Stenzek
7353cae707
GLInterface: Implement core and shared context creation for WGL
2016-11-29 20:04:32 +10:00
Stenzek
b81dee8b9a
OGL: Support full-resolution frame dumping
2016-11-28 21:54:56 +10:00
Stenzek
93221e7f48
OGL: Move frame rendering procedures to seperate methods
2016-11-28 21:54:56 +10:00
Stenzek
1c1a686f63
Vulkan: Support full resolution frame dumping
2016-11-28 21:54:55 +10:00
Stenzek
adc22a5e20
VideoCommon: Add method for calculating full-scale framedump rectangles
2016-11-28 21:54:54 +10:00
Stenzek
8d48319414
Vulkan: Validate the pipeline cache before using it
...
This ensures that if a user changes adapters or vendors we're not passing
invalid data to the driver.
2016-11-28 21:21:55 +10:00
Stenzek
9604b336c8
Vulkan: Don't destroy the device's pipeline cache on MSAA mode change
...
The user could switch back again, and this would mean this data would be
lost. Disk space is cheap, and it's not going to be much.
2016-11-28 21:21:54 +10:00
Stenzek
aac66a1b61
Vulkan: Implement a pipeline UID cache
...
This stores enough information to recreate the pipeline, including the
shader UIDs, blend/depth/rasterization state, primitive and vertex format.
2016-11-28 21:21:53 +10:00
Stenzek
6db0ee9561
VideoCommon: Remove backbuffer size parameters from methods
...
We have the s_backbuffer_{width,height} fields to represent this, so
there's no point in passing them as parameters every time.
2016-11-28 20:14:59 +10:00
Stenzek
a83ca587c6
VideoCommon: Make FramebufferManagerBase::GetTargetSize public
2016-11-28 20:14:59 +10:00
Stenzek
a0a62c0f46
VideoConfig: Add option for full-resolution frame dumping
2016-11-28 20:14:59 +10:00
Stenzek
6d0b9b816f
VideoCommon: Support dumping frames to images
...
This is mainly for potential Android fifoci usage, and thus is not
exposed anywhere in the UI. To enable, set DumpFramesAsImages under
Settings in GFX.ini.
2016-11-23 12:07:49 +10:00
Stenzek
1aecf90765
DolphinNoGUI: Receive resize events and notify backend when this occurs
...
Solves the viewport sizing issue for backends that cannot tell when the
window is resized (Vulkan on NV at least).
2016-11-19 23:03:20 +10:00
Stenzek
d6d3341183
D3D: Fix strided XFB copies
2016-11-19 20:29:51 +10:00
Stenzek
6b88a854a7
Vulkan: Handle strided XFB copies
...
Where src_rect.width * 2 != dst_stride.
2016-11-19 20:29:47 +10:00
Stenzek
725ef4c5dc
Vulkan: Pass target_rect to framebuffer draw methods
...
Fixes the black borders in frame dumps when the window was not sized to
the framebuffer aspect ratio.
2016-11-18 22:55:22 +10:00
Stenzek
3c92b35422
Vulkan: Use multiple command pools, one per frame
...
Instead of resetting two command buffers, now we only have to call
vkResetCommandPool once at the start of a frame.
NV's recommends using one pool per frame/thread. May offer a very small
boost in performance on some systems.
2016-11-15 00:40:15 +10:00
Stenzek
e83bf5705a
JitArm64: Don't emit normal block exit for branch-to-self instructions
...
The normal block exit is redundant as the exception exit will jump out of
the block first, meaning this code is never executed.
2016-11-15 00:28:14 +10:00
Stenzek
c657134c86
Jit64: Treat branch-to-self instruction as an idle loop
2016-11-15 00:26:59 +10:00
Stenzek
bd67adb362
Vulkan: Use correct sample count for EFB pokes with MSAA enabled
2016-11-14 20:24:16 +10:00
Stenzek
89176fe2ab
Vulkan: Fix crash on EFB poke
2016-11-14 20:19:59 +10:00
Stenzek
160fee6791
Vulkan: Handle maxImageCount of zero when creating swap chain
...
anv seems to set this to zero, which is fine according to the spec, but
we were using it as a maximum, which was resulting in a swap chain
without any buffers being created.
