:09 +02:00
Philip Rebohle 9be454fd3e
[dxvk] Use new pipeline layout for barrier tracking and other things 2022-06-28 14:34:07 +02:00
Philip Rebohle d5e53d3271
[dxvk] Add resource binding code using new pipeline layouts 2022-06-28 14:32:31 +02:00
Philip Rebohle ef55a7c2a0
[dxvk] Add method to create shader module using new binding layout 2022-06-28 14:32:31 +02:00
Philip Rebohle d65ceb82cf
[dxvk] Rework binding ID patching 2022-06-28 14:32:31 +02:00
Philip Rebohle 955e0cca62
[dxvk] Use native integer size for DxvkBindingSet
May make things a tad faster in 64-bit applications.
2022-06-28 14:32:31 +02:00
Philip Rebohle 16eba45987
[dxvk] Implement 64-bit tzcnt 2022-06-28 14:32:31 +02:00
Philip Rebohle 038ee0416d
[dxvk] Add setRange method to DxvkBindingSet 2022-06-28 14:32:31 +02:00
Philip Rebohle ec5ea71174
[dxvk] Store and create pipeline layout objects in pipeline manager
This should help avoid a lot of duplication.
2022-06-28 14:32:31 +02:00
Philip Rebohle 79ecd4e94a
[dxbc] Pass new binding info struct to shaders 2022-06-28 14:32:31 +02:00
Philip Rebohle cb57c2f7fb
[d3d11] Pass new binding info struct to video processor shader 2022-06-28 14:32:31 +02:00
Philip Rebohle e795f3f33a
[dxso] Pass new binding info struct to shaders 2022-06-28 14:32:31 +02:00
Philip Rebohle ecbada30f5
[d3d9] Pass new binding info struct to fixed-function shaders 2022-06-28 14:32:31 +02:00
Philip Rebohle 5edd8e92a8
[d3d9] Pass new binding info struct to various helper shaders 2022-06-28 14:32:31 +02:00
Philip Rebohle 7b8b50bca6
[hud] Pass new shader binding info struct to HUD shaders 2022-06-28 14:32:31 +02:00
Philip Rebohle 89dadc8453
[dxvk] Pass new shader binding info struct to swapchain blit shaders 2022-06-28 14:32:31 +02:00
Philip Rebohle 70a95d9085
[dxvk] Add DxvkBindingLayout to DxvkShader class
Supposed to replace the old descriptor model eventually.
2022-06-28 14:32:31 +02:00
Philip Rebohle 53519e2bd5
[dxvk] Remove old resource binding methods 2022-06-28 14:32:30 +02:00
Philip Rebohle 4cc559d690
[d3d9] Pass shader stage when binding shader resources 2022-06-28 14:32:30 +02:00
Philip Rebohle 0fb1227792
[d3d9] Pass shader stage when binding format conversion resources 2022-06-28 14:32:30 +02:00
Philip Rebohle d05864cbcb
[d3d11] Pass shader stage for binding video processor resources 2022-06-28 14:32:30 +02:00
Philip Rebohle a1bbc77c04
[d3d11] Pass shader stage for binding resources in context methods 2022-06-28 14:32:30 +02:00
Philip Rebohle ce0a2f08f0
[hud] Pass shader stage when binding HUD shader resources 2022-06-28 14:32:30 +02:00
Philip Rebohle fe03327ecd
[dxvk] Pass shader stage when binding HUD resources 2022-06-28 14:32:30 +02:00
Philip Rebohle 10eabb34da
[dxvk] Add shader stage parameter to binding methods 2022-06-28 14:32:30 +02:00
Philip Rebohle 3751edbe0c
[dxvk] Introduce DxvkBindingLayout and related classes
This is intended to replace the legacy DxvkPipelineLayout, and can support
multiple descriptor sets.
2022-06-28 14:32:30 +02:00
Philip Rebohle 67d03aabd0
[dxvk] Make recycler a ring buffer
Ensures that recycled objects actually get reused soon. Somewhat
important for memory efficiency in descriptor pools.
2022-06-28 14:32:30 +02:00
Philip Rebohle 384a665700
[dxvk] Optimize util::pipelineStages 2022-06-28 14:31:52 +02:00
Justin Kim 972de7c9fb
converting flags from VkShaderStageFlags to VkPipelineStageFlags in commitGraphicsBarriers (#2696) 2022-06-28 13:09:14 +02:00
Federico Dossena dee36be20d
Added config for A Way Out (#2694) 2022-06-25 11:17:42 +02:00
pchome 27163a6a29 [util] Fix built-in config options loging 2022-06-22 18:21:31 +02:00
Blisto91 661f8b5b56 [util] Add Forged Alliance Forever to Supreme Commander 2022-06-20 11:59:40 +01:00
Robin Kertels bd29fbd95d [util] Enable sampler type spec constants for SWTOR
Co-authored-by: Blisto91 <47954800+Blisto91@users.noreply.github.com>
2022-06-13 21:34:28 +01:00
Blisto91 968f0cdbc3 [util] Strict float emulation for Supreme Commander 2022-06-12 20:16:56 +01:00
Robin Kertels 4f56e72d56 [util] Enable strict float emulation for Sonic Adventure 2 2022-06-11 01:31:48 +02:00
Georg Lehmann 2e4caa4c14 [util] Stop using deprecated std::iterator. 2022-06-09 22:07:52 +02:00
Philip Rebohle 9e5c61bf88
[dxvk] Create state cache threads on demand 2022-06-07 11:46:06 +02:00
Andrew Eikum 279b4b7ec2 [d3d9] Defer surface creation if no HWND is given to device
Planetary Annihilation: TITANS creates a device with a NULL HWND and
requires it to succeed.
