Shader Model 6 wave ops require instructions not available by extension in SPIR-V 1.0, and device support info is found in VkPhysicalDeviceSubgroupProperties, which is also has no equivalent by extension in Vulkan 1.0.
--
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/776
Vulkan test needs a DSV clear. Also I guess we want a configurable clear value.
--
v3: vkd3d-shader/dxil: Support SV_Depth, SV_DepthGreaterEqual and SV_DepthLessEqual.
tests/hlsl: Add an SV_Depth test.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/777
This would eliminate the todo for the precise mad() test in !718. Maybe we need test results on nvidia and intel to decide if we actually want this.
--
v8: vkd3d-shader/ir: Implement MAD in two operations if flagged as precise.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/723
Another round of optimizations for the CFG structurizer. Here we begin getting rid of many of the loops that were generated during structurization to represent forward jumps, but that are better expressed with selection constructs when possible. More optimizations will follow.
--
v2: vkd3d-shader/ir: Remove loops that terminate with a `break'.
vkd3d-shader/ir: Count how many jumps target each loop.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/764
I originally planned to somehow pass hybrid/native arch info from configure instead of hardcoding it in makedep, but I'm no longer convinced it's worth extra per-arch variables. I will change it if that would be preferred.
CC @bylaws
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/5486
Vulkan test needs a DSV clear. Also I guess we want a configurable clear value.
--
v2: vkd3d-shader/dxil: Support SV_Depth, SV_DepthGreaterEqual and SV_DepthLessEqual.
tests/hlsl: Add an SV_Depth test.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/777
--
v34: vkd3d-shader/dxil: Validate control point count.
This merge request has too many patches to be relayed via email.
Please visit the URL below to see the contents of the merge request.
https://gitlab.winehq.org/wine/vkd3d/-/merge_requests/648
__int64 has an inconsistent alignment: 4 bytes on Unix, 8 bytes on PE.
Fix this by using INT64/UINT64 which has consistent alignment on both
Unix and PE side.
This allows adding 64-bit fields at unaligned offset in the future,
without triggering struct layout mismatch between Unix and PE due to
inserted padding (which exists on PE but not on Unix).
--
https://gitlab.winehq.org/wine/wine/-/merge_requests/5483