Merge pull request #2401 from KhronosGroup/pr-2400

Land PR 2400.
KhronosGroup · Oct 30, 2024 · 2f08260 · 2f08260
2 parents f93223f + 492fa9c
commit 2f08260
Show file tree

Hide file tree

Showing 9 changed files with 1,478 additions and 14 deletions.
diff --git a/reference/opt/shaders-msl/mesh/mesh-shader-basic-lines.msl3.spv14.vk.nocompat.mesh b/reference/opt/shaders-msl/mesh/mesh-shader-basic-lines.msl3.spv14.vk.nocompat.mesh
@@ -0,0 +1,193 @@
+#pragma clang diagnostic ignored "-Wmissing-prototypes"
+#pragma clang diagnostic ignored "-Wmissing-braces"
+
+#include <metal_stdlib>
+#include <simd/simd.h>
+
+using namespace metal;
+
+template<typename T, size_t Num>
+struct spvUnsafeArray
+{
+    T elements[Num ? Num : 1];
+
+    thread T& operator [] (size_t pos) thread
+    {
+        return elements[pos];
+    }
+    constexpr const thread T& operator [] (size_t pos) const thread
+    {
+        return elements[pos];
+    }
+
+    device T& operator [] (size_t pos) device
+    {
+        return elements[pos];
+    }
+    constexpr const device T& operator [] (size_t pos) const device
+    {
+        return elements[pos];
+    }
+
+    constexpr const constant T& operator [] (size_t pos) const constant
+    {
+        return elements[pos];
+    }
+
+    threadgroup T& operator [] (size_t pos) threadgroup
+    {
+        return elements[pos];
+    }
+    constexpr const threadgroup T& operator [] (size_t pos) const threadgroup
+    {
+        return elements[pos];
+    }
+
+    object_data T& operator [] (size_t pos) object_data
+    {
+        return elements[pos];
+    }
+    constexpr const object_data T& operator [] (size_t pos) const object_data
+    {
+        return elements[pos];
+    }
+};
+
+void spvSetMeshOutputsEXT(uint gl_LocalInvocationIndex, threadgroup uint2& spvMeshSizes, uint vertexCount, uint primitiveCount)
+{
+    if (gl_LocalInvocationIndex == 0)
+    {
+        spvMeshSizes.x = vertexCount;
+        spvMeshSizes.y = primitiveCount;
+    }
+}
+
+struct gl_MeshPerPrimitiveEXT
+{
+    uint gl_PrimitiveID [[primitive_id]];
+    uint gl_Layer [[render_target_array_index]];
+    uint gl_ViewportIndex [[viewport_array_index]];
+    bool gl_CullPrimitiveEXT [[primitive_culled]];
+};
+
+struct gl_MeshPerVertexEXT
+{
+    float4 gl_Position [[position]];
+    float gl_PointSize;
+    float gl_ClipDistance [[clip_distance]] [1];
+};
+
+struct BlockOut
+{
+    float4 a;
+    float4 b;
+};
+
+struct BlockOutPrim
+{
+    float4 a;
+    float4 b;
+};
+
+constant uint3 gl_WorkGroupSize [[maybe_unused]] = uint3(2u, 3u, 4u);
+
+struct spvPerVertex
+{
+    float4 gl_Position [[position]];
+    float gl_PointSize;
+    float gl_ClipDistance [[clip_distance]] [1];
+    float gl_ClipDistance_0 [[user(clip0)]];
+    float4 vOut [[user(locn0)]];
+    float4 outputs_a [[user(locn2)]];
+    float4 outputs_b [[user(locn3)]];
+};
+
+struct spvPerPrimitive
+{
+    uint gl_PrimitiveID [[primitive_id]];
+    uint gl_Layer [[render_target_array_index]];
+    uint gl_ViewportIndex [[viewport_array_index]];
+    bool gl_CullPrimitiveEXT [[primitive_culled]];
+    float4 vPrim [[user(locn1)]];
+    float4 prim_outputs_a [[user(locn4)]];