2022-06-03 13:07:43 +00:00
Philip Rebohle c596738205
[hud] Fix incorrect array length for VS resources 2022-06-02 19:57:22 +02:00
Blisto91 1862e4dc8d [util] Report Nvidia VendorId for Myst V
Game was made before ATI Technologies was bought by AMD and so doesn't recognize AMD as a GPU vendor, which for some reason makes it bug out.
It also works when it sees the word "Radeon" in the device description, which is why this issue doesn't show on amdvlk or wined3d.
2022-06-02 15:11:17 +02:00
Mike Lothian de0f81fcdc [spirv] Add utility include
This fixes a compile issue with GCC 12.1

FAILED: src/spirv/libspirv.a.p/spirv_compression.cpp.obj
i686-w64-mingw32-g++ -Isrc/spirv/libspirv.a.p -Isrc/spirv -I../dxvk-9999/src/spirv -I../dxvk-9999/include -fdiagnostics-color=always -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -Wnon-virtual-dtor -std=c++17 -O0 -DNOMINMAX -D_WIN32_WINNT=0xa00 -msse -msse2 -msse3 -mfpmath=sse -Wimplicit-fallthrough -O3 -march=native -pipe -flto=16 -mno-avx -MD -MQ src/spirv/libspirv.a.p/spirv_compression.cpp.obj -MF src/spirv/libspirv.a.p/spirv_compression.cpp.obj.d -o src/spirv/libspirv.a.p/spirv_compression.cpp.obj -c ../dxvk-9999/src/spirv/spirv_compression.cpp
In file included from ../dxvk-9999/src/spirv/../util/util_flags.h:5,
                 from ../dxvk-9999/src/spirv/spirv_include.h:7,
                 from ../dxvk-9999/src/spirv/spirv_instruction.h:6,
                 from ../dxvk-9999/src/spirv/spirv_code_buffer.h:8,
                 from ../dxvk-9999/src/spirv/spirv_compression.h:5,
                 from ../dxvk-9999/src/spirv/spirv_compression.cpp:1:
../dxvk-9999/src/spirv/../util/util_bit.h:300:33: warning: ‘template<class _Category, class _Tp, class _Distance, class _Pointer, class _Reference> struct std::iterator’ is deprecated [-Wdeprecated-declarations]
  300 |     class iterator: public std::iterator<std::input_iterator_tag,
      |                                 ^~~~~~~~
In file included from /usr/lib/gcc/i686-w64-mingw32/12.1.0/include/g++-v12/bits/stl_algobase.h:65,
                 from /usr/lib/gcc/i686-w64-mingw32/12.1.0/include/g++-v12/vector:60,
                 from ../dxvk-9999/src/spirv/spirv_compression.h:3:
/usr/lib/gcc/i686-w64-mingw32/12.1.0/include/g++-v12/bits/stl_iterator_base_types.h:127:34: note: declared here
  127 |     struct _GLIBCXX17_DEPRECATED iterator
      |                                  ^~~~~~~~
../dxvk-9999/src/spirv/spirv_code_buffer.h: In member function ‘size_t dxvk::SpirvCodeBuffer::endInsertion()’:
../dxvk-9999/src/spirv/spirv_code_buffer.h:214:19: error: ‘exchange’ is not a member of ‘std’
  214 |       return std::exchange(m_ptr, m_code.size());
      |                   ^~~~~~~~

Signed-off-by: Mike Lothian <mike@fireburn.co.uk>
2022-06-02 15:10:53 +02:00
Philip Rebohle 5d0273f520
[dxvk] Remove interpolation decorations for replaced shader inputs 2022-06-01 14:11:09 +02:00
Philip Rebohle 6d3ba1b7d7
[dxvk] Perform validation on render pass formats read from state cache 2022-06-01 14:11:09 +02:00
Philip Rebohle 80e125a130
[dxvk] Perform more extensive validation on pipeline state vectors 2022-06-01 14:11:06 +02:00
Blisto91 4ff7687dea [util] Force SM1 for the Halo CE SPV3 launcher 2022-05-31 15:45:23 +02:00
Robin Kertels 7bcf3e1062
[util] Force SM1 for the Warhammer ROR Launcher (#2579)
Co-authored-by: Philip Rebohle <25567304+doitsujin@users.noreply.github.com>
2022-05-31 15:19:30 +02:00
Winter Snowfall 8c96830752
Add d3d9.deferSurfaceCreation workaround for Scrapland (Remastered) (#2574) 2022-05-31 15:17:31 +02:00
noneistaken 93eb86aa3e Blend HUD text shadow and center correctly
This fixes incorrect blending between HUD text, which would cause a narrow transparent gap between the text center and border/shadow that is visible at large text sizes.
