 Henrik Rydgard
|
5b1b4329fd
|
GL: some extra error checks for texture loading
|
2017-01-29 14:19:58 +01:00 |
|
|
Henrik Rydgard
|
d7da981a62
|
Make the Dangan Ronpa hack use the compat system instead of hacky code
|
2017-01-29 00:12:20 +01:00 |
|
|
Henrik Rydgard
|
e59cafe177
|
Don't rely on gstate_c.spline or bezier during shader compile, as that doesn't work with the cache. Should fix #9249
|
2017-01-28 22:34:10 +01:00 |
|
|
Henrik Rydgård
|
d5d9d87dc6
|
Update the shader cache version to avoid obsolete shaders sticking around. Should fix #9249.
|
2017-01-28 22:11:48 +01:00 |
|
|
Henrik Rydgard
|
182d87cc15
|
Minor cleanup in GPU dispatch tables
|
2017-01-28 12:11:56 +01:00 |
|
|
Henrik Rydgard
|
ab6bc4cb16
|
Restores a lost optimization from uv prescale - avoiding flushing. Should fix #9219.
Also #ifdef away some invalid-address checks in hot paths in the GPU.
|
2017-01-28 11:39:34 +01:00 |
|
|
Henrik Rydgard
|
9c55e1e0de
|
Only enable RAM Clears for the SOCOM games that require it.
Should remove the performance impact of #8994 which is bigger than
expected, it seems (cache pollution?)
|
2017-01-28 10:11:43 +01:00 |
|
|
xebra
|
a460acf411
|
[spline/bezier]Fix to build GLES.
|
2017-01-27 13:32:36 +09:00 |
|
|
xebra
|
0e6c2b00cf
|
[spline/bezier]Fix compatibility for GLES.
|
2017-01-27 13:32:35 +09:00 |
|
|
Henrik Rydgard
|
98e0ccf1e1
|
Vertex JIT: Add some missing functions to ARM/ARM64 decoders.
|
2017-01-25 20:13:29 +01:00 |
|
|
Henrik Rydgård
|
524854ac16
|
Merge pull request #9233 from xebra/spline_bezier
[spline/bezier]Remove wasted code.
|
2017-01-24 23:33:51 +01:00 |
|
|
xebra
|
133a9d78a8
|
[spline/bezier]Remove wasted code.
|
2017-01-25 06:35:36 +09:00 |
|
|
Henrik Rydgard
|
d85888dc2b
|
Table-driven dirty-flags for Vulkan and D3D9. Minor Vulkan optimization.
|
2017-01-24 18:12:20 +01:00 |
|
|
Henrik Rydgard
|
1fe32d380b
|
Update lots of dirty-flag table entries
|
2017-01-24 18:12:20 +01:00 |
|
|
Henrik Rydgard
|
c70265d7f2
|
Use table-driven dirtying instead of callbacks in GPU_GLES
|
2017-01-24 18:12:20 +01:00 |
|
|
Henrik Rydgard
|
e4cb44c243
|
Merge the texture flags into the dirty flags
|
2017-01-24 18:12:20 +01:00 |
|
|
Henrik Rydgard
|
c289a2c7bf
|
Split TextureChanged into two bools, so we can later convert to dirtyflags
|
2017-01-24 18:12:20 +01:00 |
|
|
Henrik Rydgard
|
7d60ec73be
|
Move framebufChanged into the dirty flag field
|
2017-01-24 18:12:20 +01:00 |
|
|
Henrik Rydgard
|
b423998c36
|
Move uniform dirty-flag to gstate_c, opening up for other uses.
|
2017-01-24 18:12:20 +01:00 |
|
|
xebra
|
3ecff41908
|
[spline/bezier]Fix to avoid temporarily texture animation bug with hardware tessellation.
|
2017-01-24 14:39:19 +09:00 |
|
|
Henrik Rydgard
|
f111767315
|
Workaround the lack of support for software skinning in Vulkan
|
2017-01-23 22:06:30 +01:00 |
|
|
Henrik Rydgard
|
69267d4cd6
|
Minor optimizations
|
2017-01-23 21:14:27 +01:00 |
|
|
Henrik Rydgard
|
ae37df0a8c
|
Further centralization of EstimatePerVertexCost, now moved to GPUCommon
|
2017-01-23 21:11:39 +01:00 |
|
|
Henrik Rydgard
|
daf02f1de6
|
Centralize the matrix load functions, no need to duplicate them for each backend.
|
2017-01-23 21:11:39 +01:00 |
|
|
Henrik Rydgard
|
3c221a7ebd
|
Extract the dirty flag from the shader managers so code that dirties uniform flags can be shared.