+    float4 prim_outputs_b [[user(locn5)]];
+};
+
+using spvMesh_t = mesh<spvPerVertex, spvPerPrimitive, 24, 22, topology::line>;
+
+static inline __attribute__((always_inline))
+void _4(threadgroup spvUnsafeArray<uint2, 22>& gl_PrimitiveLineIndicesEXT, thread uint& gl_LocalInvocationIndex, threadgroup spvUnsafeArray<gl_MeshPerPrimitiveEXT, 22>& gl_MeshPrimitivesEXT, thread uint3& gl_GlobalInvocationID, threadgroup spvUnsafeArray<gl_MeshPerVertexEXT, 24>& gl_MeshVerticesEXT, threadgroup spvUnsafeArray<float4, 24>& vOut, threadgroup spvUnsafeArray<BlockOut, 24>& outputs, threadgroup spvUnsafeArray<float4, 22>& vPrim, thread uint3& gl_WorkGroupID, threadgroup spvUnsafeArray<BlockOutPrim, 22>& prim_outputs, threadgroup uint2& spvMeshSizes)
+{
+    spvSetMeshOutputsEXT(gl_LocalInvocationIndex, spvMeshSizes, 24u, 22u);
+    float3 _158 = float3(gl_GlobalInvocationID);
+    float _159 = _158.x;
+    gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_Position = float4(_159, _158.yz, 1.0);
+    gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_PointSize = 2.0;
+    gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_ClipDistance[0] = 4.0;
+    vOut[gl_LocalInvocationIndex] = float4(_159, _158.yz, 2.0);
+    outputs[gl_LocalInvocationIndex].a = float4(5.0);
+    outputs[gl_LocalInvocationIndex].b = float4(6.0);
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    if (gl_LocalInvocationIndex < 22u)
+    {
+        vPrim[gl_LocalInvocationIndex] = float4(float3(gl_WorkGroupID), 3.0);
+        prim_outputs[gl_LocalInvocationIndex].a = float4(0.0);
+        prim_outputs[gl_LocalInvocationIndex].b = float4(1.0);
+        gl_PrimitiveLineIndicesEXT[gl_LocalInvocationIndex] = uint2(0u, 1u) + uint2(gl_LocalInvocationIndex);
+        int _206 = int(gl_GlobalInvocationID.x);
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_PrimitiveID = _206;
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_Layer = _206 + 1;
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_ViewportIndex = _206 + 2;
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_CullPrimitiveEXT = short((gl_GlobalInvocationID.x & 1u) != 0u);
+    }
+}
+
+[[mesh]] void main0(uint gl_LocalInvocationIndex [[thread_index_in_threadgroup]], uint3 gl_GlobalInvocationID [[thread_position_in_grid]], uint3 gl_WorkGroupID [[threadgroup_position_in_grid]], spvMesh_t spvMesh)
+{
+    threadgroup uint2 spvMeshSizes;
+    threadgroup spvUnsafeArray<uint2, 22> gl_PrimitiveLineIndicesEXT;
+    threadgroup spvUnsafeArray<gl_MeshPerPrimitiveEXT, 22> gl_MeshPrimitivesEXT;
+    threadgroup spvUnsafeArray<gl_MeshPerVertexEXT, 24> gl_MeshVerticesEXT;
+    threadgroup spvUnsafeArray<float4, 24> vOut;
+    threadgroup spvUnsafeArray<BlockOut, 24> outputs;
+    threadgroup spvUnsafeArray<float4, 22> vPrim;
+    threadgroup spvUnsafeArray<BlockOutPrim, 22> prim_outputs;
+    threadgroup spvUnsafeArray<float, 16> shared_float;
+    if (gl_LocalInvocationIndex == 0) spvMeshSizes.y = 0u;
+    _4(gl_PrimitiveLineIndicesEXT, gl_LocalInvocationIndex, gl_MeshPrimitivesEXT, gl_GlobalInvocationID, gl_MeshVerticesEXT, vOut, outputs, vPrim, gl_WorkGroupID, prim_outputs, spvMeshSizes);