2022-05-30 18:23:19 +02:00
Zhiyi Zhang 8147844aef [dxgi] Return DXGI_ERROR_INVALID_CALL for invalid IDXGIAdapter3::RegisterVideoMemoryBudgetChangeNotificationEvent() parameters
According to wine tests 14237e321b

Signed-off-by: Zhiyi Zhang <zzhang@codeweavers.com>
2022-05-25 18:44:20 +02:00
WinterSnowfall 51e56c0420 [util] Enforce a maxAvailableMemory limit for Majesty 2 2022-05-23 07:54:37 +00:00
Paul Gofman cc11bb8240 [d3d9] Don't adjust window position and size in windowed mode. 2022-05-23 07:53:13 +00:00
Paul Gofman cde0fbe7b0 [util] Enable d3d9.deferSurfaceCreation for Small Radios Big Televisions 2022-05-19 02:51:41 +00:00
Joshua Ashton 5b7406fed5 [d3d9] Fix crash when using StretchRect with NULL rts 2022-05-10 02:02:35 +01:00
Philip Rebohle 6ddd0cfe0a
[d3d11] Do not add UINT formats to all UAV format lists
Analogous to vkd3d-proton changes. Allows drivers to use compression more
frequently.
2022-04-25 17:37:10 +02:00
Philip Rebohle a85f7c48a5
[d3d11] Make ClearUnorderedAccessViewUINT fallback more generic 2022-04-25 17:34:45 +02:00
Philip Rebohle 304d0549d5
[dxbc] Implement range check for private array reads
We already do this for stores.
2022-04-21 18:06:41 +02:00
Robin Kertels 3004026db8 [d3d9] Only bind RT if we actually write to it
The alternative render path for shadow maps in Dead Space relies on this.
2022-04-20 00:58:17 +01:00
Robin Kertels bdafa16e39 [d3d9] Only bind depth buffer if the depth or stencil test is enabled 2022-04-20 00:58:17 +01:00
Joshua Ashton d6cbd99141 [util] Add missing include to thread.h 2022-04-19 18:14:58 +01:00
Joshua Ashton 08e2ec6c98 [util] Implement env helpers on non-Windows platforms 2022-04-19 17:08:17 +01:00
Joshua Ashton 147f3738f9 [d3d9] Use strlcpy helper 2022-04-19 17:08:17 +01:00
Joshua Ashton dc6b7fa4a7 [util] Add strlcpy helper
strncpy is not safe.
2022-04-19 17:08:17 +01:00
Joshua Ashton 9ee0f51870 [util] Implement thread helpers on non-Windows platforms 2022-04-19 17:08:06 +01:00
Robin Kertels 9302d33ac7 [d3d9+util] Enable invariant position by default 2022-04-18 12:40:46 +01:00
Philip Rebohle 35e8f4676b [dxvk] Remove old shader creation code 2022-04-18 13:20:16 +02:00
Philip Rebohle a731f5daae [d3d11] Use new DxvkShader constructor for video shaders 2022-04-18 13:20:16 +02:00
Philip Rebohle 69817a84dc [d3d11] Use new DxvkShaderCreateInfo struct to retrieve shader info 2022-04-18 13:20:16 +02:00
Philip Rebohle 427f51eda7 [d3d9] Use new DxvkShader constructor for format conversion 2022-04-18 13:20:16 +02:00
Philip Rebohle 98ec79f6fa [d3d9] Use new DxvkShader constructor for SWVP emulation 2022-04-18 13:20:16 +02:00
Philip Rebohle ab95b61e44 [d3d9] Use new DxvkShader constructor for fixed-function shaders 2022-04-18 13:20:16 +02:00
Philip Rebohle 10bab0c182 [dxso] Use new DxvkShader constructor 2022-04-18 13:20:16 +02:00
Philip Rebohle 8993560cde [dxbc] Use new DxvkShader constructor 2022-04-18 13:20:16 +02:00
Philip Rebohle 81d88a484d [dxvk] Use new DxvkShader constructor for HUD shaders 2022-04-18 13:20:16 +02:00
Philip Rebohle 62e706a6bb [dxvk] Use new DxvkShader constructor for swap chain shaders 2022-04-18 13:20:16 +02:00
Philip Rebohle d0b52f3ac5 [dxvk] Use new DxvkShaderCreateInfo struct to retrieve shader info 2022-04-18 13:20:16 +02:00
Philip Rebohle b0db58f098 [dxvk] Introduce new way to create DxvkShader objects 2022-04-18 13:20:16 +02:00
Philip Rebohle 736f743ae4
[spirv] Implement faster in-memory compression for shaders
Seems to be anything up to 3x as fast to decode than the previous code,
with the compression ratio being slightly worse. Encoding seems faster
as well.
