Overhaul global inst deduplication and cpp/cuda backend. (#2654)

* Overhaul global inst deduplication and cpp/cuda backend.

* Update IR documentation.

---------

Co-authored-by: Yong He <yhe@nvidia.com>

4 files changed, 22 insertions, 13 deletions

diff --git a/tests/pipeline/rasterization/mesh/component-write.slang.glsl b/tests/pipeline/rasterization/mesh/component-write.slang.glsl
index 5234fb62f..143947504 100644
--- a/tests/pipeline/rasterization/mesh/component-write.slang.glsl
+++ b/tests/pipeline/rasterization/mesh/component-write.slang.glsl
@@ -2,26 +2,30 @@
 #extension GL_EXT_mesh_shader : require
 layout(row_major) uniform;
 layout(row_major) buffer;
-const vec3  colors_0[3] = { vec3(1.00000000000000000000, 1.00000000000000000000, 0.00000000000000000000), vec3(0.00000000000000000000, 1.00000000000000000000, 1.00000000000000000000), vec3(1.00000000000000000000, 0.00000000000000000000, 1.00000000000000000000) };
-const vec2  positions_0[3] = { vec2(0.00000000000000000000, -0.50000000000000000000), vec2(0.50000000000000000000, 0.50000000000000000000), vec2(-0.50000000000000000000, 0.50000000000000000000) };
+const vec2  positions_0[3] = { vec2(0.0, -0.5), vec2(0.5, 0.5), vec2(-0.5, 0.5) };
+const vec3  colors_0[3] = { vec3(1.0, 1.0, 0.0), vec3(0.0, 1.0, 1.0), vec3(1.0, 0.0, 1.0) };
 layout(location = 0)
 out vec3  _S1[3];
 
+
 out gl_MeshPerVertexEXT
 {
     vec4 gl_Position;
 } gl_MeshVerticesEXT[3];
 
+out uvec3  gl_PrimitiveTriangleIndicesEXT[1];
+
 layout(local_size_x = 3, local_size_y = 1, local_size_z = 1) in;
 layout(max_vertices = 3) out;
 layout(max_primitives = 1) out;
 layout(triangles) out;
 void main()
 {
+
     SetMeshOutputsEXT(3U, 1U);
     if(gl_LocalInvocationIndex < 3U)
     {
-        gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_Position = vec4(positions_0[gl_LocalInvocationIndex], 0.00000000000000000000, 1.00000000000000000000);
+        gl_MeshVerticesEXT[gl_LocalInvocationIndex].gl_Position = vec4(positions_0[gl_LocalInvocationIndex], 0.0, 1.0);
         _S1[gl_LocalInvocationIndex] = colors_0[gl_LocalInvocationIndex];
     }
     else
@@ -34,6 +38,6 @@ void main()
     else
     {
     }
+
     return;
 }
-
diff --git a/tests/pipeline/rasterization/mesh/hello.slang.glsl b/tests/pipeline/rasterization/mesh/hello.slang.glsl
index 3a0848dcb..0b3d4acb3 100644
--- a/tests/pipeline/rasterization/mesh/hello.slang.glsl
+++ b/tests/pipeline/rasterization/mesh/hello.slang.glsl
@@ -2,12 +2,12 @@
 #extension GL_EXT_mesh_shader : require
 layout(row_major) uniform;
 layout(row_major) buffer;
-const vec3  colors_0[3] = { vec3(1.0, 1.0, 0.0), vec3(0.0, 1.0, 1.0), vec3(1.0, 0.0, 1.0) };
 const vec2  positions_0[3] = { vec2(0.0, -0.5), vec2(0.5, 0.5), vec2(-0.5, 0.5) };
-
+const vec3  colors_0[3] = { vec3(1.0, 1.0, 0.0), vec3(0.0, 1.0, 1.0), vec3(1.0, 0.0, 1.0) };
 layout(location = 0)
 out vec3  _S1[3];
 
