[spirv] Add mesh/amplification shader tests for use of buffers and fn calls (#2390)
- Also cleans up SPIR-V.rst documentation issues.
This commit is contained in:
Sahil Parmar
Ehsan
Родитель
c8f7a6c970
Коммит
83c1bf8e92
|
|
@ -3104,14 +3104,14 @@ Callable Stage
|
|||
Mesh and Amplification Shaders
|
||||
------------------------------
|
||||
|
||||
DirectX adds 2 new shader stages for using MeshShading pipeline namely Mesh and Amplification.
|
||||
Amplification shaders corresponds to Task Shaders in Vulkan.
|
||||
|
||||
| DirectX adds 2 new shader stages for using MeshShading pipeline namely Mesh and Amplification.
|
||||
| Amplification shaders corresponds to Task Shaders in Vulkan.
|
||||
|
|
||||
| Refer to following HLSL and SPIR-V specs for details:
|
||||
| https://docs.microsoft.com/<TBD>
|
||||
| https://github.com/KhronosGroup/SPIRV-Registry/blob/master/extensions/NV/SPV_NV_mesh_shader.asciidoc
|
||||
|
||||
This section describes how Mesh and Amplification shaders are translated to SPIR-V for Vulkan.
|
||||
|
|
||||
| This section describes how Mesh and Amplification shaders are translated to SPIR-V for Vulkan.
|
||||
|
||||
Entry Point Attributes
|
||||
~~~~~~~~~~~~~~~~~~~~~~
|
||||
|
|
@ -3120,18 +3120,19 @@ shaders and are translated to SPIR-V execution modes according to the table belo
|
|||
|
||||
.. table:: Mapping from HLSL attribute to SPIR-V execution mode
|
||||
|
||||
+--------------------+----------------+-------------------------+
|
||||
+-------------------+--------------------+-------------------------+
|
||||
| HLSL Attribute | Value | SPIR-V Execution Mode |
|
||||
+====================+================+=========================+
|
||||
| | ``point`` | ``OutputPoints`` |
|
||||
| +----------------+-------------------------+
|
||||
| ``outputtopology`` | ``line`` | ``OutputLinesNV`` |
|
||||
| (Mesh shader) +----------------+-------------------------+
|
||||
+===================+====================+=========================+
|
||||
|``outputtopology`` | ``point`` | ``OutputPoints`` |
|
||||
| +--------------------+-------------------------+
|
||||
|``(Mesh shader)`` | ``line`` | ``OutputLinesNV`` |
|
||||
| +--------------------+-------------------------+
|
||||
| | ``triangle`` | ``OutputTrianglesNV`` |
|
||||
+--------------------+----------------+-------------------------+
|
||||
+-------------------+--------------------+-------------------------+
|
||||
| ``numthreads`` | ``X, Y, Z`` | ``LocalSize X, Y, Z`` |
|
||||
| | (X*Y*Z <= 128) | |
|
||||
+--------------------+----------------+-------------------------+
|
||||
| | | |
|
||||
| | ``(X*Y*Z <= 128)`` | |
|
||||
+-------------------+--------------------+-------------------------+
|
||||
|
||||
Intrinsics
|
||||
~~~~~~~~~~
|
||||
|
|
@ -3140,24 +3141,29 @@ and are translated to SPIR-V intrinsics according to the table below:
|
|||
|
||||
.. table:: Mapping from HLSL intrinsics to SPIR-V intrinsics
|
||||
|
||||
+-------------------------+--------------------+-----------------------------------------+