2022-04-11 02:55:12 +02:00
Blisto91 3672375bc1 [util] Use cached constant buffers for Armored Warfare 2022-04-09 15:19:49 +02:00
Robin Kertels 42c66c410e [d3d9] Calculate slice alignment when uploading straight from the mapping buffer 2022-04-06 18:52:58 +00:00
Paul Gofman 957a305ca8 [d3d9] Ignore multiple app activation window messages. 2022-04-06 18:51:50 +00:00
Paul Gofman b0ed9e30ce [d3d9] Filter window messages when processing WM_ACTIVATEAPP. 2022-04-06 18:51:50 +00:00
Robin Kertels ce87bec412 [util] Force sampler type spec const for Star Wars TFU2
The game tries to binda  2D texture to a slot that is declared
as a 3D texture in the shader. This causes one particle effect
to be completely black because DXVK does not bind the texture
2022-04-05 22:37:04 +00:00
Philip Rebohle 95a3413949
[util] Fix typo in app profiles
Accidentally broke everything.
2022-03-31 18:01:52 +02:00
Philip Rebohle e07157fe72
[d3d11] Only apply anisotropy override to linear samplers
Mirrors D3D9, more or less.
2022-03-30 14:37:33 +02:00
Philip Rebohle e7e7fa231c
[dxvk] GPU query reset path
Require VK_EXT_host_query_reset instead. This fallback path is
untested nowadays and too slow to be useful.
2022-03-30 13:32:18 +02:00
Georg Lehmann f0ccd8fe2e [util] Limit Limbo to 60 fps
Fixes: #2564
2022-03-28 21:39:28 +02:00
Derek Lesho d11f0ac77b [d3d11] Always export correct shared handle type from ::GetSharedHandle and ::CreateSharedHandle
Before we just assumed that the calls here would match the corresponding flag value (D3D11_RESOURCE_MISC_SHARED -> ::GetSharedHandle, D3D11_RESOURCE_MISC_SHARED_NTHANDLE -> ::CreateSharedHandle), but it turns out that its possible to set both flags and use both methods.  Now we always tell Vulkan to export a KMT handle if D3D11_RESOURCE_MISC_SHARED is present, and use openKmtHandle to get an NT handle when needed.
2022-03-28 21:38:42 +02:00
Philip Rebohle 1c3736da8c
[dxvk] Filter out unnecessary access flags when recording barriers
Rationale is as follows:
- srcAccess never needs to contain read flags, since any memory being
  read must have been made visible before by a write operation
- dstAccess is only relevant if srcAccess contains a write, because
  reads alone cannot modify memory and thus do not require making the
  same memory available again. An exception are layout transitions.

Doesn't really change performance in anything as far as I can tell, but
we avoid some unnecessary UBO cache flushes in compute-heavy scenarios.
2022-03-28 10:45:29 +02:00
Philip Rebohle ebdaf90fdc
[util] Enable d3d9.deferSurfaceCreation for Atelier Sophie 2
2022 and K-T are still using D3D9 for video stuff.
2022-03-28 04:37:02 +02:00
Philip Rebohle a72463c5af
[meta] Release 1.10.1 2022-03-26 12:44:25 +01:00
Philip Rebohle b4efaa4ef0
[util] Enable cached constant buffers for Frostpunk
Massively improves CPU-bound performance.
2022-03-26 12:44:25 +01:00
Robin Kertels 4e34c29c56 [d3d9] Disable culling when the app passes an invalid value 2022-03-25 17:13:12 +01:00
Philip Rebohle 991a11617a
[util] Enable d3d9.deferSurfaceCreation for Stranger of Paradise FFO
Reportedly required for VRR to work. Game still doesn't work here.
2022-03-25 16:36:17 +01:00
Philip Rebohle 0db26a0456
[dxbc] Actually do the skip range check thing properly
Turns out the first attempt only worked because my test case didn't
do any dynamically indexed stores at all, but broke everything else.
Oops.
2022-03-24 12:53:29 +01:00
Philip Rebohle 3ecd13cec2
[dxbc] Only emit temp array range check for dynamically indexed stores
Generates less code and makes things slightly more readable.
2022-03-24 12:47:40 +01:00
Philip Rebohle c590736fec
[dxbc] Generate smallest possible vectors for local arrays
FXC is buggy and always emits vec4 in the array declaration,
so we'll have to analyze the used components ourselves.
2022-03-24 12:47:40 +01:00
Robin Kertels b2f22d5719 [d3d9] UpdateTexture: Handle automatic mip gen properly 2022-03-24 02:31:42 +00:00
Philip Rebohle 119c1ececb
[util] Set frame latency to 1 for God of War
Frame pacing is horrible otherwise, as of the 1.0.9 update.
2022-03-24 03:05:29 +01:00
Philip Rebohle 6b8e8afd5b
[dxvk] Zero-initialize newly allocated buffer slices on creation
Fixes random flicker in God of War. Since patch 1.0.9, the game's lighting
system relies on MAP_DISCARD returning a zero-initialized memory slices for
its constant buffers, or some lights would get skipped in various compute
passes. Changing the memset to e.g. write 0xFF instead of 0 shows this issue.