2016-11-11 23:33:40 +10:00
Stenzek
38c3ca5cd4
Vulkan: Implement asynchronous frame dumping
2016-11-10 01:07:46 +10:00
Stenzek
c09ce029df
Vulkan: Don't use FIFO_RELAXED present mode for vsync.
2016-11-07 19:22:27 +10:00
Stenzek
c880c37244
Vulkan: Rename screenshot buffer to frame dump buffer.
...
Name makes more sense given the methods it calls in the base class.
2016-11-03 22:38:48 +10:00
Stenzek
9aed27cdcf
Vulkan: Combine frame dumping and present into one command buffer.
...
Small optimization that should make things slightly more efficient when
frame dumping is enabled.
2016-11-03 22:38:48 +10:00
Stenzek
690a6deeb3
Vulkan: Fix swapped top/bottom images in TAB stereo mode
2016-11-03 22:33:24 +10:00
Stenzek
70eb904536
Vulkan: Fix incorrect geometry shader input/output usage
2016-11-03 22:33:24 +10:00
Stenzek
5250f3c6a4
Vulkan: Add missing call to ObjectCache::RecompileSharedShaders
...
This was causing issues when the stereo mode was changed at runtime.
2016-11-03 22:33:24 +10:00
Stenzek
d67877d27e
Vulkan: Fix fast clear path not being used in all cases
2016-11-03 22:33:24 +10:00
Stenzek
5182e6b549
Vulkan: Implement virtual/real XFB support
2016-11-03 22:33:24 +10:00
Stenzek
3593fa27ab
Vulkan: Move CopyRectangleFromTexture to TextureCache
2016-11-03 22:01:55 +10:00
Stenzek
01b3c0f036
Vulkan: Make TextureCache::TCacheEntry public
2016-11-03 22:01:54 +10:00
Stenzek
b066d51dfa
Vulkan: Remove parameters/members of single-instance classes
...
There's not a lot of point in passing these around or storing them
(texture cache/state tracker mainly) as there will only ever be a single
instance of the class.
Also adds downcast helpers such as Vulkan::Renderer::GetInstance().
2016-11-03 22:01:54 +10:00
Stenzek
176b00ded7
Vulkan: Handle forced aspect ratio changes at runtime
2016-10-08 18:59:46 +10:00
Stenzek
010514bd39
Vulkan: Use correct source format to determine palette size
...
Fixes blur in fortune street fifologs.
2016-10-06 21:55:27 +10:00
Stenzek
b6d09c61ed
TextureCache: Don't re-use pooled textures within the same frame
...
This is an issue because a driver may have to maintain two copies of a
texture if it batches all uploads together at the start of a frame.
In the Vulkan backend, we do something similar to avoid breaking out of a
render pass to copy a texture from the streaming buffer to the destination
image.
This was causing issues in the sms-bubbles fifolog, where an EFB copy to
the same address of a previously-used texture caused the previous texture
to be re-used again for a different image later on in the frame, causing
the original contents to be discarded.
2016-10-06 20:23:37 +10:00
Stenzek
b39ac950eb
Vulkan: Don't save borders to screenshots/frame dumps
...
This matches the behavior on GL, making for easier comparisons.
2016-10-05 22:28:51 +10:00
Stenzek
400ba3c7e5
Vulkan: Stop dumping frames on shutdown if left enabled
2016-10-05 22:28:51 +10:00
Stenzek
c422fb7e82
Vulkan: Set alpha channel of swap chain buffers to 1.0.
...
Copying the alpha channel from the game causes issues with frame dumping,
since we're using a buffer directly from the GPU as a source for AVIDump.
2016-10-05 22:02:04 +10:00
Stenzek
db09c05eec
Vulkan: Fix bug with palette converted EFB copies
...