|
||||
+---------------------------+--------------------+-----------------------------------------+
|
||||
| HLSL Intrinsic | Parameters | SPIR-V Intrinsic |
|
||||
+=========================+====================+=========================================+
|
||||
+===========================+====================+=========================================+
|
||||
| ``SetMeshOutputCounts`` | ``numVertices`` | ``PrimitiveCountNV numPrimitives`` |
|
||||
| (Mesh shader) | ``numPrimitives`` | |
|
||||
+-------------------------+--------------------+-----------------------------------------+
|
||||
| | ``ThreadX`` | |
|
||||
| ``DispatchMesh`` | ``ThreadY`` | ``OpControlBarrier`` |
|
||||
| (Amplification shader) | ``ThreadZ`` | ``TaskCountNV ThreadX*ThreadY*ThreadZ`` |
|
||||
| | | |
|
||||
| ``(Mesh shader)`` | ``numPrimitives`` | |
|
||||
+---------------------------+--------------------+-----------------------------------------+
|
||||
| ``DispatchMesh`` | ``ThreadX`` | ``OpControlBarrier`` |
|
||||
| | | |
|
||||
| ``(Amplification shader)``| ``ThreadY`` | ``TaskCountNV ThreadX*ThreadY*ThreadZ`` |
|
||||
| | | |
|
||||
| | ``ThreadZ`` | |
|
||||
| | | |
|
||||
| | ``MeshPayload`` | |
|
||||
+-------------------------+--------------------+-----------------------------------------+
|
||||
+---------------------------+--------------------+-----------------------------------------+
|
||||
|
||||
| *For DispatchMesh intrinsic, we also emit MeshPayload as output block with PerTaskNV decoration
|
||||
| Note : For ``DispatchMesh`` intrinsic, we also emit ``MeshPayload`` as output block with ``PerTaskNV`` decoration
|
||||
|
||||
Mesh Interface Variables
|
||||
~~~~~~~~~~~~~~~~~~~~~~~~
|
||||
Interface variables are defined for Mesh shaders using HLSL modifiers.
|
||||
Following table gives high level overview of the mapping:
|
||||
| Interface variables are defined for Mesh shaders using HLSL modifiers.
|
||||
| Following table gives high level overview of the mapping:
|
||||
|
|
||||
|
||||
.. table:: Mapping from HLSL modifiers to SPIR-V definitions
|
||||
|
||||
|
|
@ -3165,9 +3171,11 @@ Following table gives high level overview of the mapping:
|
|||
| HLSL modifier | SPIR-V definition |
|
||||
+=================+=========================================================================+
|
||||
| ``indices`` | Maps to SPIR-V intrinsic ``PrimitiveIndicesNV`` |
|
||||
| | |
|
||||
| | Defines SPIR-V Execution Mode ``OutputPrimitivesNV <array-size>`` |
|
||||
+-----------------+-------------------------------------------------------------------------+
|
||||
| ``vertices`` | Maps to per-vertex out attributes |
|
||||
| | |
|
||||
| | Defines existing SPIR-V Execution Mode ``OutputVertices <array-size>`` |
|
||||
+-----------------+-------------------------------------------------------------------------+
|
||||
| ``primitives`` | Maps to per-primitive out attributes with ``PerPrimitiveNV`` decoration |
|
||||
|
|
|
|||
|
|
@ -0,0 +1,202 @@
|
|||
// Run: %dxc -T ms_6_5 -E main
|
||||
|
||||