2022-03-24 02:46:25 +01:00
Philip Rebohle e440fa26ab
[dxbc] Handle fallthrough around default properly 2022-03-23 15:32:45 +01:00
Philip Rebohle 8823e4bb3d
[dxgi] Work around swapchain use-after-free bugs
Affects Divinity: Original Sin Enhanced Edition. Requires Wine hack to
delay memory deallocation to not crash during resolution changes.
2022-03-23 15:09:39 +01:00
Philip Rebohle 586948df1e
[d3d11] Get strong reference to swap chain in swap chain back buffers 2022-03-22 19:48:13 +01:00
Philip Rebohle 42edb62df8
[d3d11] Use smart pointer for swap chain back buffer 2022-03-22 19:28:49 +01:00
Philip Rebohle b015cf0bb2
[dxbc] Support switch-case fallthrough
Apparently this is a thing in Shader Model 4, although FXC cannot emit it.
2022-03-22 17:32:43 +01:00
Philip Rebohle 0d54f7161c
[util] Enable cached dynamic resources for AC3 and AC4
Without it, AC3 chugs along at 40 FPS on my 5950X.
2022-03-19 20:12:20 +01:00
Oleg Kuznetsov b36ca2c758 [d3d9] Fix Visual Studio build to resolve 'operator !=' is ambiguous error for RECT 2022-03-17 11:35:19 +01:00
Oleg Kuznetsov 5156994440 [dxvk] Add a config option to enable debug utils in addition to DXVK_PERF_EVENTS=1 2022-03-17 11:35:19 +01:00
Joshua Ashton cf1cee04b8 [d3d11] Register annotation interfaces with D3D9
Some apps try use the D3DPERF_ functions for debug markers/annotations.

This utilizes the DXVK_RegisterAnnotation hidden functions to share the interfaces.

Co-authored-by: Oleg Kuznetsov <okouznetsov@nvidia.com>
2022-03-17 11:35:19 +01:00
Joshua Ashton 937a60c882 [d3d9] Add hidden exports for registering annotations
Adds DXVK_RegisterAnnotation at ordinal 28257 and DXVK_UnRegisterAnnotation at ordinal 28258.
2022-03-17 11:35:19 +01:00
Joshua Ashton 787a979514 [d3d9] Implement D3D9UserDefinedAnnotation 2022-03-17 11:35:19 +01:00
Joshua Ashton 47b1ab52ce [d3d9] Implement D3D9GlobalAnnotationList 2022-03-17 11:35:19 +01:00
Joshua Ashton 5d54d79865 [d3d11] Use IDXVKUserDefinedAnnotation 2022-03-17 11:35:19 +01:00
Joshua Ashton a010397f34 [util] Move DecodeD3DColor to util
This will be used in the D3D11UserDefinedAnnotation implementation to handle PIX calls which contain a color.
2022-03-17 11:35:19 +01:00
Joshua Ashton d5d5c1a8bc [dxvk] Define IDXVKUserDefinedAnnotation
Something common to share for perf markers between D3D9 and D3D11.

Inherits from the public D3D11 interface.
2022-03-17 11:35:19 +01:00
Joshua Ashton 0bc972697b [d3d11] Fix D3D11UserDefinedAnnotation declaration
Mark it as final too.
2022-03-17 11:35:19 +01:00
Georg Lehmann 2550cff149 [build] Cleanup build system.
No changes except dropping support for msvc before 15.3.
2022-03-17 01:38:58 +00:00
Georg Lehmann f1efc9dc9f [dxso] Emit spirv OpCross if we can. 2022-03-16 19:13:09 +00:00
Georg Lehmann 630fee59fc [dxso] Implement zerowins for Lerp.
Fixes #2545.
2022-03-16 19:13:09 +00:00
Joshua Ashton 1f88ee595f [d3d9] Don't expose D32 format
Not supported anywhere except REF device it seems... *sigh*

Supercedes: #2547
2022-03-16 19:11:16 +00:00
Philip Rebohle 4f8da62c34
[dxvk] Fix color write mask normalization
Previously we'd set too many bits by accident here. Also, we should
not modify partial write masks to include unnecessary bits. Only do
this if we can actually promote to a full write mask for consistency.
2022-03-16 19:31:37 +01:00
Liam Middlebrook f92c6ae859 [dxgi] Add DXVK_ENABLE_NVAPI envvar
Add a new environment variable DXVK_ENABLE_NVAPI as an environment-level
override for 'nvapiHack'. This will allow for DLSS (and other
NvAPI-backed features) to be available without the user manually writing
a configuration file, allowing for more seamless integration with
Proton's launch script.
2022-03-16 13:06:25 +01:00
Robin Kertels 115385d1d9 [d3d9] Update buffer seq number in FlushBuffer
How did I miss this?!
2022-03-15 01:24:54 +00:00
Robin Kertels 5f4fc56226 [d3d9] Update texture sequence number AFTER using it 2022-03-13 17:13:00 +00:00
Robin Kertels a6357a254d [d3d9] Fix CS thread synchronization for directly mapped buffers 2022-03-13 17:13:00 +00:00
Philip Rebohle 9eb2393d44
[dxvk] Normalize color write masks for non-RGBA formats 2022-03-13 16:24:53 +01:00
Philip Rebohle 4c429f044f
[util] Add another weeb game to the list of workarounds
Sophie is apparently D3D9 an we already have Lydie and Suelle in there,
so it's just this on missing from that series.