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    if (spvMeshSizes.y == 0)
+    {
+        return;
+    }
+    spvMesh.set_primitive_count(spvMeshSizes.y);
+    const uint spvThreadCount = (gl_WorkGroupSize.x * gl_WorkGroupSize.y * gl_WorkGroupSize.z);
+    const uint spvVI = gl_LocalInvocationIndex;
+    if (gl_LocalInvocationIndex < spvMeshSizes.x)
+    {
+        spvPerVertex spvV = {};
+        spvV.gl_Position = gl_MeshVerticesEXT[spvVI].gl_Position;
+        spvV.gl_PointSize = gl_MeshVerticesEXT[spvVI].gl_PointSize;
+        spvV.gl_ClipDistance[0] = gl_MeshVerticesEXT[spvVI].gl_ClipDistance[0];
+        spvV.gl_ClipDistance_0 = gl_MeshVerticesEXT[spvVI].gl_ClipDistance[0];
+        spvV.outputs_a = outputs[spvVI].a;
+        spvV.outputs_b = outputs[spvVI].b;
+        spvMesh.set_vertex(spvVI, spvV);
+    }
+    const uint spvPI = gl_LocalInvocationIndex;
+    if (gl_LocalInvocationIndex < spvMeshSizes.y)
+    {
+        spvMesh.set_index(spvPI * 2u + 0u, gl_PrimitiveLineIndicesEXT[spvPI].x);
+        spvMesh.set_index(spvPI * 2u + 1u, gl_PrimitiveLineIndicesEXT[spvPI].y);
+        spvPerPrimitive spvP = {};
+        spvP.gl_PrimitiveID = gl_MeshPrimitivesEXT[spvPI].gl_PrimitiveID;
+        spvP.gl_Layer = gl_MeshPrimitivesEXT[spvPI].gl_Layer;
+        spvP.gl_ViewportIndex = gl_MeshPrimitivesEXT[spvPI].gl_ViewportIndex;
+        spvP.gl_CullPrimitiveEXT = gl_MeshPrimitivesEXT[spvPI].gl_CullPrimitiveEXT;
+        spvP.vPrim = vPrim[spvPI];
+        spvP.prim_outputs_a = prim_outputs[spvPI].a;
+        spvP.prim_outputs_b = prim_outputs[spvPI].b;
+        spvMesh.set_primitive(spvPI, spvP);
+    }
+}
+
diff --git a/reference/opt/shaders-msl/mesh/mesh-shader-basic-triangle.msl3.spv14.vk.nocompat.mesh b/reference/opt/shaders-msl/mesh/mesh-shader-basic-triangle.msl3.spv14.vk.nocompat.mesh
@@ -0,0 +1,201 @@
+#pragma clang diagnostic ignored "-Wmissing-prototypes"
+#pragma clang diagnostic ignored "-Wmissing-braces"
+
+#include <metal_stdlib>
+#include <simd/simd.h>
+
+using namespace metal;
+
+template<typename T, size_t Num>
+struct spvUnsafeArray
+{
+    T elements[Num ? Num : 1];
+
+    thread T& operator [] (size_t pos) thread
+    {
+        return elements[pos];
+    }
+    constexpr const thread T& operator [] (size_t pos) const thread
+    {
+        return elements[pos];
+    }
+
+    device T& operator [] (size_t pos) device
+    {
+        return elements[pos];
+    }
+    constexpr const device T& operator [] (size_t pos) const device
+    {
+        return elements[pos];
+    }
+
+    constexpr const constant T& operator [] (size_t pos) const constant
+    {
+        return elements[pos];
+    }
+
+    threadgroup T& operator [] (size_t pos) threadgroup
+    {
+        return elements[pos];
+    }
+    constexpr const threadgroup T& operator [] (size_t pos) const threadgroup
+    {
+        return elements[pos];
+    }
+
+    object_data T& operator [] (size_t pos) object_data
+    {
+        return elements[pos];
+    }
+    constexpr const object_data T& operator [] (size_t pos) const object_data
+    {
+        return elements[pos];
+    }
+};
+
+void spvSetMeshOutputsEXT(uint gl_LocalInvocationIndex, threadgroup uint2& spvMeshSizes, uint vertexCount, uint primitiveCount)