// CHECK: OpCapability MeshShadingNV
|
||||
// CHECK: OpExtension "SPV_NV_mesh_shader"
|
||||
// CHECK: OpEntryPoint MeshNV %main "main"
|
||||
|
||||
// CHECK: OpName %UserVertex "UserVertex"
|
||||
struct UserVertex {
|
||||
// CHECK: OpMemberName %UserVertex 0 "position"
|
||||
// CHECK: OpMemberName %UserVertex 1 "texcoord"
|
||||
// CHECK: OpMemberName %UserVertex 2 "color"
|
||||
float3 position;
|
||||
float2 texcoord;
|
||||
float3 color;
|
||||
};
|
||||
|
||||
// CHECK: OpName %Mesh "Mesh"
|
||||
struct Mesh {
|
||||
// CHECK: OpMemberName %Mesh 0 "firstSubmesh"
|
||||
// CHECK: OpMemberName %Mesh 1 "submeshCount"
|
||||
// CHECK: OpMemberName %Mesh 2 "dummy"
|
||||
uint firstSubmesh;
|
||||
uint submeshCount;
|
||||
uint dummy[2];
|
||||
};
|
||||
|
||||
// CHECK: OpName %SubMesh "SubMesh"
|
||||
struct SubMesh {
|
||||
// CHECK: OpMemberName %SubMesh 0 "vertexCount"
|
||||
// CHECK: OpMemberName %SubMesh 1 "vertexOffset"
|
||||
// CHECK: OpMemberName %SubMesh 2 "primitiveCount"
|
||||
// CHECK: OpMemberName %SubMesh 3 "indexOffset"
|
||||
// CHECK: OpMemberName %SubMesh 4 "boundingBox"
|
||||
uint vertexCount;
|
||||
uint vertexOffset;
|
||||
uint primitiveCount;
|
||||
uint indexOffset;
|
||||
float4 boundingBox[8];
|
||||
};
|
||||
|
||||
// CHECK: OpDecorate %userVertices DescriptorSet 0
|
||||
// CHECK: OpDecorate %userVertices Binding 0
|
||||
// CHECK: OpDecorate %userIndices DescriptorSet 0
|
||||
// CHECK: OpDecorate %userIndices Binding 1
|
||||
// CHECK: OpDecorate %meshes DescriptorSet 0
|
||||
// CHECK: OpDecorate %meshes Binding 2
|
||||
// CHECK: OpDecorate %submeshes DescriptorSet 0
|
||||
// CHECK: OpDecorate %submeshes Binding 3
|
||||
// CHECK: OpDecorate %UBO DescriptorSet 0
|
||||
// CHECK: OpDecorate %UBO Binding 4
|
||||
|
||||
// CHECK: OpMemberDecorate %UserVertex 0 Offset 0
|
||||
// CHECK: OpMemberDecorate %UserVertex 1 Offset 16
|
||||
// CHECK: OpMemberDecorate %UserVertex 2 Offset 32
|
||||
// CHECK: OpDecorate %_runtimearr_UserVertex ArrayStride 48
|
||||
// CHECK: OpMemberDecorate %type_RWStructuredBuffer_UserVertex 0 Offset 0
|
||||
// CHECK: OpDecorate %type_RWStructuredBuffer_UserVertex BufferBlock
|
||||
|
||||
// CHECK: OpDecorate %_runtimearr_uint ArrayStride 4
|
||||
// CHECK: OpMemberDecorate %type_RWStructuredBuffer_uint 0 Offset 0
|
||||
// CHECK: OpDecorate %type_RWStructuredBuffer_uint BufferBlock
|
||||
|
||||
// CHECK: OpMemberDecorate %Mesh 0 Offset 0
|
||||
// CHECK: OpMemberDecorate %Mesh 1 Offset 4
|
||||
// CHECK: OpMemberDecorate %Mesh 2 Offset 8
|
||||
// CHECK: OpDecorate %_runtimearr_Mesh ArrayStride 16
|
||||
// CHECK: OpMemberDecorate %type_RWStructuredBuffer_Mesh 0 Offset 0
|
||||
// CHECK: OpDecorate %type_RWStructuredBuffer_Mesh BufferBlock
|
||||
|
||||
// CHECK: OpMemberDecorate %SubMesh 0 Offset 0
|
||||
// CHECK: OpMemberDecorate %SubMesh 1 Offset 4
|
||||
// CHECK: OpMemberDecorate %SubMesh 2 Offset 8
|
||||
// CHECK: OpMemberDecorate %SubMesh 3 Offset 12
|
||||
// CHECK: OpMemberDecorate %SubMesh 4 Offset 16
|