2022-03-13 05:38:06 +01:00
Philip Rebohle caa83247d9
[d3d9] Mark backend image as shared for shared resources
Otherwise, the backend may not transition the image to the correct
layout after each submission.
2022-03-13 02:32:56 +01:00
Robin Kertels bd1a2aa5a1 [dxvk] Force dedicated allocation for exportable images
The Nvidia driver does not set prefers-/requiresDedicatedAllocation
for exportable images on its own.

This makes DXVK ignore the dedicated allocation struct ptr
which also contains VkExportMemoryAllocateInfo or
VkImportMemoryWin32HandleInfoKHR.
2022-03-13 00:43:41 +01:00
Philip Rebohle 59e3c96976
[d3d9] Fix texture formats that can be exported 2022-03-12 22:24:05 +01:00
Derek Lesho 1d1d8adee6
[d3d9] Fix shared handle check for exporting images
Co-authored-by: Philip Rebohle <philip.rebohle@tu-dortmund.de>
2022-03-12 21:35:42 +01:00
Philip Rebohle c87660480d
[d3d11] Reimplement R11G11B10 UAV clears without R32 views 2022-03-12 16:23:42 +01:00
Philip Rebohle ee28f145dc
[d3d11] Explicitly handle R32-compatible UAV formats 2022-03-12 16:23:39 +01:00
Derek Lesho 6640cc350c
[d3d9] Add support for shared IDirect3DTexture9 resources. 2022-03-12 16:23:06 +01:00
Derek Lesho 48046e009c
[d3d11] Add support for shared ID3D11Texture2D resources. 2022-03-12 16:23:06 +01:00
Derek Lesho f9d94c82a5
[util] Add helpers for shared resource metadata access. 2022-03-12 16:23:06 +01:00
Derek Lesho 02d3e0d31f
[dxvk] Add shared handle access to DxvkImage memory.
Based off preliminary work from Josh.
2022-03-12 16:23:06 +01:00
Derek Lesho b5014401e7
[dxvk] Enable VK_KHR_EXTERNAL_MEMORY_WIN32 if available. 2022-03-12 16:23:03 +01:00
Philip Rebohle 5d4c8a6b51 [dxso] Fix ExpP instruction on Shader Model 2+ 2022-03-11 22:44:18 +00:00
Joshua Ashton 7151fbd598 [dxvk] Use new [[fallthrough]] attribute 2022-03-10 00:47:38 +01:00
Joshua Ashton b96f65be31 [dxbc] Use new [[fallthrough]] attribute 2022-03-10 00:47:38 +01:00
Joshua Ashton 63bd41f785 [d3d11] Add fallthrough comment to PickFormats
Silences a warning
2022-03-10 00:47:38 +01:00
Joshua Ashton dc6c350cf2 [d3d9] Add fallthrough comment to SetRenderState
Silences a warning
2022-03-10 00:47:38 +01:00
Joshua Ashton f4ab14cafa [dxso] Fix potential fallthrough in RasterizerOut 2022-03-10 00:47:38 +01:00
Joshua Ashton a5b0f2898a [dxso] Fix ExpP fallthrough 2022-03-10 00:47:38 +01:00
Joshua Ashton b2d96ee087 [d3d9] Fix fallthroughs in PickFormats 2022-03-10 00:47:38 +01:00
Joshua Ashton 53b9a723ac [d3d9] Set initial dirty state flags
We had a bug where initial state values caused the data to not get sent to the backend.

Let's fix that going forward and dirty everything we possibly can on device creation.
2022-03-09 23:33:12 +00:00
Joshua Ashton d4d4f0ca18 [d3d9] Fix default initialization of some state values
If we end up being the same as what we are, we don't dirty initially.
2022-03-09 23:33:08 +00:00
Philip Rebohle 89b1f025eb
[d3d10] Forward OpenSharedResource to D3D11 implementation
Trivial since the requested IID is passed by the application.
2022-03-04 19:13:30 +01:00
Philip Rebohle 23eead9e99
[meta] Release 1.10 2022-03-04 17:15:51 +01:00
Oschowa bb178baaf1 [util] Enable cached vertex and index buffers for The Evil Within
Large performance win.
2022-03-04 02:19:17 +01:00
Philip Rebohle f14bccc8c3
[util] Use CPU-cached constant buffers for Anno 1800
Sigh.
2022-03-03 17:33:14 +01:00
Philip Rebohle 1b196c24a4
[dxvk] Invalidate buffer in clearBuffer if possible 2022-03-01 21:39:35 +01:00
Philip Rebohle 5822a8de9f
[util] Enable cached vertex and index buffers for FFXIV
Fixes some weird performance issues on the Garlemald map. Doesn't seem
to affect performance in other areas.
2022-02-28 23:33:27 +01:00
Philip Rebohle 2bceeff5f2
[d3d11] Replace apitrace mode option with something more granular
And enable it only for vertex and index buffers in Nier Replicant.