|
2017-01-23 21:11:39 +01:00 |
|
|
Henrik Rydgard
|
097638f4fd
|
Fix crash bug
|
2017-01-23 21:09:20 +01:00 |
|
|
Henrik Rydgard
|
59d80a3785
|
Share the set of dirty flags between the backend, to let use share more code.
|
2017-01-23 20:39:17 +01:00 |
|
|
Henrik Rydgard
|
63c00ec528
|
Don't need to take vertex format into account when applying UV scale for spline/bezier
|
2017-01-23 17:44:13 +01:00 |
|
|
Henrik Rydgard
|
bd4436c6eb
|
Centralize EstimatePerVertexCost and Execute_BlockTransferStart
|
2017-01-23 17:22:17 +01:00 |
|
|
Henrik Rydgard
|
d9acd27126
|
Rename GLES files to match the convention the other backends use.
|
2017-01-23 17:08:58 +01:00 |
|
|
Henrik Rydgård
|
f111eede85
|
Some GLSL syntax fixes
|
2017-01-23 14:31:44 +01:00 |
|
|
xebra
|
7530e35ab6
|
Fix 64bits shader dirty flags properly.
|
2017-01-23 14:13:28 +01:00 |
|
|
xebra
|
de253a6aef
|
[spline/bezier]Revert #ab9c1d4 about spline/bezier.
|
2017-01-23 14:12:46 +01:00 |
|
|
xebra
|
deea258383
|
[spline/bezier]Fix to avoid crashing when activated hardware tessellation with dx9 and vulkan.
|
2017-01-23 14:12:03 +01:00 |
|
|
xebra
|
334afbd1ae
|
[spline/bezier]Fix again about compatibility with GLES. from @LunaMoo.
|
2017-01-23 14:12:03 +01:00 |
|
|
xebra
|
67503ed160
|
[spline/bezier]Small fixes for compatibility with GLES 2.0.
|
2017-01-23 14:12:03 +01:00 |
|
|
xebra
|
5aeaa95f2c
|
[spline/bezier]Implement hardware tessellation on OpenGL.
|
2017-01-23 14:12:02 +01:00 |
|
|
xebra
|
38293adc78
|
[spline/bezier]Expand shader dirty flags to 64 bit, and add some flags for hardware tessellation.
|
2017-01-23 14:12:02 +01:00 |
|
|
Henrik Rydgard
|
ab9c1d4dc0
|
Centralize the spline/bezier/bbox execute functions into GPUCommon
|
2017-01-21 20:43:05 +01:00 |
|
|
Henrik Rydgård
|
5af8b152e1
|
Further buildfixes
|
2016-12-27 11:59:12 +01:00 |
|
|
Henrik Rydgård
|
674937a520
|
Extend some enums with potentially useful values
|
2016-12-27 11:59:12 +01:00 |
|
|
Henrik Rydgard
|
dc5c0c0d2d
|
Centralize ProcessEvent
|
2016-12-21 20:02:05 +01:00 |
|
|
Henrik Rydgard
|
7634ff7f09
|
Centralize a few more functions, fix DX9
|
2016-12-21 18:51:19 +01:00 |
|
|
Henrik Rydgard
|
98ebf9ed6f
|
Centralize PerformStencil* and Invalidate* functions
|
2016-12-21 18:33:08 +01:00 |
|
|
Henrik Rydgard
|
866d4a645f
|
Centralize "PerformMemory*" functions
|
2016-12-21 18:26:06 +01:00 |
|
|
Henrik Rydgard
|
e15cba0e1b
|
Centralize "Resized()"
|
2016-12-21 18:13:58 +01:00 |
|
|
Henrik Rydgard
|
7e46a153c0
|
Centralize DoBlockTransfer between the GPUs.
Required adding a pointer to the framebuffer manager and texture cache in
GPUCommon, which took a little bit of refactoring.
|
2016-12-21 18:07:17 +01:00 |
|
|
Henrik Rydgard
|
0fa2f2c7ae
|
Remove more now-unused code
|
2016-12-20 13:50:07 +01:00 |
|
|
Henrik Rydgard
|
1dbeca0618
|
Only support decoding UV to float in vertex formats. Supporting u8 and 16 was more complexity than it's worth.
|
2016-12-20 13:42:54 +01:00 |
|
|
Henrik Rydgard
|
9d7983eee8
|
Remove the ability to turn off prescaleUV entirely. It's simply better to do it in the vertex decoder.
|
2016-12-20 13:27:44 +01:00 |
|