||||
// CHECK: OpDecorate %_runtimearr_SubMesh ArrayStride 144
|
||||
// CHECK: OpMemberDecorate %type_RWStructuredBuffer_SubMesh 0 Offset 0
|
||||
// CHECK: OpDecorate %type_RWStructuredBuffer_SubMesh BufferBlock
|
||||
|
||||
// CHECK: OpMemberDecorate %type_UBO 0 Offset 0
|
||||
// CHECK: OpMemberDecorate %type_UBO 0 MatrixStride 16
|
||||
// CHECK: OpMemberDecorate %type_UBO 0 ColMajor
|
||||
// CHECK: OpDecorate %type_UBO Block
|
||||
|
||||
// CHECK: %UserVertex = OpTypeStruct %v3float %v2float %v3float
|
||||
// CHECK: %_runtimearr_UserVertex = OpTypeRuntimeArray %UserVertex
|
||||
// CHECK: %type_RWStructuredBuffer_UserVertex = OpTypeStruct %_runtimearr_UserVertex
|
||||
// CHECK: %_ptr_Uniform_type_RWStructuredBuffer_UserVertex = OpTypePointer Uniform %type_RWStructuredBuffer_UserVertex
|
||||
[[vk::binding(0, 0)]]
|
||||
RWStructuredBuffer<UserVertex> userVertices;
|
||||
|
||||
// CHECK: %_runtimearr_uint = OpTypeRuntimeArray %uint
|
||||
// CHECK: %type_RWStructuredBuffer_uint = OpTypeStruct %_runtimearr_uint
|
||||
// CHECK: %_ptr_Uniform_type_RWStructuredBuffer_uint = OpTypePointer Uniform %type_RWStructuredBuffer_uint
|
||||
[[vk::binding(1, 0)]]
|
||||
RWStructuredBuffer<uint> userIndices;
|
||||
|
||||
// CHECK: %_arr_uint_uint_2 = OpTypeArray %uint %uint_2
|
||||
// CHECK: %Mesh = OpTypeStruct %uint %uint %_arr_uint_uint_2
|
||||
// CHECK: %_runtimearr_Mesh = OpTypeRuntimeArray %Mesh
|
||||
// CHECK: %type_RWStructuredBuffer_Mesh = OpTypeStruct %_runtimearr_Mesh
|
||||
// CHECK: %_ptr_Uniform_type_RWStructuredBuffer_Mesh = OpTypePointer Uniform %type_RWStructuredBuffer_Mesh
|
||||
[[vk::binding(2, 0)]]
|
||||
RWStructuredBuffer<Mesh> meshes;
|
||||
|
||||
// CHECK: %uint_8 = OpConstant %uint 8
|
||||
// CHECK: %v4float = OpTypeVector %float 4
|
||||
// CHECK: %_arr_v4float_uint_8 = OpTypeArray %v4float %uint_8
|
||||
// CHECK: %SubMesh = OpTypeStruct %uint %uint %uint %uint %_arr_v4float_uint_8
|
||||
// CHECK: %_runtimearr_SubMesh = OpTypeRuntimeArray %SubMesh
|
||||
// CHECK: %type_RWStructuredBuffer_SubMesh = OpTypeStruct %_runtimearr_SubMesh
|
||||
// CHECK: %_ptr_Uniform_type_RWStructuredBuffer_SubMesh = OpTypePointer Uniform %type_RWStructuredBuffer_SubMesh
|
||||
[[vk::binding(3, 0)]]
|
||||
RWStructuredBuffer<SubMesh> submeshes;
|
||||
|
||||
// CHECK: %mat4v4float = OpTypeMatrix %v4float 4
|
||||
// CHECK: %type_UBO = OpTypeStruct %mat4v4float
|
||||
// CHECK: %_ptr_Uniform_type_UBO = OpTypePointer Uniform %type_UBO
|
||||
[[vk::binding(4, 0)]]
|
||||
cbuffer UBO {
|
||||
row_major float4x4 mvp;
|
||||
}
|
||||
|
||||
struct PerVertex {
|
||||
float4 position : SV_Position;
|
||||
float2 texcoord : TEXCOORD;
|
||||
float3 color : COLOR;
|
||||
};
|
||||
|
||||
struct PerPrimitive {
|
||||
float4 primcolor : PCOLOR;
|
||||
};
|
||||
|
||||
struct SubMeshes {
|
||||
uint submeshID[256] : SUBMESH;
|
||||
};
|
||||
|
||||
static const uint vertsPerPrim = 3U;
|
||||
|
||||
// CHECK: %userVertices = OpVariable %_ptr_Uniform_type_RWStructuredBuffer_UserVertex Uniform