2022-02-28 22:44:10 +01:00
Robin Kertels b42c07253e [d3d9] Fix sysmem readback 2022-02-24 07:49:55 +00:00
Robin Kertels a81c653b42 [d3d9] Fix synchronization after readback 2022-02-22 21:49:28 +00:00
Philip Rebohle a37cd46491
[util] Set maxDynamicImageBufferSize for Total War: Warhammer III
Massively increases performance since the game otherwise keeps
uploading a huge 48 MiB texture in every frame.
2022-02-22 05:48:25 +01:00
Philip Rebohle dbd4dad095
[d3d11] Introduce d3d11.maxDynamicImageBufferSize option 2022-02-22 05:31:27 +01:00
Philip Rebohle 33cf522647
[d3d11] Use appropriate memory types for directly mapped images 2022-02-22 05:31:20 +01:00
Philip Rebohle 177e70949c
[d3d11] Apply apitrace mode to image upload buffers 2022-02-22 01:30:43 +01:00
Robin Kertels 29d8812709 [d3d9] Don't set NeedsReadback for POOL_SYSMEM textures
... or in SetRenderTarget because we always do readback for render targets.
2022-02-21 13:43:04 +00:00
Robin Kertels b163a99bde [d3d9] Clean up texture locking
We had two code paths that largely did the same.
2022-02-21 00:08:41 +00:00
Robin Kertels dde83717f1 [d3d9] Rename WrittenByGPU to NeedsReadback 2022-02-21 00:08:41 +00:00
Robin Kertels 40444c1f50 [d3d9] Unify texture uploads 2022-02-21 00:08:41 +00:00
Robin Kertels 8b9db26536 [util] Add computeMipLevelOffset 2022-02-21 00:08:41 +00:00
Philip Rebohle 49e5357a87
[hud] Greatly simplify frame time graph rendering 2022-02-20 15:58:29 +01:00
Philip Rebohle 0709c5f5c7
[hud] Greatly simplify text rendering in the HUD 2022-02-20 15:58:29 +01:00
Philip Rebohle 8ce83cabca
[dxvk] Only mark transfer buffers as transient
Otherwise we may accidentally catch things like uniform buffers as well.
2022-02-20 15:58:26 +01:00
Robin Kertels fd23dcef64 [d3d9] Disable direct buffer mapping for RE games 2022-02-20 03:30:03 +00:00
Robin Kertels cf4c5c3422 [d3d9] Add option to disable direct buffer mapping 2022-02-20 03:30:03 +00:00
Robin Kertels 6d71eea516 [d3d9] Store buffer map mode in D3D9CommonBuffer 2022-02-20 03:30:03 +00:00
Robin Kertels 97ae14b6a0 [d3d9] Synchronize only to given sequence number in WaitForResources 2022-02-20 03:29:31 +00:00
Robin Kertels 917a8d00a2 [d3d9] Track last staging resource usage with a sequence number 2022-02-20 03:29:31 +00:00
Robin Kertels c12cd1952c [d3d9] Handle different mip chain lengths in UpdateTexture 2022-02-20 03:29:04 +00:00
Philip Rebohle 80f744549f
[dxvk] Get rid of spinlock when allocating GPU events
This is not performance-critical
2022-02-20 01:16:04 +01:00
Philip Rebohle 0ade12dc83
[dxvk] Use lock-free list for render pass instances
And replace the spin lock with a regular mutex.
2022-02-20 00:04:06 +01:00
Philip Rebohle 477cb617ac
[dxvk] Use lock-free list for compute pipeline lookup 2022-02-19 17:36:42 +01:00
Philip Rebohle 67e2ee1b26
[dxvk] Use lock-free list for graphics pipeline lookup
And use a proper mutex if we do have to synchronize,
so that we can avoid busy-waits.
2022-02-19 17:36:42 +01:00
Philip Rebohle a4fe43462c
[dxvk] Introduce lock-free list 2022-02-19 17:36:39 +01:00
Philip Rebohle c9750f6657
[dxvk] Don't use spinlocks for CS chunk pool
No reason to anymore since SRWLocks are fast enough here.
2022-02-19 16:57:09 +01:00
Philip Rebohle 520d62903a
[dxvk] Remove null check when setting vertex stride
Move the responsibility to the front-end instead.
2022-02-19 13:29:59 +01:00
Philip Rebohle 8e11630370
[d3d11] Set zero stride when binding null vertex buffer 2022-02-19 13:20:36 +01:00
Philip Rebohle 146fbd492f
[dxvk] Free existing staging buffer before creating a new one 2022-02-18 18:13:28 +01:00
Philip Rebohle 17a1b0ad44
[d3d11] Consider empty CS chunks when tracking resources
Avoids deadlocks if we track multiple resources and flush in between.
2022-02-18 14:21:53 +01:00
Philip Rebohle 9af3dce304
[dxvk] Track buffer as used in initBuffer
Git ate my commit when I was testing something...