This happened when the source texture was an EFB copy, therefore it had
not been populated prior to the draw command buffer being executed, and
the conversion was occurring in the init command list.
2016-10-04 22:30:37 +10:00
Stenzek
28e5fa8d26
Vulkan: Handle both destination alpha and logic ops being enabled
...
Same way as GL with the dual-pass fallback. Not highly accurate, but does
fix the Kirby shadow bug.
2016-10-03 19:11:50 +10:00
Stenzek
f595fe080f
Vulkan: Fix bug with fractional LOD bias and min/max LOD
2016-10-03 19:11:48 +10:00
Stenzek
5e29508b8f
Vulkan: Fix vsync behavior when throttler is temp disabled
2016-10-03 19:11:48 +10:00
Stenzek
b193282830
Vulkan: Correct logic for handling target and window size changes
...
Should fix a possible reference to deleted framebuffers, as well as fixing
the issues with the render area being correct if the game's source area
changes, or auto-scaling is enabled.
2016-10-03 19:11:47 +10:00
Stenzek
7d14b9b48b
Vulkan: Add missing call to TextureCache::OnConfigChanged
...
This was preventing certain settings from being updated when changed at
runtime.
2016-10-03 19:11:47 +10:00
Stenzek
4a8766cec4
Vulkan: Fix resource leaks present at shutdown and mode changes
...
Infrequent, but still happened.
2016-10-03 19:11:47 +10:00
Stenzek
1286c309e3
Vulkan: Fix compilation on 32-bit targets
2016-10-03 19:11:47 +10:00
Stenzek
a8194cff3c
VideoNull: Set all fields in backend_info
...
A few of these were missing, which could cause the adapter list to remain
visible after switching to null, for example.
2016-10-01 02:40:03 +10:00
Stenzek
5f66cf5ed7
Vulkan: Only submit init/upload command buffer when it has commands
...
This way we're not submitting empty buffers when it's unnecessary.
2016-10-01 02:40:03 +10:00
Stenzek
bac8c2d441
Vulkan: Work around indexed fragment output bug on AMD drivers
2016-10-01 02:40:02 +10:00
Stenzek
f6cdc38c8b
Vulkan: Use render-pass based clears where possible
2016-10-01 02:40:02 +10:00
Stenzek
c290398320
Vulkan: Ensure fast path is used for non-RGBA formats when clearing
2016-10-01 02:40:02 +10:00
Stenzek
f4944f006d
Vulkan: Support frame dumping/screenshots
2016-10-01 02:40:02 +10:00
Stenzek
4a4f6cc135
Android: Add video backend choice to settings
2016-10-01 02:40:02 +10:00
Stenzek
77a128ab87
Implement experimental Vulkan backend
2016-10-01 02:40:01 +10:00
Stenzek
fdd954e7e7
Common: Add a Semaphore wrapper class
2016-10-01 01:09:12 +10:00
Stenzek
828aac7890
VideoBackends: Make TextureCache::CompileShaders return a bool
2016-10-01 01:09:12 +10:00
Stenzek
6a99cbd9fc
VideoCommon: Call Renderer::SurfaceChanged on render parent resize
...
This is needed because for some reason the WSI for NV Vulkan drivers
doesn't return VK_ERROR_OUT_OF_DATE_KHR, so there is no other way to know
that a resize has occured apart from polling, which is a poor solution for
X11 (since it is blocking).
2016-10-01 01:09:12 +10:00
Stenzek
5346078791
VideoCommon: Add config fields for multithreading and validation layers
2016-10-01 01:09:12 +10:00
Stenzek
09638e714e
VideoCommon: Extend DriverDetails to support both OpenGL and Vulkan
2016-10-01 01:09:12 +10:00
Stenzek
a71381e80a
VideoCommon: Add APIType entry for Vulkan
2016-10-01 01:09:11 +10:00
Stenzek
75e4e42e56
ShaderGen: Work around issue with glslang and dynamic vector subscripts
...
Seems that vec[eye] will select vec.x no matter what the value of eye is.
