Commit Graph

3807 Commits

Author SHA1 Message Date
Connor McLaughlin
1b1662773e
Merge pull request #8048 from stenzek/vulkan-negative-scissor-rect
Vulkan: Don't set a negative offset in scissor rect
2019-04-28 23:37:51 +10:00
Connor McLaughlin
d2d8d7ce90
Merge pull request #8051 from JosJuice/efb-constexpr
Turn EFB_WIDTH/EFB_HEIGHT into constexpr
2019-04-28 20:37:25 +10:00
JosJuice
ced2306fc5 Turn EFB_WIDTH/EFB_HEIGHT into constexpr
https://bugs.dolphin-emu.org/issues/11692#note-5

Also change const into constexpr while we're at it.
2019-04-28 11:50:17 +02:00
Connor McLaughlin
f7199397a0
Merge pull request #8047 from stenzek/d3d-feature-level-10
D3D11: Fix Dolphin crashing on feature level 10.0 devices
2019-04-28 17:07:37 +10:00
Stenzek
53af27b133 FramebufferManager: Fix invalid transitions for 1xIR in EFB cache 2019-04-28 16:13:49 +10:00
Stenzek
51154d6907 FramebufferShaderGen: Don't emit SV_SampleIndex when not using SSAA 2019-04-28 15:26:21 +10:00
Stenzek
d5b3595314 TextureCache: Fix incorrect format string in GetXFBTexture() 2019-04-28 14:32:37 +10:00
iwubcode
c2abf092e9 VideoCommon: Do not reset the gather pipe (WPAR buffer) as this does not reflect what actually happens on real hardware. On hardware, there's no way for the GPU to signal to reset the WPAR buffer. Instead, the WPAR buffer will be reset automatically upon receiving 32 bytes or when spr[WPAR] is written 2019-04-26 22:20:53 -05:00
Connor McLaughlin
59c0e51f5b
Merge pull request #8000 from stenzek/more-videocommon-cleanup
Additional cleanup/fixes from VideoCommon merge
2019-04-22 12:50:22 +10:00
Connor McLaughlin
93be178f28
Merge pull request #7857 from stenzek/opengl-stereo
OGL: Fix stereoscopy
2019-04-22 12:41:12 +10:00
Connor McLaughlin
44d5a71e27
Merge pull request #7965 from jordan-woyak/condvar-fixes
Minor changes to usages of std::condition_variable.
2019-04-21 23:50:53 +10:00
Stenzek
9577d0641b ShaderCache: Fix crash at shutdown on shared shader compilation fail 2019-04-21 14:28:14 +10:00
Stenzek
f8c1ba409c Replace EFBRectangle/TargetRectangle with MathUtil::Rectangle 2019-04-21 14:28:14 +10:00
Connor McLaughlin
5b4ebccf6b
Merge pull request #8012 from stenzek/bounding-box-shaders
PixelShaderGen: Don't emit bounding box shader code for old UIDs when disabled
2019-04-21 13:27:05 +10:00
Stenzek
b09a0e1a60 TextureCache: Check for out-of-range partial copy rect after scaling 2019-04-21 12:54:32 +10:00
Stenzek
3791262d96 TextureCache: Use linear filtering on y-scaled and >1xIR VRAM copies 2019-04-21 12:41:15 +10:00
Stenzek
708bd3d9f7 TextureCache: Simplify XFB reconstruction
This also better handles in-memory interlaced XFB data placed by the CPU
by considering the stride from the VI.
2019-04-21 12:41:15 +10:00
Stenzek
dbaba0062a TextureDecoder: Move XFB decoding to Common
This was previously missing for generic (which is used on ARM).
2019-04-21 12:41:15 +10:00
Stenzek
eddde3e6c8 ShaderGen: Use interface blocks when geometry shaders are supported
We don't use explicit locations in OpenGL currently, so this breaks
when we use alternative names in the geometry shaders.
2019-04-20 23:44:33 +10:00
Stenzek
c6b45c5ca9 ShaderCache: Fix crash if pipeline from uid cache fails creation 2019-04-20 23:16:29 +10:00
Stenzek
96aa762d1e PixelShaderGen: Don't emit bounding box shader code for old UIDs when disabled
If bounding box is enabled when a UID cache is created, then later disabled,
we shouldn't emit the bounding box portion of the shader.