+
 out gl_MeshPerVertexEXT
 {
     vec4 gl_Position;
@@ -21,6 +21,7 @@ layout(max_primitives = 1) out;
 layout(triangles) out;
 void main()
 {
+
     SetMeshOutputsEXT(3U, 1U);
     if(gl_LocalInvocationIndex < 3U)
     {
@@ -30,6 +31,7 @@ void main()
     else
     {
     }
+
     if(gl_LocalInvocationIndex < 1U)
     {
         gl_PrimitiveTriangleIndicesEXT[gl_LocalInvocationIndex] = uvec3(0U, 1U, 2U);
@@ -37,6 +39,6 @@ void main()
     else
     {
     }
+
     return;
 }
-
diff --git a/tests/pipeline/rasterization/mesh/hello.slang.hlsl b/tests/pipeline/rasterization/mesh/hello.slang.hlsl
index a10dc6884..06871fedd 100644
--- a/tests/pipeline/rasterization/mesh/hello.slang.hlsl
+++ b/tests/pipeline/rasterization/mesh/hello.slang.hlsl
@@ -4,16 +4,15 @@
 #endif
 #pragma warning(disable: 3557)
 
-static const float3  colors_0[int(3)] = { float3(1.0, 1.0, 0.0), float3(0.0, 1.0, 1.0), float3(1.0, 0.0, 1.0) };
 static const float2  positions_0[int(3)] = { float2(0.0, -0.5), float2(0.5, 0.5), float2(-0.5, 0.5) };
+static const float3  colors_0[int(3)] = { float3(1.0, 1.0, 0.0), float3(0.0, 1.0, 1.0), float3(1.0, 0.0, 1.0) };
 struct Vertex_0
 {
     float4 pos_0 : SV_Position;
     float3 color_0 : Color;
 };
 
-[shader("mesh")]
-[numthreads(3, 1, 1)]
+[shader("mesh")][numthreads(3, 1, 1)]
 [outputtopology("triangle")]
 void main(uint tig_0 : SV_GROUPINDEX, vertices vertices out Vertex_0  verts_0[int(3)], indices indices out uint3  triangles_0[int(1)])
 {
@@ -26,6 +25,7 @@ void main(uint tig_0 : SV_GROUPINDEX, vertices vertices out Vertex_0  verts_0[in
     else
     {
     }
+
     if(tig_0 < 1U)
     {
         triangles_0[tig_0] = uint3(0U, 1U, 2U);
@@ -33,6 +33,6 @@ void main(uint tig_0 : SV_GROUPINDEX, vertices vertices out Vertex_0  verts_0[in
     else
     {
     }
+
     return;
 }
-
diff --git a/tests/pipeline/rasterization/mesh/primitive-output.slang.glsl b/tests/pipeline/rasterization/mesh/primitive-output.slang.glsl
index 35efcf4af..cfe266bec 100644
--- a/tests/pipeline/rasterization/mesh/primitive-output.slang.glsl
+++ b/tests/pipeline/rasterization/mesh/primitive-output.slang.glsl
@@ -2,9 +2,10 @@
 #extension GL_EXT_mesh_shader : require
 layout(row_major) uniform;
 layout(row_major) buffer;
-const vec3  colors_0[3] = { vec3(1.0, 1.0, 0.0), vec3(0.0, 1.0, 1.0), vec3(1.0, 0.0, 1.0) };
 const vec2  positions_0[3] = { vec2(0.0, -0.5), vec2(0.5, 0.5), vec2(-0.5, 0.5) };
+const vec3  colors_0[3] = { vec3(1.0, 1.0, 0.0), vec3(0.0, 1.0, 1.0), vec3(1.0, 0.0, 1.0) };
 out uvec3  gl_PrimitiveTriangleIndicesEXT[1];
+
 layout(location = 0)
 out vec3  _S1[3];
 
@@ -28,6 +29,7 @@ layout(max_primitives = 1) out;
 layout(triangles) out;
 void main()
 {
+
     SetMeshOutputsEXT(3U, 1U);
     if(gl_LocalInvocationIndex < 3U)
     {
@@ -37,6 +39,7 @@ void main()
     else
     {
     }
+
     if(gl_LocalInvocationIndex < 1U)
     {
         gl_PrimitiveTriangleIndicesEXT[gl_LocalInvocationIndex] = uvec3(0U, 1U, 2U);
@@ -47,6 +50,6 @@ void main()
     else
     {
     }
+
     return;
 }
-