+{
+    if (gl_LocalInvocationIndex == 0)
+    {
+        spvMeshSizes.x = vertexCount;
+        spvMeshSizes.y = primitiveCount;
+    }
+}
+
+struct gl_MeshPerVertexEXT
+{
+    float4 gl_Position [[position]];
+    float gl_PointSize;
+    float gl_ClipDistance [[clip_distance]] [1];
+};
+
+struct BlockOut
+{
+    float4 a;
+    float4 b;
+};
+
+struct BlockOutPrim
+{
+    float4 a;
+    float4 b;
+};
+
+struct gl_MeshPerPrimitiveEXT
+{
+    uint gl_PrimitiveID [[primitive_id]];
+    uint gl_Layer [[render_target_array_index]];
+    uint gl_ViewportIndex [[viewport_array_index]];
+    bool gl_CullPrimitiveEXT [[primitive_culled]];
+};
+
+constant uint3 gl_WorkGroupSize [[maybe_unused]] = uint3(2u, 3u, 4u);
+
+struct TaskPayload
+{
+    float a;
+    float b;
+    int c;
+};
+
+struct spvPerVertex
+{
+    float4 gl_Position [[position]];
+    float gl_PointSize;
+    float gl_ClipDistance [[clip_distance]] [1];
+    float gl_ClipDistance_0 [[user(clip0)]];
+    float4 vOut [[user(locn0)]];
+    float4 outputs_a [[user(locn2)]];
+    float4 outputs_b [[user(locn3)]];
+};
+
+struct spvPerPrimitive
+{
+    float4 vPrim [[user(locn1)]];
+    float4 prim_outputs_a [[user(locn4)]];
+    float4 prim_outputs_b [[user(locn5)]];
+    uint gl_PrimitiveID [[primitive_id]];
+    uint gl_Layer [[render_target_array_index]];
+    uint gl_ViewportIndex [[viewport_array_index]];
+    bool gl_CullPrimitiveEXT [[primitive_culled]];
+};
+
+using spvMesh_t = mesh<spvPerVertex, spvPerPrimitive, 24, 22, topology::triangle>;
+
+static inline __attribute__((always_inline))
+void _4(threadgroup spvUnsafeArray<gl_MeshPerVertexEXT, 24>& gl_MeshVerticesEXT, thread uint& gl_LocalInvocationIndex, thread uint3& gl_GlobalInvocationID, threadgroup spvUnsafeArray<float4, 24>& vOut, threadgroup spvUnsafeArray<BlockOut, 24>& outputs, threadgroup spvUnsafeArray<float4, 22>& vPrim, thread uint3& gl_WorkGroupID, threadgroup spvUnsafeArray<BlockOutPrim, 22>& prim_outputs, threadgroup spvUnsafeArray<uint3, 22>& gl_PrimitiveTriangleIndicesEXT, threadgroup spvUnsafeArray<gl_MeshPerPrimitiveEXT, 22>& gl_MeshPrimitivesEXT, threadgroup uint2& spvMeshSizes)
+{
+    spvSetMeshOutputsEXT(gl_LocalInvocationIndex, spvMeshSizes, 24u, 22u);
+    float3 _27 = float3(gl_GlobalInvocationID);
+    float _29 = _27.x;
+    gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_Position = float4(_29, _27.yz, 1.0);
+    gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_PointSize = 2.0;
+    gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_ClipDistance[0] = 4.0;
+    vOut[gl_LocalInvocationIndex] = float4(_29, _27.yz, 2.0);
+    outputs[gl_LocalInvocationIndex].a = float4(5.0);
+    outputs[gl_LocalInvocationIndex].b = float4(6.0);
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    if (gl_LocalInvocationIndex < 22u)
+    {
+        vPrim[gl_LocalInvocationIndex] = float4(float3(gl_WorkGroupID), 3.0);
+        prim_outputs[gl_LocalInvocationIndex].a = float4(0.0);
+        prim_outputs[gl_LocalInvocationIndex].b = float4(1.0);
+        gl_PrimitiveTriangleIndicesEXT[gl_LocalInvocationIndex] = uint3(0u, 1u, 2u) + uint3(gl_LocalInvocationIndex);