|
||||
// CHECK: %userIndices = OpVariable %_ptr_Uniform_type_RWStructuredBuffer_uint Uniform
|
||||
// CHECK: %meshes = OpVariable %_ptr_Uniform_type_RWStructuredBuffer_Mesh Uniform
|
||||
// CHECK: %submeshes = OpVariable %_ptr_Uniform_type_RWStructuredBuffer_SubMesh Uniform
|
||||
// CHECK: %UBO = OpVariable %_ptr_Uniform_type_UBO Uniform
|
||||
|
||||
[outputtopology("triangle")]
|
||||
[numthreads(32, 1, 1)]
|
||||
void main(
|
||||
out indices uint3 primIndices[128],
|
||||
out vertices PerVertex verts[128],
|
||||
out primitives PerPrimitive prims[128],
|
||||
in payload SubMeshes taskmem,
|
||||
in uint gid : SV_GroupID,
|
||||
in uint tid : SV_GroupThreadID
|
||||
)
|
||||
{
|
||||
uint task = taskmem.submeshID[gid];
|
||||
// CHECK: %submesh = OpVariable %_ptr_Function_SubMesh_0 Function
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_SubMesh %submeshes %int_0 [[task:%\d+]]
|
||||
// CHECK: OpStore %submesh [[submeshVal:%\d+]]
|
||||
SubMesh submesh = submeshes[task];
|
||||
// CHECK: OpAccessChain %_ptr_Function_uint %submesh %int_0
|
||||
uint numPackedVertices = submesh.vertexCount;
|
||||
// CHECK: OpAccessChain %_ptr_Function_uint %submesh %int_2
|
||||
uint numPackedPrimitives = submesh.primitiveCount;
|
||||
|
||||
SetMeshOutputCounts(numPackedVertices, numPackedPrimitives);
|
||||
|
||||
for (uint i = 0U; i < numPackedVertices; i += 32U) {
|
||||
uint vid = i + tid;
|
||||
// CHECK: OpAccessChain %_ptr_Function_uint %submesh %int_1
|
||||
uint svid = vid + submesh.vertexOffset;
|
||||
if (vid >= numPackedVertices) continue;
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_v2float %userVertices %int_0 [[svid_1:%\d+]] %int_1
|
||||
verts[vid].texcoord = userVertices[svid].texcoord;
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_v3float %userVertices %int_0 [[svid_2:%\d+]] %int_2
|
||||
verts[vid].color = userVertices[svid].color;
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_v3float %userVertices %int_0 [[svid_0:%\d+]] %int_0
|
||||
float3 position = userVertices[svid].position;
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_mat4v4float %UBO %int_0
|
||||
verts[vid].position = mul(mvp, float4(position, 1.0));
|
||||
}
|
||||
|
||||
GroupMemoryBarrier();
|
||||
|
||||
for (uint j = 0U; j < numPackedPrimitives; j += 32U) {
|
||||
uint pid = j + tid;
|
||||
uint didxoff = vertsPerPrim * pid;
|
||||
// CHECK: OpAccessChain %_ptr_Function_uint %submesh %int_3
|
||||
uint sidxoff = submesh.indexOffset + didxoff;
|
||||
if (pid >= numPackedPrimitives) continue;
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_uint %userIndices %int_0 [[sidxoff_0:%\d+]]
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_uint %userIndices %int_0 [[sidxoff_1:%\d+]]
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_uint %userIndices %int_0 [[sidxoff_2:%\d+]]
|
||||
primIndices[pid] = uint3(userIndices[sidxoff], userIndices[sidxoff+1], userIndices[sidxoff+2]);
|
||||
// CHECK: OpAccessChain %_ptr_Function_uint %submesh %int_1