2022-02-17 23:42:31 +01:00
Philip Rebohle 6ae5488157
[d3d9] Use initBuffer method 2022-02-17 20:58:31 +01:00
Philip Rebohle 18a960a1e1
[d3d11] Use initBuffer method 2022-02-17 20:58:31 +01:00
Philip Rebohle 080aa1a989
[dxvk] Add initBuffer method 2022-02-17 20:58:31 +01:00
Philip Rebohle f7e021f791
[dxvk] Add command buffer parameter to cmdFillBuffer 2022-02-17 20:58:31 +01:00
Philip Rebohle 102bceb30c
[dxvk] Remove unused clear methods 2022-02-17 20:58:31 +01:00
Philip Rebohle 752f98cf00
[d3d9] Use initImage to clear uninitialized image resources 2022-02-17 20:58:31 +01:00
Philip Rebohle 6d0e5dfd21
[d3d11] Use initImage to clear uninitialized image resources 2022-02-17 20:58:31 +01:00
Philip Rebohle 464c6810b3
[dxvk] Repurpose initImage method
This is now supposed to clear images of any type, and only to be
used for resource initialization after creation.
2022-02-17 20:58:31 +01:00
Philip Rebohle 34fd16b8f2
[d3d11] Add implicit flush after tracking sequence numbers
Flushing early when using a tracked resource may reduce stalls.
2022-02-16 20:49:00 +01:00
Philip Rebohle 2dfdc5ac3b
[d3d11] Enable stall tracking for timestamp queries
Because games are dumb and don't understand that the GPU doesn't
work synchronously with the render thread.
2022-02-16 19:59:11 +01:00
Philip Rebohle d45f5a8d79
[d3d11] Handle subresource field in copy/move operations
Derp.
2022-02-15 13:00:10 +01:00
Philip Rebohle e70f9f92a6
[util] Enable apitrace mode for Nier Replicant
Game is broken and reads back dynamic vertex/index buffers over PCI-E.
2022-02-14 09:20:41 +01:00
Philip Rebohle 4e464327ef
[util] Bump maxImplicitDiscardSize for Quantum Break
Otherwise we're synchronizing and frame times are garbage.
2022-02-14 04:21:46 +01:00
Philip Rebohle e1b3bc45ce
[d3d11] Add d3d11.maxImplicitDiscardSize option 2022-02-14 04:17:56 +01:00
Philip Rebohle 63bf928ab5
[hud] Display GPU synchronization in HUD 2022-02-14 03:28:45 +01:00
Philip Rebohle 4d9b464f7c
[d3d9] Use new waitForResource method 2022-02-14 03:15:47 +01:00
Philip Rebohle 2f80f8847e
[d3d11] Use new waitForResource method 2022-02-14 03:15:47 +01:00
Philip Rebohle 0c99b17081
[dxvk] Introduce DxvkDevice::waitForResource
Blocks on the queue thread's condition variable instead of busy-waiting,
and tracks synchronization with new stat counters. Cleanup is rearranged
to minimize delays before signals and resources are notified.
2022-02-14 03:15:47 +01:00
Philip Rebohle 25ebf94873
[dxvk] Display barrier count in draw call HUD item 2022-02-14 01:10:15 +01:00
Philip Rebohle 316e252a27
[dxvk] Add stat counter for pipeline barriers 2022-02-14 01:08:55 +01:00
Philip Rebohle b5078a7ec0
[dxvk] Reduce context staging buffer size to 4 MiB
Same idea as before, just create a temporary buffer for larger resources.

This can avoid frequent Vulkan memory allocations and deallocations since
many small buffers are more likely to fit into a single memory chunk than
a small number of large buffers, thus reducing the overall memory footprint.
2022-02-14 01:01:34 +01:00
Philip Rebohle 08ecd49c66
[dxvk] Don't suballocate large staging buffer allocations
Otherwise we'll risk wasting almost half the staging buffer memory.
Creating a temporary buffer is cheap enough, so just do that.
2022-02-14 01:00:48 +01:00
Philip Rebohle b9201db554
[dxvk] Remove unused trimStagingBuffers method 2022-02-13 02:08:20 +01:00
Philip Rebohle 95b7e6c030
[dxvk] Rework HUD font texture initialization
We really shouldn't need a separate context for this.
2022-02-13 02:08:20 +01:00
Philip Rebohle 425fce9200
[dxvk] Introduce transient memory flag for staging buffers
Potentially reduces fragmentation by putting short-lived staging buffers
and sysmem resources created by the application into different memory pools.
2022-02-13 02:08:20 +01:00
Philip Rebohle 9d4be00fa7
[dxvk] Allow large sysmem allocations on 64-bit platforms again
Since we frequently discard staging buffers now, having larger chunks
is actually beneficial again.
2022-02-13 02:08:20 +01:00
Philip Rebohle d262bebd90
[dxvk] Remove DxvkStagingDataAlloc
Unused and overly clunky.
2022-02-13 02:08:20 +01:00
Philip Rebohle 8518572d13
[dxvk] Use DxvkStagingBuffer in DxvkContext 2022-02-13 02:08:20 +01:00
Philip Rebohle 1b88bc624a
[dxvk] Remove unused updateImage function 2022-02-13 02:08:20 +01:00
Philip Rebohle 3b833988fe
[dxvk] Use staging buffer for gamma ramp uploads 2022-02-13 02:08:19 +01:00