+        int _116 = int(gl_GlobalInvocationID.x);
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_PrimitiveID = _116;
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_Layer = _116 + 1;
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_ViewportIndex = _116 + 2;
+        gl_MeshPrimitivesEXT[gl_LocalInvocationIndex].gl_CullPrimitiveEXT = short((gl_GlobalInvocationID.x & 1u) != 0u);
+    }
+}
+
+[[mesh]] void main0(uint gl_LocalInvocationIndex [[thread_index_in_threadgroup]], uint3 gl_GlobalInvocationID [[thread_position_in_grid]], uint3 gl_WorkGroupID [[threadgroup_position_in_grid]], spvMesh_t spvMesh)
+{
+    threadgroup uint2 spvMeshSizes;
+    threadgroup spvUnsafeArray<gl_MeshPerVertexEXT, 24> gl_MeshVerticesEXT;
+    threadgroup spvUnsafeArray<float4, 24> vOut;
+    threadgroup spvUnsafeArray<BlockOut, 24> outputs;
+    threadgroup spvUnsafeArray<float4, 22> vPrim;
+    threadgroup spvUnsafeArray<BlockOutPrim, 22> prim_outputs;
+    threadgroup spvUnsafeArray<uint3, 22> gl_PrimitiveTriangleIndicesEXT;
+    threadgroup spvUnsafeArray<gl_MeshPerPrimitiveEXT, 22> gl_MeshPrimitivesEXT;
+    threadgroup spvUnsafeArray<float, 16> shared_float;
+    if (gl_LocalInvocationIndex == 0) spvMeshSizes.y = 0u;
+    _4(gl_MeshVerticesEXT, gl_LocalInvocationIndex, gl_GlobalInvocationID, vOut, outputs, vPrim, gl_WorkGroupID, prim_outputs, gl_PrimitiveTriangleIndicesEXT, gl_MeshPrimitivesEXT, spvMeshSizes);
+    threadgroup_barrier(mem_flags::mem_threadgroup);
+    if (spvMeshSizes.y == 0)
+    {
+        return;
+    }
+    spvMesh.set_primitive_count(spvMeshSizes.y);
+    const uint spvThreadCount = (gl_WorkGroupSize.x * gl_WorkGroupSize.y * gl_WorkGroupSize.z);
+    const uint spvVI = gl_LocalInvocationIndex;
+    if (gl_LocalInvocationIndex < spvMeshSizes.x)
+    {
+        spvPerVertex spvV = {};
+        spvV.gl_Position = gl_MeshVerticesEXT[spvVI].gl_Position;
+        spvV.gl_PointSize = gl_MeshVerticesEXT[spvVI].gl_PointSize;
+        spvV.gl_ClipDistance[0] = gl_MeshVerticesEXT[spvVI].gl_ClipDistance[0];
+        spvV.gl_ClipDistance_0 = gl_MeshVerticesEXT[spvVI].gl_ClipDistance[0];
+        spvV.outputs_a = outputs[spvVI].a;
+        spvV.outputs_b = outputs[spvVI].b;
+        spvMesh.set_vertex(spvVI, spvV);
+    }
+    const uint spvPI = gl_LocalInvocationIndex;
+    if (gl_LocalInvocationIndex < spvMeshSizes.y)
+    {
+        spvMesh.set_index(spvPI * 3u + 0u, gl_PrimitiveTriangleIndicesEXT[spvPI].x);
+        spvMesh.set_index(spvPI * 3u + 1u, gl_PrimitiveTriangleIndicesEXT[spvPI].y);
+        spvMesh.set_index(spvPI * 3u + 2u, gl_PrimitiveTriangleIndicesEXT[spvPI].z);
+        spvPerPrimitive spvP = {};
+        spvP.vPrim = vPrim[spvPI];
+        spvP.prim_outputs_a = prim_outputs[spvPI].a;
+        spvP.prim_outputs_b = prim_outputs[spvPI].b;
+        spvP.gl_PrimitiveID = gl_MeshPrimitivesEXT[spvPI].gl_PrimitiveID;
+        spvP.gl_Layer = gl_MeshPrimitivesEXT[spvPI].gl_Layer;
+        spvP.gl_ViewportIndex = gl_MeshPrimitivesEXT[spvPI].gl_ViewportIndex;
+        spvP.gl_CullPrimitiveEXT = gl_MeshPrimitivesEXT[spvPI].gl_CullPrimitiveEXT;
+        spvMesh.set_primitive(spvPI, spvP);
+    }
+}
+