2016-10-01 01:09:11 +10:00
Stenzek
7f3a876aae
ShaderGen: posmtx should be a 4-component unsigned byte
...
This is a global change across backends, so should be tested for
regressions.
2016-10-01 01:09:11 +10:00
Stenzek
d9c034e8cc
ShaderGen: Specify attribute/output locations/bindings explicitly
...
This also shifts the SSBO index from index 3 to index 0.
2016-10-01 01:09:11 +10:00
Stenzek
9f541e490d
OGL: Handle case where both constant alpha and logic op is enabled
2016-09-30 23:18:14 +10:00
Stenzek
951fc44d86
Externals: Add glslang from Vulkan SDK v1.0.21.1
2016-08-11 22:40:32 +10:00
Stenzek
49e57df2b6
Externals: Add Vulkan headers from SDK v1.0.21.1
2016-08-11 22:40:32 +10:00
Stenzek
bce8097712
D3D11: Support texture dumping of non-zero mipmap levels
2016-08-10 23:45:25 +10:00
Stenzek
e169d54f3c
D3D11: Fix CPU EFB color reads when MSAA is enabled
...
Also swaps the byte order from RGBA->BGRA to match GL/D3D12, and what
the read handler is expecting.
Depth reads will now return the minimum depth of all samples, instead of
the average of all samples.
2016-05-19 22:51:00 +10:00
Stenzek
89e54fbd6c
OGL: Work around slowdown of glMapBufferRange with SSBO on NVIDIA drivers
...
Using glMapBufferRange to read back the contents of the SSBO is extremely
slow on NVIDIA drivers. This is more noticeable at higher internal
resolutions. Using glGetBufferSubData instead does not seem to exhibit
this slowdown.
2016-05-19 21:24:09 +10:00
Stenzek
ccf9470241
D3D12: Specify read/write ranges when calling Map/Unmap
2016-05-08 23:18:59 +10:00
Stenzek
fde7dee652
D3D12: Fix invalid CopyTextureRegion call in CopyRectangleFromTexture
...
This was occuring when the source texture was larger than the destination
texture, but the source rect was <= dest rect, so the copy is valid.
2016-05-08 23:18:58 +10:00
Stenzek
9bff187547
D3D12: Cleanup startup/shutdown process
...
Sorts out references that cause some modules to be kept around after
backend shutdown.
Should also solve the issue with errors being thrown due to the config
being loaded after device creation, leading to the incorrect device being
used in a multi-adapter system.
2016-05-08 23:18:58 +10:00
Stenzek
4269abdc3e
D3D12: Implement perf query support
2016-05-08 23:18:57 +10:00
Stenzek
25d5da0ea3
D3D12: Remove D3D11 header references
2016-05-08 23:18:56 +10:00
Stenzek
6f3573dda8
D3D12: Implement XFB encoding/decoding (support Real XFB)
2016-05-08 23:18:51 +10:00
Stenzek
3372bfa6ab
D3D12: Remove feature level checks
...
We don't create a device below feature level 11_0 anyway, so no point
checking, we can just assume support.
2016-05-08 12:08:25 +10:00
Stenzek
063761fbd2
D3D12: Don't add padding when allocating within empty StreamBuffer
...
Resources are already aligned to an address larger than any of our
requirements, anyway.
2016-05-08 12:08:25 +10:00
Stenzek
0c27aae7d3
D3D12: Improve output of shader compiler errors
...
These were completely broken due to lack of c_str(). We also output
warnings now as well (these can be useful).
2016-05-08 12:08:25 +10:00
Stenzek
acfa93372e
D3D12: Refactoring and cleanups
...
Moves render target restoring to RestoreAPIState, this also means no need
to manually restore after allocating in a buffer that caused execution,
because the manager restores it for us.
Remove a method that wasn't used from D3DUtil.cpp, and fixes a few errors
in EFB poke drawing.