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_uint %userIndices %int_0 [[ind:%\d+]]
|
||||
uint providx = submesh.vertexOffset + userIndices[sidxoff + vertsPerPrim - 1U];
|
||||
// CHECK: OpAccessChain %_ptr_Uniform_v3float %userVertices %int_0 [[providx:%\d+]] %int_2
|
||||
prims[pid].primcolor = float4(userVertices[providx].color, 1.0);
|
||||
}
|
||||
}
|
||||
|
||||
|
|
@ -0,0 +1,135 @@
|
|||
// Run: %dxc -T as_6_5 -E main -fspv-target-env=vulkan1.1
|
||||
|
||||
// CHECK: OpCapability MeshShadingNV
|
||||
// CHECK: OpExtension "SPV_NV_mesh_shader"
|
||||
// CHECK: OpEntryPoint TaskNV %main "main"
|
||||
|
||||
struct SubMesh {
|
||||
uint vertexCount;
|
||||
uint vertexOffset;
|
||||
uint primitiveCount;
|
||||
uint indexOffset;
|
||||
float4 boundingBox[8];
|
||||
};
|
||||
|
||||
struct Mesh {
|
||||
uint firstSubmesh;
|
||||
uint submeshCount;
|
||||
uint dummy[2];
|
||||
};
|
||||
|
||||
struct UserVertex {
|
||||
float3 position;
|
||||
float2 texcoord;
|
||||
float3 color;
|
||||
};
|
||||
|
||||
[[vk::binding(0, 0)]]
|
||||
RWStructuredBuffer<UserVertex> userVertices;
|
||||
|
||||
[[vk::binding(1, 0)]]
|
||||
RWStructuredBuffer<uint> userIndices;
|
||||
|
||||
[[vk::binding(2, 0)]]
|
||||
RWStructuredBuffer<Mesh> meshes;
|
||||
|
||||
[[vk::binding(3, 0)]]
|
||||
RWStructuredBuffer<SubMesh> submeshes;
|
||||
|
||||
[[vk::binding(4, 0)]]
|
||||
cbuffer UBO {
|
||||
row_major float4x4 mvp;
|
||||
}
|
||||
|
||||
groupshared uint passedSubmeshes;
|
||||
struct SubMeshes {
|
||||
uint submeshID[256] : SUBMESH;
|
||||
};
|
||||
groupshared SubMeshes sharedSubMeshes;
|
||||
|
||||
// CHECK: %_arr_v4float_uint_8_0 = OpTypeArray %v4float %uint_8
|
||||
// CHECK: %SubMesh_0 = OpTypeStruct %uint %uint %uint %uint %_arr_v4float_uint_8_0
|
||||
// CHECK: %_ptr_Function_SubMesh_0 = OpTypePointer Function %SubMesh_0
|
||||
// CHECK: [[funcType:%\d+]] = OpTypeFunction %bool %_ptr_Function_SubMesh_0
|
||||
|
||||
bool TestSubmesh(SubMesh submesh) {
|
||||
uint clip = 0x0U;
|
||||
|
||||
for (uint bbv = 0U ; bbv < 8U; bbv++) {
|
||||
float4 pos= mul(mvp, submesh.boundingBox[bbv]);
|
||||
if (pos.x <= pos.w) clip |= 0x1U;
|
||||
if (pos.y <= 0.3333 * pos.w) clip |= 0x2U;
|
||||
if (pos.z <= pos.w) clip |= 0x4U;
|
||||
if (pos.x >= -pos.w) clip |= 0x8U;
|
||||
if (pos.y >= -pos.w) clip |= 0x10U;
|
||||
if (pos.z >= -pos.w) clip |= 0x20U;
|
||||
}
|
||||
return (clip == 0x3FU);
|
||||
}
|
||||
|
||||
[numthreads(32, 1, 1)]
|
||||
void main(
|
||||
in uint tid : SV_GroupThreadID,
|
||||
in uint mid : SV_GroupID
|
||||
)
|
||||
{
|
||||
uint firstSubmesh = meshes[mid].firstSubmesh;
|
||||
uint submeshCount = meshes[mid].submeshCount;
|
||||
passedSubmeshes = 0U;
|
||||
GroupMemoryBarrier();
|
||||
for (uint i = 0U; i < submeshCount; i += 32U) {
|
||||
uint smid = firstSubmesh + i + tid;
|
||||
if (smid >= firstSubmesh + submeshCount) continue;
|
||||
|
||||
// CHECK: %submesh = OpVariable %_ptr_Function_SubMesh_0 Function
|
||||
// CHECK: %passed = OpVariable %_ptr_Function_bool Function
|
||||