Fixes pipeline creation errors on D3D12 backend for this case.
2019-04-20 23:16:14 +10:00
Connor McLaughlin
c26f53bf84
Merge pull request #7999 from stenzek/pipeline-cache-data
Implement pipeline data cache for OpenGL and D3D12
2019-04-20 12:56:02 +10:00
Jordan Woyak
779e618046 VideoCommon: Change free-look's middle-mouse action to roll the camera. 2019-04-17 05:10:11 -05:00
Stenzek
00b83b4196 ShaderCache: Utilize pipeline cache data where possible 2019-04-16 00:39:43 +10:00
Stenzek
61a656570e AbstractPipeline: Support returning "cache data"
"Cache data" can be used to assist a driver with creating pipelines by
using previously-compiled shader ISA.
2019-04-16 00:09:47 +10:00
Jordan Woyak
d34a9afe04 Minor changes to usages of std::condition_variable. 2019-04-06 17:39:25 -05:00
spycrab
e44433c2f6
Merge pull request #7955 from Techjar/netplay-golf-mode
NetPlay golf mode
2019-04-05 17:25:23 +02:00
Techjar
6c393f9ff4 Add imgui golf mode overlay 2019-04-05 07:01:03 -04:00
Techjar
a318c55cec PixelShaderGen: Fix bounding box coordinates being offset by 1 pixel 2019-04-03 10:09:05 -04:00
Stenzek
f2d8c8d2a8 D3D12: Fix rare case where command list was executed with open queries 2019-04-01 20:49:24 +10:00
Connor McLaughlin
32e330eb35
Merge pull request #7950 from stenzek/out-of-range-efb-copies
BPStructs: Gracefully handle out-of-range EFB copies
2019-04-01 13:08:11 +10:00
Stenzek
113bd60fe7 Implement D3D12 backend 2019-04-01 11:24:55 +10:00
Stenzek
378b605669 BPStructs: Gracefully handle out-of-range EFB copies
Since the copy X and Y coordinates/sizes are 10-bit, the game can configure a
copy region up to 1024x1024. Hardware tests have found that the number of bytes
written does not depend on the configured stride, instead it is based on the
size registers, writing beyond the length  of a single row. The data written
for the pixels which lie outside the EFB bounds does not wrap around instead
returning different colors based on the pixel format of the EFB.

This suggests it's not based on coordinates, but instead on memory addresses.
The effect of a within-bounds size but out-of-bounds offset
(e.g. offset 320,0, size 640,480) are the same.

As it would be difficult to emulate the exact behavior of out-of-bounds reads,
instead of writing the junk data, we don't write anything to RAM at all for
over-sized copies, and clamp to the EFB borders for over-offset copies.
2019-03-31 22:05:53 +10:00
Stenzek
3c455de799 FramebufferManager: Fix broken EFB peeks in GL with MSAA on 2019-03-30 03:14:29 +10:00
Stenzek
087b11e780 TextureCacheBase: Fix possible crash on shutdown with deferred EFB copies 2019-03-29 20:54:44 +10:00
Connor McLaughlin
f3fadd7302
Merge pull request #7869 from stenzek/d3dcommon
D3D: Move sharable D3D11/D3D12 code to common library
2019-03-29 20:40:06 +10:00
Stenzek
d66d778bae PixelShaderGen: Use subgroup reduction operations for bounding box 2019-03-29 20:06:56 +10:00
Connor McLaughlin
1304e963ad
Merge pull request #7895 from stenzek/framebuffer-corruption
FramebufferManager: Don't discard in ReinterpretPixelData
2019-03-29 20:03:13 +10:00
Connor McLaughlin
154eeae8ae
Merge pull request #7915 from stenzek/bbox-scaled-update
VideoBackends: Scale bounding box rectangle in the pixel shader
2019-03-29 19:59:25 +10:00
Stenzek
3b86c93285 VertexManagerBase: Increase vertex/uniform buffer sizes
ZTP was uploading 10MB+ of uniforms per frame, reducing paralellism by
forcing GPU waits.
2019-03-29 19:52:38 +10:00
Stenzek
2a4bca8b4a VideoBackends: Verify validity before doing full renderer init
Prevents attempting to create invalid MSAA configurations, for example.
2019-03-29 19:52:38 +10:00
Stenzek
e9cfcf4632 ShaderCache: Don't create invalid logic op pipelines
uint_output will never be set unless logic ops are enabled. Also clears
the bits so we're not compiling unused pixel shaders on the other
backends.
2019-03-28 20:35:02 +10:00
Stenzek
5732769a3c ShaderCache: Set correct primitive in dummy ubershader pipelines 2019-03-28 20:35:02 +10:00
Stenzek
7dd9d74eec RenderBase: Add EFBHasAlphaChannel() helper 2019-03-28 20:35:02 +10:00
JMC47
b4b0ebae27
Merge pull request #7934 from stenzek/stretched-efb-peeks
FramebufferManager: Fix EFB peek cache stretching last tile
2019-03-27 10:53:47 -04:00
Stenzek
92fa6c34cd FramebufferManager: Fix EFB pokes being offset by 1 in D3D 2019-03-27 23:28:34 +10:00
Stenzek
c28393d6f9 FramebufferManager: Fix EFB peek cache stretching last tile 2019-03-27 22:12:40 +10:00
spycrab
f7897778ff NetPlayChatUI: Add activate chat hotkey 2019-03-26 18:13:32 +01:00
Stenzek
16294acd2a VideoBackends: Scale bounding box rectangle in the pixel shader 2019-03-25 18:47:58 +10:00
Stenzek
cff60f5ba0 FramebufferManager: Only clear tile array when valid
When looking into the Faron Woods fifolog, I noticed this code was quite
high in the profile (~10%). Clearing 4096 entries from the vector isn't
needed every draw, we only need to do this when the cache was actually
valid in the first place.

Should provide a slight general performance boost.
2019-03-24 15:03:04 +10:00