2016-05-08 12:08:25 +10:00
Stenzek
7ec1fce741
D3D12: Fix error with >1xIR/MSAA EFB depth access
2016-05-08 12:08:25 +10:00
Stenzek
ac1cd8279b
D3D12: Implement GPU-based bounding box
2016-05-08 12:08:25 +10:00
Stenzek
32599559db
D3D12: Use helper method for binding EFB render targets
2016-05-08 12:08:25 +10:00
Stenzek
984da2d624
D3D12: Use signed ints for viewport origin
...
Fixes black screen when crop is enabled.
2016-05-08 12:08:25 +10:00
Stenzek
a8c4d6c242
D3D12: Allow large texture uploads (>64MiB) by using temporary buffer
...
This is not optimal, but for those texture packs with extremely large
images, it won't crash. Releasing after the frame completes is an option
too, however, there is the risk of running out of memory by doing this.
2016-05-08 12:08:25 +10:00
Stenzek
e6b2212ec0
ShaderGen: Only specify storage qualifier in interface block when needed
...
Drivers that don't support GL_ARB_shading_language_420pack require that
the storage qualifier be specified even when inside an interface block.
AMD's driver throws a compile error when "centroid in/out" is used within
an interface block.
Our previous behavior was to include the storage qualifier regardless, but
this wasn't working on AMD, therefore we should check for the presence of
the extension and include based on this, instead.
2016-03-30 00:42:50 +10:00
Stenzek
53cf42fb06
D3D11: Fix some cases where render target switches desynced StateManager
...
This was occuring in certain EFB copy patterns, leaving the textures
unbound for the next draw call.
2016-03-26 00:01:26 +10:00
Stenzek
63e4e07683
D3D11: Simplify MSAA depth texture resolving
...
This also fixes EFB depth buffer copies when MSAA is enabled.
2016-03-26 00:00:39 +10:00
Stenzek
0a96e2f531
D3D11: Fix texture dumping, for both single and multi-mip textures
2016-02-28 17:24:47 +10:00
Stenzek
c793459b88
D3D12: Fixed issue where EFB copies could end up corrupted after reset
...
Also prevents previously-released textures from ending up in a descriptor
table.
2016-02-28 17:24:46 +10:00
Stenzek
74275bdfe3
D3D12: Don't keep screenshot/encoder buffers mapped
...
Readback heaps do not support persistent mapping. See D3D12 docs.
2016-02-28 17:24:09 +10:00
Stenzek
c4d79d6db3
D3D12: Add helper method for setting both viewport and scissor rect
...
Simplfies making changes, as well as keeping the two in sync.
2016-02-28 17:18:46 +10:00
Stenzek
13e143de38
D3D12: Optionally prevent StreamBuffer from executing command list
...
This applies to callers that do not have full knowledge of the command
list state, and thus, cannot restore it should allocations cause command
list execution. Instead we reallocate a new buffer. Should not happen
often enough for this to be a concern, as it's mainly for the utility
classes.
2016-02-28 17:18:46 +10:00
Stenzek
04257029e0
D3D12: Don't enumerate outputs, it's not used anywhere
...
The D3D12 backend does not support exclusive fullscreen.
2016-02-28 17:18:45 +10:00
Stenzek
9efe66509d
D3D12: Fix crash/errors when switching MSAA modes while running
2016-02-28 17:18:44 +10:00
Stenzek
1d909ec7a4
D3D12: Implement non-blocking EFB access when EFB has not been modified
2016-02-28 17:18:43 +10:00
Stenzek
6bbf836ea9
D3D12: Simplify and fix MSAA EFB depth copy path
2016-02-28 17:18:43 +10:00
Stenzek
649b94338e
D3D12: Cleanup/refactoring of teardown process
2016-02-28 17:18:42 +10:00
Stenzek
759b77474d
D3D12: Use std::thread for worker thread
...
Using CreateThread can create issues if any CRT calls are made, as
thread-specific data may not be initialized. Additionally, TerminateThread
is not a good idea for similar reasons, and may not free CRT resources.
2016-02-28 17:18:41 +10:00