// CHECK: %param_var_submesh = OpVariable %_ptr_Function_SubMesh_0 Function
|
||||
SubMesh submesh = submeshes[smid];
|
||||
bool passed = true;
|
||||
|
||||
// CHECK: [[submeshValue:%\d+]] = OpLoad %SubMesh_0 %submesh
|
||||
// CHECK: OpStore %param_var_submesh [[submeshValue]]
|
||||
// CHECK: [[rv:%\d+]] = OpFunctionCall %bool %TestSubmesh %param_var_submesh
|
||||
// CHECK: [[cond:%\d+]] = OpLogicalNot %bool [[rv]]
|
||||
// CHECK: OpSelectionMerge %if_merge_0 None
|
||||
// CHECK: OpBranchConditional [[cond]] %if_true_0 %if_merge_0
|
||||
// CHECK: %if_true_0 = OpLabel
|
||||
// CHECK: OpStore %passed %false
|
||||
// CHECK: OpBranch %if_merge_0
|
||||
// CHECK: %if_merge_0 = OpLabel
|
||||
if (!TestSubmesh(submesh)) passed = false;
|
||||
|
||||
if (passed) {
|
||||
uint ballot = WaveActiveBallot(passed).x;
|
||||
uint laneMaskLT = (1 << WaveGetLaneIndex()) - 1;
|
||||
uint lowerThreads = ballot & laneMaskLT;
|
||||
uint slot = passedSubmeshes + WavePrefixCountBits(passed);
|
||||
sharedSubMeshes.submeshID[slot] = smid;
|
||||
if (lowerThreads == 0U) {
|
||||
passedSubmeshes += WaveActiveCountBits(passed);
|
||||
}
|
||||
}
|
||||
GroupMemoryBarrier();
|
||||
}
|
||||
DispatchMesh(passedSubmeshes, 1, 1, sharedSubMeshes);
|
||||
}
|
||||
|
||||
/* bool TestSubmesh(SubMesh submesh) { ... } */
|
||||
|
||||
// CHECK: %TestSubmesh = OpFunction %bool None [[funcType]]
|
||||
// CHECK: %submesh_0 = OpFunctionParameter %_ptr_Function_SubMesh_0
|
||||
|
||||
// CHECK: %bb_entry_0 = OpLabel
|
||||
|
||||
// CHECK: %clip = OpVariable %_ptr_Function_uint Function
|
||||
// CHECK: %bbv = OpVariable %_ptr_Function_uint Function
|
||||
// CHECK: %pos = OpVariable %_ptr_Function_v4float Function
|
||||
|
||||
// CHECK: %for_check_0 = OpLabel
|
||||
// CHECK: %for_body_0 = OpLabel
|
||||
// CHECK: %for_merge_0 = OpLabel
|
||||
|
||||
// CHECK: [[clipValue:%\d+]] = OpLoad %uint %clip
|
||||
// CHECK: [[retValue:%\d+]] = OpIEqual %bool [[clipValue]] %uint_63
|
||||
// CHECK: OpReturnValue [[retValue]]
|
||||
// CHECK: OpFunctionEnd
|
||||
|
|
@ -2044,6 +2044,9 @@ TEST_F(FileTest, MeshShadingNVMeshLine) {
|
|||
TEST_F(FileTest, MeshShadingNVMeshPoint) {
|
||||
runFileTest("meshshading.nv.point.mesh.hlsl");
|
||||
}
|
||||
TEST_F(FileTest, MeshShadingNVMeshBuffer) {
|
||||
runFileTest("meshshading.nv.buffer.mesh.hlsl");
|
||||
}
|
||||
TEST_F(FileTest, MeshShadingNVMeshError1) {
|
||||
runFileTest("meshshading.nv.error1.mesh.hlsl", Expect::Failure);
|
||||
}
|
||||
|
|
@ -2089,6 +2092,10 @@ TEST_F(FileTest, MeshShadingNVMeshError14) {
|
|||
TEST_F(FileTest, MeshShadingNVAmplification) {
|
||||
runFileTest("meshshading.nv.amplification.hlsl");
|
||||
}
|
||||
TEST_F(FileTest, MeshShadingNVAmplificationFunCall) {
|
||||
useVulkan1p1();
|
||||
runFileTest("meshshading.nv.fncall.amplification.hlsl");
|
||||
}
|
||||
TEST_F(FileTest, MeshShadingNVAmplificationError1) {
|
||||
runFileTest("meshshading.nv.error1.amplification.hlsl", Expect::Failure);
|
||||
}
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче