Capabilities generator inclusive join and misc (#4237)

7 files changed, 237 insertions, 174 deletions

diff --git a/source/core/core.natvis b/source/core/core.natvis
index 2448e2c88..f2547b3fe 100644
--- a/source/core/core.natvis
+++ b/source/core/core.natvis
@@ -55,62 +55,44 @@
 </Type>
   
 <Type Name="Slang::Dictionary&lt;*,*&gt;">
+    <DisplayString>{{ {map.m_values} }}</DisplayString>
+    <Expand>
+        <Item Name="[items]">map.m_values</Item>
+        <Item Name="map">map</Item>
+    </Expand>
+</Type>
+
+<Type Name="Slang::HashSet&lt;*&gt;">
+    <DisplayString>{{ {dict} }}</DisplayString>
+    <Expand>
+        <Item Name="dict">dict</Item>
+    </Expand>
+</Type>
+    
+<Type Name="Slang::OrderedHashSet&lt;*&gt;">
+    <DisplayString>{{ size={dict._count} }}</DisplayString>
+    <Expand>
+        <LinkedListItems>
+            <Size>m_dict._count</Size>
+            <HeadPointer>m_dict.kvPairs.head</HeadPointer>
+            <NextPointer>next</NextPointer>
+            <ValueNode>value</ValueNode>
+        </LinkedListItems>
+    </Expand>
+</Type>
+
+<Type Name="Slang::OrderedDictionary&lt;*,*&gt;">
     <DisplayString>{{ size={m_count} }}</DisplayString>
     <Expand>
-        <Item Name="[size]">m_count</Item>
-        <Item Name="[capacity]">m_bucketCountMinusOne + 1</Item>
-        <CustomListItems MaxItemsPerView="5000" ExcludeView="Test">
-            <Variable Name="iBucket" InitialValue="0" />
-            <Variable Name="pBucket" InitialValue="m_hashMap" />
-            <Variable Name="isDeleted" InitialValue="0" />
-            <Variable Name="isEmpty" InitialValue="0" />
+        <LinkedListItems>
             <Size>m_count</Size>
-            <Exec>pBucket = m_hashMap</Exec>
-            <Loop>
-                <If Condition="iBucket &gt;= m_bucketCountMinusOne">
-                    <Break/>
-                </If>
-                <Exec>
-                    isDeleted = m_marks.m_buffer.m_count &gt; (iBucket*2+1)/32
-                    ? ((m_marks.m_buffer.m_buffer[(iBucket*2+1)/32]&amp;(1&lt;&lt;(iBucket*2+1)%32)) != 0)
-                      : 0
-                 </Exec>
-                <Exec>
-                    isEmpty = m_marks.m_buffer.m_count &gt; (iBucket*2)/32
-                    ? ((m_marks.m_buffer.m_buffer[(iBucket*2)/32]&amp;(1&lt;&lt;(iBucket*2)%32)) == 0)
-                    : 1
-                </Exec>
-                <If Condition="isDeleted+isEmpty==0">
-                    <Item>*(m_hashMap + iBucket)</Item>
-                </If>
-                <Exec>iBucket++</Exec>
-            </Loop>
-        </CustomListItems>
+            <HeadPointer>m_kvPairs.head</HeadPointer>
+            <NextPointer>next</NextPointer>
+            <ValueNode>value</ValueNode>
+        </LinkedListItems>
     </Expand>
 </Type>
 
-    <Type Name="Slang::OrderedHashSet&lt;*&gt;">
-        <DisplayString>{{ size={dict._count} }}</DisplayString>
-        <Expand>
-            <LinkedListItems>
-                <Size>m_dict._count</Size>
-                <HeadPointer>m_dict.kvPairs.head</HeadPointer>
-                <NextPointer>next</NextPointer>
-                <ValueNode>value</ValueNode>
-            </LinkedListItems>
-        </Expand>
-    </Type>
-    <Type Name="Slang::OrderedDictionary&lt;*,*&gt;">
-        <DisplayString>{{ size={m_count} }}</DisplayString>
-        <Expand>
-            <LinkedListItems>
-                <Size>m_count</Size>
-                <HeadPointer>m_kvPairs.head</HeadPointer>
-                <NextPointer>next</NextPointer>
-                <ValueNode>value</ValueNode>
-            </LinkedListItems>
-        </Expand>
-    </Type>
 <Type Name="Slang::RefPtr&lt;*&gt;">
     <SmartPointer Usage="Minimal">pointer</SmartPointer>
     <DisplayString Condition="pointer == 0">empty</DisplayString>
diff --git a/source/core/slang-uint-set.h b/source/core/slang-uint-set.h
index 22ca457b0..7e08500fd 100644
--- a/source/core/slang-uint-set.h
+++ b/source/core/slang-uint-set.h
@@ -14,19 +14,9 @@
 namespace Slang
 {
 
-template<typename T>
-constexpr static Index computeElementShift()
+constexpr Index intLog2(unsigned x)
 {
-    Index currentShift = 0;
-    Index currentShiftValue = 1;
-
-    while (currentShiftValue != sizeof(T) * 8)
-    {
-        currentShift++;
-        currentShiftValue *= 2;
-    }
-
-    return currentShift;
+    return x == 1 ? 0 : 1 + intLog2(x >> 1);
 }
 
 static inline Index bitscanForward(uint64_t in)
@@ -65,13 +55,12 @@ public:
     
     constexpr static Index kElementSize = sizeof(Element) * 8; ///< The number of bits in an element. This also determines how many values a element can hold.
     constexpr static Index kElementMask = kElementSize - 1; ///< Mask to get shift from an index
-    constexpr static Index kElementShift = computeElementShift<Element>(); ///< How many bits to shift to get Element index from an index. 5 for 2^5=32 elements in a uint32_t. 6 for 2^6=64 in a uint64_t.
+    constexpr static Index kElementShift = intLog2(sizeof(Element)*8); ///< How many bits to shift to get Element index from an index. 5 for 2^5=32 elements in a uint32_t. 6 for 2^6=64 in a uint64_t.
 
     UIntSet() {}
     UIntSet(const UIntSet& other) { m_buffer = other.m_buffer; }
     UIntSet(UIntSet && other) { *this = (_Move(other)); }
     UIntSet(UInt maxVal) { resizeAndClear(maxVal); }
-    UIntSet(List<UIntSet::Element> buffer) { m_buffer = buffer; }
 
     UIntSet& operator=(UIntSet&& other);
     UIntSet& operator=(const UIntSet& other);
@@ -81,7 +70,7 @@ public:
     /// Return the count of all bits directly represented
     Int getCount() const { return Int(m_buffer.getCount()) * kElementSize; }
 
-    List<Element>& getBuffer() { return m_buffer; }
+    const List<Element>& getBuffer() const { return m_buffer; }
 
         /// Resize such that val can be stored and clear contents
     void resizeAndClear(UInt val);
@@ -101,6 +90,9 @@ public:
         /// Add a value
     inline void add(UInt val);
     inline void add(const UIntSet& val);
+    inline void addRange(const List<UInt>& other);
+
+    inline void addRawElement(Element val, Index bitOffset);
 
         /// Remove a value
     inline void remove(UInt val);
@@ -145,37 +137,38 @@ public:
     {
         friend class UIntSet;
     private:
-        const List<Element>* context;
-        Index block = 0;
-        Element processedElement = 0;
-        uint64_t LSB = 0;
+        const List<Element>* m_context;
+        Index m_block = 0;
+        Element m_processedElement = 0;
+        uint64_t m_LSB = 0;
 
         void clearLSB()
         {
-            LSB = bitscanForward(processedElement);
-            processedElement &= processedElement - 1;
+            m_LSB = bitscanForward(m_processedElement);
+            m_processedElement &= m_processedElement - 1;
         }
-    public:
-        Iterator(const List<Element>* inContext)
+
+        Iterator(const List<Element>* context)
         {
-            context = inContext;
+            m_context = context;
         }
+    public:
 
         Element operator*()
         {
-            return Element(LSB + (kElementSize * block));
+            return Element(m_LSB + (kElementSize * m_block));
         }
 
         Iterator& operator++()
         {
-            while (processedElement == 0)
+            while (m_processedElement == 0)
             {
-                block++;
-                if (block >= context->getCount())
+                m_block++;
+                if (m_block >= m_context->getCount())
                 {
                     return *this;
                 }
-                processedElement = (*context)[block];
+                m_processedElement = (*m_context)[m_block];
             }
             clearLSB();
             return *this;
@@ -186,8 +179,8 @@ public:
         }
         bool operator==(const Iterator& other) const
         {
-            return other.block == this->block
-                && other.processedElement == this->processedElement;
+            return other.m_block == this->m_block
+                && other.m_processedElement == this->m_processedElement;
         }
         bool operator!=(const Iterator& other) const
         {
@@ -200,19 +193,21 @@ public:
         if (m_buffer.getCount() == 0)
             return tmp;
 
-        tmp.processedElement = m_buffer[0];
-        if (tmp.processedElement == 0)
+        tmp.m_processedElement = m_buffer[0];
+        if (tmp.m_processedElement == 0)
+        {
             tmp++;
+            return tmp;
+        }
 
         tmp.clearLSB();
-
         return tmp;
     }
     Iterator end() const
     {
         Iterator tmp(&m_buffer);
-        tmp.block = m_buffer.getCount();
-        tmp.processedElement = 0;
+        tmp.m_block = m_buffer.getCount();
+        tmp.m_processedElement = 0;
         return tmp;
     }
 
@@ -307,6 +302,19 @@ inline void UIntSet::add(const UIntSet& other)
         m_buffer[i] |= other.m_buffer[i];
 }
 
+inline void UIntSet::addRange(const List<UInt>& other)
+{
+    for (auto i : other)
+        add(i);
+}
+
+inline void UIntSet::addRawElement(Element other, Index elementIndex)
+{
+    if(this->m_buffer.getCount() <= elementIndex)
+        resizeBackingBufferDirectly(elementIndex+1);
+    m_buffer[elementIndex] |= other;
+}
+
 template<typename T>
 List<T> UIntSet::getElements() const
 {
diff --git a/source/slang/hlsl.meta.slang b/source/slang/hlsl.meta.slang
index 1458467b9..13b1f0131 100644
--- a/source/slang/hlsl.meta.slang
+++ b/source/slang/hlsl.meta.slang
@@ -3665,7 +3665,7 @@ struct InputPatch
 };
 
 __generic<T, let N : int>
-[require(glsl_hlsl_spirv, shader_stages_domain_hull)]
+[require(glsl_hlsl_spirv, domain_hull)]
 __magic_type(HLSLOutputPatchType)
 __intrinsic_type($(kIROp_HLSLOutputPatchType))
 struct OutputPatch
diff --git a/source/slang/slang-capabilities.capdef b/source/slang/slang-capabilities.capdef
index 5a0df9f9b..7567dd617 100644
--- a/source/slang/slang-capabilities.capdef
+++ b/source/slang/slang-capabilities.capdef
@@ -20,10 +20,22 @@
 // There are three types of capability definitions:
 // - `def`: this will introduce an new capability atom. If there is an inheritance clause,
 //    the capability name will expand to all inherited atoms plus the newly introduced atom.
-// - `abstract`: an abstract capability does not introduce an actual atom, but it defines
-//   an implicit conflict group such that two capabilities inheriting the same abstract
-//   capability cannot be satisfied simultaneously. When used in an expression, `abstract`
-//   capability also expands into a disjunction (or) of all inherited capabilities.
+// - `abstract`: an abstract capability does not introduce an actual atom. Primarily used to 
+//   make disjunctions using the following rules:
+//      1. Each type of abstract atom defines a "keyhole". `target` and `stage` are distinct "keyholes".
+//      2. Any atom immediately derived off of an abstract atom is a "key atom". 
+//      3. Every conjunction may only populate each "keyhole" once, else the set is incompatible.
+//          * If joining ('+') incompatible sets, an invalid capability is made.
+//          * If Inclusive joining ('|') incompatible sets, a disjunction is made.
+//      4. ex: `hlsl + glsl` both populate the same "keyhole" of `target`, this is incompatible. 
+//      5. ex: `hlsl + _sm_6_0` both populate same keyholes of `hlsl` and `hlsl`, this is compatible.
+//      6. ex: `hlsl | glsl` creates a disjunction of 2 sets: `hlsl` and `glsl`. This is due to the 
+//         2 sets being incompatible.
+//      7. Having a not populated "keyhole" means your set is compatible with any "key atom" of 
+//         distinct "keyhole".
+//      8. ex: `vertex + glsl` works because 'vertex' has a unpopulated `target` "keyhole", and 
+//         therefore is compatible with all `target` "key atoms"
+
 // - `alias`: this defines an alias and does not introduce actual atoms.
 
 // Several capabilities represent the target formats in which we generate code.
@@ -173,27 +185,23 @@ alias tesscontrol = hull;
 alias tesseval = domain;
 alias amplification_mesh = amplification | mesh;
 alias raytracing_stages = raygen | intersection | anyhit | closesthit | miss | callable;
-alias raytracing_stages_intersection = intersection;
-alias raytracing_stages_raygen = raygen;
-alias raytracing_stages_anyhit_closesthit = anyhit | closesthit;
-alias raytracing_stages_raygen_closesthit_miss = raygen | closesthit | miss;
-alias raytracing_stages_anyhit_closesthit_intersection = anyhit | closesthit | intersection;
-alias raytracing_stages_anyhit_closesthit_intersection_miss = anyhit | closesthit | intersection | miss;
-alias raytracing_stages_raygen_closesthit_miss_callable = raygen | closesthit | miss | callable;
-alias shader_stages_compute_tesscontrol_tesseval = compute | tesscontrol | tesseval;
-alias shader_stages_compute_fragment = compute | fragment;
-alias shader_stages_compute_fragment_geometry_vertex = compute | fragment | geometry | vertex;
-alias shader_stages_domain_hull = domain | hull;
-alias raytracing_stages_fragment = raytracing_stages | fragment;
-alias raytracing_stages_compute = raytracing_stages | compute;
-alias raytracing_stages_compute_amplification_mesh = raytracing_stages_compute | amplification_mesh;
-alias raytracing_stages_compute_fragment = raytracing_stages | shader_stages_compute_fragment;
-alias raytracing_stages_compute_fragment_geometry_vertex = raytracing_stages | shader_stages_compute_fragment_geometry_vertex;
+alias anyhit_closesthit = anyhit | closesthit;
+alias raygen_closesthit_miss = raygen | closesthit | miss;
+alias anyhit_closesthit_intersection = anyhit | closesthit | intersection;
+alias anyhit_closesthit_intersection_miss = anyhit | closesthit | intersection | miss;
+alias raygen_closesthit_miss_callable = raygen | closesthit | miss | callable;
+alias compute_tesscontrol_tesseval = compute | tesscontrol | tesseval;
+alias compute_fragment = compute | fragment;
+alias compute_fragment_geometry_vertex = compute | fragment | geometry | vertex;
+alias domain_hull = domain | hull;
+alias raytracingstages_fragment = raytracing_stages | fragment;
+alias raytracingstages_compute = raytracing_stages | compute;
+alias raytracingstages_compute_amplification_mesh = raytracingstages_compute | amplification_mesh;
+alias raytracingstages_compute_fragment = raytracing_stages | compute_fragment;
+alias raytracingstages_compute_fragment_geometry_vertex = raytracing_stages | compute_fragment_geometry_vertex;
 
 // SPIRV extensions.
 
-def SOURCE_EXT_GL_NV_compute_shader_derivatives : spirv_1_0;
-
 def SPV_EXT_fragment_shader_interlock : spirv_1_0;
 def SPV_KHR_fragment_shader_barycentric : spirv_1_0;
 def SPV_EXT_fragment_fully_covered : spirv_1_0;
@@ -342,7 +350,7 @@ alias GL_KHR_shader_subgroup_shuffle = _GL_KHR_shader_subgroup_shuffle | spvGrou
 alias GL_KHR_shader_subgroup_shuffle_relative = _GL_KHR_shader_subgroup_shuffle_relative | spvGroupNonUniformShuffle;
 alias GL_KHR_shader_subgroup_vote = _GL_KHR_shader_subgroup_vote | spvGroupNonUniformVote;
 alias GL_KHR_shader_subgroup_quad = _GL_KHR_shader_subgroup_quad | spvGroupNonUniformQuad;
-alias GL_NV_compute_shader_derivatives = _GL_NV_compute_shader_derivatives | SOURCE_EXT_GL_NV_compute_shader_derivatives | SPV_NV_compute_shader_derivatives | _sm_6_6;
+alias GL_NV_compute_shader_derivatives = _GL_NV_compute_shader_derivatives | SPV_NV_compute_shader_derivatives | _sm_6_6;
 alias GL_ARB_shader_image_size = _GL_ARB_shader_image_size | spvImageQuery | metal;
 alias GL_ARB_shader_texture_image_samples = _GL_ARB_shader_texture_image_samples | spvImageQuery | metal;
 alias GL_NV_shader_atomic_fp16_vector = _GL_NV_shader_atomic_fp16_vector + _GL_NV_gpu_shader5 | spirv_1_0;
@@ -364,8 +372,8 @@ alias rayquery = GL_EXT_ray_query | _sm_6_5;
 alias raytracing_motionblur =  raytracing + motionblur | cuda;
 alias ser_motion = ser + motionblur;
 alias shaderclock = GL_EXT_shader_realtime_clock | hlsl_nvapi | cpp | cuda;
-alias meshshading_internal = spvMeshShadingEXT + _sm_6_5 + _GL_EXT_mesh_shader;
-alias meshshading = amplification + meshshading_internal | mesh + meshshading_internal;
+alias meshshading_internal = GL_EXT_mesh_shader | _sm_6_5;
+alias meshshading = amplification_mesh + meshshading_internal;
 alias fragmentshaderinterlock = _GL_ARB_fragment_shader_interlock | hlsl_nvapi | spvFragmentShaderPixelInterlockEXT;
 alias atomic64 = GL_EXT_shader_atomic_int64 | _sm_6_6 | cpp | cuda;
 alias atomicfloat = GL_EXT_shader_atomic_float | _sm_6_0 + hlsl_nvapi | cpp | cuda;
@@ -375,7 +383,7 @@ alias groupnonuniform = GL_KHR_shader_subgroup_ballot + GL_KHR_shader_subgroup_s
                         | _sm_6_0 | cuda;
 alias fragmentshaderbarycentric = GL_EXT_fragment_shader_barycentric | _sm_6_1;
 alias shadermemorycontrol = glsl | spirv_1_0 | _sm_5_0;
-alias shadermemorycontrol_compute = raytracing_stages_compute + shadermemorycontrol;
+alias shadermemorycontrol_compute = raytracingstages_compute + shadermemorycontrol;
 alias subpass = fragment + any_gfx_target;
 alias waveprefix = _sm_6_5 | _cuda_sm_7_0 | GL_KHR_shader_subgroup_arithmetic;
 alias bufferreference = GL_EXT_buffer_reference;
@@ -564,7 +572,7 @@ alias GLSL_460 = _GLSL_460
                | cpp
                ;
 
-alias GLSL_410_SPIRV_1_0 = _GLSL_410 + GLSL_400 | GLSL_400;
+alias GLSL_410_SPIRV_1_0 = _GLSL_410 | spirv_1_0;
 alias GLSL_420_SPIRV_1_0 = _GLSL_420 + GLSL_410_SPIRV_1_0 | GLSL_410_SPIRV_1_0;
 alias GLSL_430_SPIRV_1_0 = _GLSL_430 + GLSL_420_SPIRV_1_0 | GLSL_420_SPIRV_1_0;
 
@@ -587,41 +595,41 @@ alias METAL_2_4 = metallib_2_4;
 
 alias sm_2_0_GLSL_140 = _GLSL_140 + sm_4_0 | sm_4_0;
 alias sm_2_0_GLSL_400 = _GLSL_400 + sm_4_0 | sm_4_0;
-alias appendstructuredbuffer = sm_5_0 + raytracing_stages_compute_fragment;
+alias appendstructuredbuffer = sm_5_0 + raytracingstages_compute_fragment;
 alias atomic_hlsl = _sm_4_0;
 alias atomic_hlsl_nvapi = _sm_4_0 + hlsl_nvapi;
 alias atomic_hlsl_sm_6_6 = _sm_6_6;
 alias byteaddressbuffer = sm_4_0;
-alias byteaddressbuffer_rw = sm_4_0 + raytracing_stages_compute_fragment;
-alias consumestructuredbuffer = sm_5_0 + raytracing_stages_compute_fragment;
+alias byteaddressbuffer_rw = sm_4_0 + raytracingstages_compute_fragment;
+alias consumestructuredbuffer = sm_5_0 + raytracingstages_compute_fragment;
 alias fragmentprocessing = fragment + _sm_5_0
                          | fragment + glsl_spirv
-                         | raytracing_stages_compute_amplification_mesh + GL_NV_compute_shader_derivatives
+                         | raytracingstages_compute_amplification_mesh + GL_NV_compute_shader_derivatives
                          ;
 alias fragmentprocessing_derivativecontrol = fragment + _sm_5_0
                                            | fragment + GL_ARB_derivative_control
-                                           | raytracing_stages_compute_amplification_mesh + GL_NV_compute_shader_derivatives
+                                           | raytracingstages_compute_amplification_mesh + GL_NV_compute_shader_derivatives
                                            ;
 alias getattributeatvertex = fragment + _sm_6_1 | fragment + GL_EXT_fragment_shader_barycentric;
-alias memorybarrier_compute = raytracing_stages_compute + sm_5_0;
+alias memorybarrier_compute = raytracingstages_compute + sm_5_0;
 alias glsl_barrier = hlsl + memorybarrier_compute 
-                   | glsl_spirv + shader_stages_compute_tesscontrol_tesseval
+                   | glsl_spirv + compute_tesscontrol_tesseval
                    ;
 alias structuredbuffer = sm_4_0;
-alias structuredbuffer_rw = sm_4_0 + raytracing_stages_compute_fragment;
+alias structuredbuffer_rw = sm_4_0 + raytracingstages_compute_fragment;
 alias texture_sm_4_1 = sm_4_1
                      ;
 alias texture_sm_4_1_samplerless = cpp + texture_sm_4_1
                               | cuda + texture_sm_4_1
                               | glsl + texture_sm_4_1 + GL_EXT_samplerless_texture_functions
-                              | hlsl + texture_sm_4_1 + raytracing_stages_compute_fragment
+                              | hlsl + texture_sm_4_1 + raytracingstages_compute_fragment
                               | spirv_1_0 + texture_sm_4_1 + GL_EXT_samplerless_texture_functions
                               | metal + texture_sm_4_1
                               ;
 alias texture_sm_4_1_compute_fragment = cpp + texture_sm_4_1
                               | cuda + texture_sm_4_1
                               | glsl + texture_sm_4_1
-                              | hlsl + texture_sm_4_1 + raytracing_stages_compute_fragment
+                              | hlsl + texture_sm_4_1 + raytracingstages_compute_fragment
                               | spirv_1_0 + texture_sm_4_1
                               | metal + texture_sm_4_1
                               ;
@@ -629,7 +637,7 @@ alias texture_sm_4_1_compute_fragment = cpp + texture_sm_4_1
 alias texture_sm_4_1_fragment = cpp + texture_sm_4_1
                               | cuda + texture_sm_4_1
                               | glsl + texture_sm_4_1
-                              | hlsl + texture_sm_4_1 + raytracing_stages_compute_fragment
+                              | hlsl + texture_sm_4_1 + raytracingstages_compute_fragment
                               | spirv_1_0 + texture_sm_4_1
                               | metal + texture_sm_4_1
                               ;
@@ -637,7 +645,7 @@ alias texture_sm_4_1_clamp_fragment = texture_sm_4_1_fragment + GL_ARB_sparse_te
 alias texture_sm_4_1_vertex_fragment_geometry = cpp + texture_sm_4_1
                                               | cuda + texture_sm_4_1
                                               | glsl + texture_sm_4_1
-                                              | hlsl + texture_sm_4_1 + raytracing_stages_compute_fragment_geometry_vertex
+                                              | hlsl + texture_sm_4_1 + raytracingstages_compute_fragment_geometry_vertex
                                               | spirv_1_0 + texture_sm_4_1
                                               | metal + texture_sm_4_1
                                               ;
@@ -665,8 +673,8 @@ alias printf = GL_EXT_debug_printf | _sm_4_0 | _cuda_sm_2_0 | cpp;
 alias texturefootprint = GL_NV_shader_texture_footprint + GLSL_450 | hlsl_nvapi + _sm_4_0;
 alias texturefootprintclamp = texturefootprint + GL_ARB_sparse_texture_clamp;
 
-alias shader5_sm_4_0 = GL_ARB_gpu_shader5 + sm_4_0 | sm_4_0;
-alias shader5_sm_5_0 = GL_ARB_gpu_shader5 + sm_4_0 | sm_5_0;
+alias shader5_sm_4_0 = GL_ARB_gpu_shader5 + _GLSL_140 + sm_4_0 | sm_4_0;
+alias shader5_sm_5_0 = GL_ARB_gpu_shader5 + _GLSL_140 + sm_4_0 | sm_5_0;
 
 alias subgroup_basic = GL_KHR_shader_subgroup_basic | _sm_6_0 | _cuda_sm_7_0;
 alias subgroup_ballot = spirv_1_0 + GL_KHR_shader_subgroup_ballot
@@ -681,7 +689,8 @@ alias subgroup_ballot_activemask = spirv_1_0 + GL_KHR_shader_subgroup_ballot
                       ;
 alias subgroup_basic_ballot = glsl + GL_KHR_shader_subgroup_basic + subgroup_ballot 
                             | spirv + GL_KHR_shader_subgroup_basic + subgroup_ballot
-                            | hlsl + subgroup_ballot | cuda + subgroup_ballot
+                            | hlsl + subgroup_ballot
+                            | cuda + subgroup_ballot
                             ;
 alias subgroup_vote = GL_KHR_shader_subgroup_vote | _sm_6_0 | _cuda_sm_7_0;
 alias shaderinvocationgroup = subgroup_vote;
@@ -705,23 +714,23 @@ alias breakpoint = GL_EXT_debug_printf | hlsl | _cuda_sm_8_0 | cpp;
 alias rayobject = raytracing | rayquery;
 alias raytracing_allstages = raytracing_stages + raytracing;
 alias raytracing_anyhit = anyhit + raytracing;
-alias raytracing_intersection = raytracing_stages_intersection + raytracing;
-alias raytracing_anyhit_closesthit = raytracing_stages_anyhit_closesthit + raytracing;
-alias raytracing_anyhit_closesthit_intersection = raytracing_stages_anyhit_closesthit_intersection + raytracing;
-alias raytracing_raygen_closesthit_miss = raytracing_stages_raygen_closesthit_miss + raytracing;
-alias raytracing_anyhit_closesthit_intersection_miss = raytracing_stages_anyhit_closesthit_intersection_miss + raytracing;
-alias raytracing_raygen_closesthit_miss_callable = raytracing_stages_raygen_closesthit_miss_callable + raytracing;
-alias raytracing_position = raytracing + GL_EXT_ray_tracing_position_fetch + anyhit + closesthit;
-alias raytracing_motionblur_anyhit_closesthit_intersection_miss = raytracing_stages_anyhit_closesthit_intersection_miss + raytracing_motionblur;
-alias raytracing_motionblur_raygen_closesthit_miss = raytracing_stages_raygen_closesthit_miss + raytracing_motionblur;
+alias raytracing_intersection = intersection + raytracing;
+alias raytracing_anyhit_closesthit = anyhit_closesthit + raytracing;
+alias raytracing_anyhit_closesthit_intersection = anyhit_closesthit_intersection + raytracing;
+alias raytracing_raygen_closesthit_miss = raygen_closesthit_miss + raytracing;
+alias raytracing_anyhit_closesthit_intersection_miss = anyhit_closesthit_intersection_miss + raytracing;
+alias raytracing_raygen_closesthit_miss_callable = raygen_closesthit_miss_callable + raytracing;
+alias raytracing_position = raytracing + GL_EXT_ray_tracing_position_fetch + anyhit_closesthit;
+alias raytracing_motionblur_anyhit_closesthit_intersection_miss = anyhit_closesthit_intersection_miss + raytracing_motionblur;
+alias raytracing_motionblur_raygen_closesthit_miss = raygen_closesthit_miss + raytracing_motionblur;
 alias rayquery_position = rayquery + GL_EXT_ray_tracing_position_fetch;
-alias ser_raygen = raytracing_stages_raygen + ser;
-alias ser_raygen_closesthit_miss = raytracing_stages_raygen_closesthit_miss + ser;
-alias ser_any_closesthit_intersection_miss = raytracing_stages_anyhit_closesthit_intersection_miss + ser;
-alias ser_anyhit_closesthit_intersection = raytracing_stages_anyhit_closesthit_intersection + ser;
-alias ser_anyhit_closesthit = raytracing_stages_anyhit_closesthit + ser;
-alias ser_motion_raygen_closesthit_miss = raytracing_stages_raygen_closesthit_miss + ser_motion;
-alias ser_motion_raygen = raytracing_stages_raygen + ser_motion;
+alias ser_raygen = raygen + ser;
+alias ser_raygen_closesthit_miss = raygen_closesthit_miss + ser;
+alias ser_any_closesthit_intersection_miss = anyhit_closesthit_intersection_miss + ser;
+alias ser_anyhit_closesthit_intersection = anyhit_closesthit_intersection + ser;
+alias ser_anyhit_closesthit = anyhit_closesthit + ser;
+alias ser_motion_raygen_closesthit_miss = raygen_closesthit_miss + ser_motion;
+alias ser_motion_raygen = raygen + ser_motion;
 
 alias all = _sm_6_7 + hlsl_nvapi
           | glsl_spirv_1_5 + sm_6_7
diff --git a/source/slang/slang-capability.cpp b/source/slang/slang-capability.cpp
index eae04f277..e77901b5b 100644
--- a/source/slang/slang-capability.cpp
+++ b/source/slang/slang-capability.cpp
@@ -25,11 +25,11 @@ enum class CapabilityNameFlavor : int32_t
     Concrete,
 
     // An abstract capability represents a class of feature
-    // where multiple different implementations might be possible.
-    // For example, "ray tracing" might be an abstract feature
-    // that a function can require, but a specific target will
-    // only be able to provide that abstract feature via some
-    // specific concrete feature (e.g., `GL_EXT_ray_tracing`).
+    // where multiple distinct implementations might be possible.
+    // 'raytracing' may be allowed with a 'raygen' "stage", but 
+    // not a 'vertex' "stage"
+    // For more information (and a clearer description of the rules), 
+    // read `slang-capabilities.capdef`
     Abstract,
 
     // An alias capability atom is one that is exactly equivalent
@@ -490,9 +490,6 @@ CapabilitySet CapabilitySet::getTargetsThisHasButOtherDoesNot(const CapabilitySe
     return newSet;
 }
 
-/// Join `this` with a compatble stage set of `CapabilityTargetSet other`.
-/// Return false when `other` is fully incompatible.
-/// incompatability is when `this->stage` is not a supported stage by `other.shaderStageSets`.
 bool CapabilityStageSet::tryJoin(const CapabilityTargetSet& other)
 {
     const CapabilityStageSet* otherStageSet = other.shaderStageSets.tryGetValue(this->stage);
@@ -506,11 +503,6 @@ bool CapabilityStageSet::tryJoin(const CapabilityTargetSet& other)
     return true;
 }
 
-/// Join a compatable target set from `this` with `CapabilityTargetSet other`.
-/// Return false when `other` is fully incompatible.
-/// incompatability is when one of 2 senarios are true:
-/// 1. `this->target` is not a supported target by `other.shaderStageSets`
-/// 2. `this` has completly disjoint shader stages from other.
 bool CapabilityTargetSet::tryJoin(const CapabilityTargetSets& other)
 {
     const CapabilityTargetSet* otherTargetSet = other.tryGetValue(this->target);
@@ -631,9 +623,6 @@ bool CapabilitySet::hasSameTargets(const CapabilitySet& other) const
 #pragma warning(push)
 #pragma warning(disable:4702)
 #endif
-/// returns true if 'this' is a better target for 'targetCaps' than 'that'
-/// isEqual: is `this` and `that` equal
-/// isIncompatible: is `this` and `that` incompatible
 bool CapabilitySet::isBetterForTarget(CapabilitySet const& that, CapabilitySet const& targetCaps, bool& isEqual) const
 {
     if (this->isEmpty() && (that.isEmpty() || that.isInvalid()))
@@ -882,6 +871,16 @@ int TEST_findTargetStage(
     return capSet.getCapabilityTargetSets()[target].shaderStageSets.containsKey(stage);
 }
 
+
+int TEST_targetCapSetWithSpecificAtomInStage(
+    CapabilitySet& capSet,
+    CapabilityAtom target,
+    CapabilityAtom stage,
+    CapabilityAtom atom)
+{
+    return capSet.getCapabilityTargetSets()[target].shaderStageSets[stage].atomSet->contains((UInt)atom);
+}
+
 int TEST_targetCapSetWithSpecificSetInStage(
     CapabilitySet& capSet,
     CapabilityAtom target,
@@ -943,8 +942,8 @@ void TEST_CapabilitySet_addAtom()
     testCapSet = CapabilitySet(CapabilityName::TEST_ADD_2);
 
     CHECK_CAPS(TEST_findTargetCapSet(testCapSet, CapabilityAtom::hlsl));
-    CHECK_CAPS(TEST_targetCapSetWithSpecificSetInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::vertex,
-        { CapabilityAtom::textualTarget, CapabilityAtom::hlsl, CapabilityAtom::vertex,
+    CHECK_CAPS(TEST_targetCapSetWithSpecificSetInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::compute,
+        { CapabilityAtom::textualTarget, CapabilityAtom::hlsl, CapabilityAtom::compute,
         CapabilityAtom::_sm_4_0, CapabilityAtom::_sm_4_1 }));
     CHECK_CAPS(TEST_targetCapSetWithSpecificSetInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment,
         { CapabilityAtom::textualTarget, CapabilityAtom::hlsl, CapabilityAtom::fragment,
@@ -959,7 +958,54 @@ void TEST_CapabilitySet_addAtom()
     CHECK_CAPS(TEST_targetCapSetWithSpecificSetInStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment,
         { CapabilityAtom::textualTarget, CapabilityAtom::glsl, CapabilityAtom::fragment,
         CapabilityAtom::_GLSL_130 }));
+
+    // ------------------------------------------------------------
+
+    testCapSet = CapabilitySet(CapabilityName::TEST_GEN_1);
+
+    CHECK_CAPS(TEST_findTargetCapSet(testCapSet, CapabilityAtom::hlsl));
+    CHECK_CAPS((int)!TEST_findTargetCapSet(testCapSet, CapabilityAtom::glsl));
+    CHECK_CAPS(TEST_findTargetStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::vertex));
+    CHECK_CAPS(TEST_findTargetStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_6_0));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_5_0));
+
+    // ------------------------------------------------------------
+
+    testCapSet = CapabilitySet(CapabilityName::TEST_GEN_2);
+
+    CHECK_CAPS(TEST_findTargetCapSet(testCapSet, CapabilityAtom::hlsl));
+    CHECK_CAPS((int)!TEST_findTargetCapSet(testCapSet, CapabilityAtom::glsl));
+    CHECK_CAPS(TEST_findTargetStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_6_5));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_5_0));
+
     // ------------------------------------------------------------
+
+    testCapSet = CapabilitySet(CapabilityName::TEST_GEN_3);
+
+    CHECK_CAPS(TEST_findTargetCapSet(testCapSet, CapabilityAtom::glsl));
+    CHECK_CAPS(TEST_findTargetStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment, CapabilityAtom::_GL_NV_shader_texture_footprint));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment, CapabilityAtom::_GL_NV_compute_shader_derivatives));
+
+    // ------------------------------------------------------------
+
+    testCapSet = CapabilitySet(CapabilityName::TEST_GEN_4);
+
+    CHECK_CAPS(TEST_findTargetCapSet(testCapSet, CapabilityAtom::glsl));
+    CHECK_CAPS(TEST_findTargetStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment, CapabilityAtom::_GL_NV_shader_texture_footprint));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::glsl, CapabilityAtom::fragment, CapabilityAtom::_GL_ARB_shader_image_size));
+
+    // ------------------------------------------------------------
+
+    testCapSet = CapabilitySet(CapabilityName::TEST_GEN_5);
+
+    CHECK_CAPS(TEST_findTargetCapSet(testCapSet, CapabilityAtom::hlsl));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_6_5));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_6_4));
+    CHECK_CAPS(TEST_targetCapSetWithSpecificAtomInStage(testCapSet, CapabilityAtom::hlsl, CapabilityAtom::fragment, CapabilityAtom::_sm_6_0));
 }
 
 void TEST_CapabilitySet_join()
@@ -1034,15 +1080,15 @@ void TEST_CapabilitySet()
 /*
 /// Test Capabilities
 
-alias TEST_ADD_1 = _sm_4_1 | _GLSL_130 | spirv_1_1 | metal
-                    ;
-
-alias TEST_ADD_2 = _sm_4_1 | _sm_4_0 + shader_stages_compute_fragment
-                    ;
-
-alias TEST_ADD_3 = _GLSL_130 + shader_stages_compute_fragment_geometry_vertex;
+alias TEST_ADD_1 = _sm_4_1 | _GLSL_130 | spirv_1_1 | metal;
+alias TEST_ADD_2 = _sm_4_1 |& _sm_4_0 + compute_fragment;
+alias TEST_ADD_3 = _GLSL_130 + compute_fragment_geometry_vertex;
 
-//
+alias TEST_GEN_1 = _sm_6_5 + fragment | _sm_6_0 + vertex;
+alias TEST_GEN_2 = _sm_6_5 + fragment;
+alias TEST_GEN_3 = GL_NV_shader_texture_footprint + GL_NV_compute_shader_derivatives + fragment | _GL_NV_shader_texture_footprint + fragment;
+alias TEST_GEN_4 = GL_ARB_shader_image_size |& GL_NV_shader_texture_footprint + fragment;
+alias TEST_GEN_5 = sm_6_0 + compute_fragment| sm_6_5;
 
 alias TEST_JOIN_1A = hlsl;
 alias TEST_JOIN_1B = glsl;
@@ -1065,6 +1111,13 @@ alias TEST_JOIN_3B = _sm_4_1 + fragment
 
 alias TEST_JOIN_4A = _GLSL_140 + _GL_EXT_texture_query_lod;
 alias TEST_JOIN_4B = _GLSL_150 + _GL_EXT_texture_shadow_lod;
+
+// Will cause capability generator failiure
+alias TEST_ERROR_GEN_1 = GL_NV_shader_texture_footprint + GL_NV_compute_shader_derivatives + fragment | _GL_NV_shader_texture_footprint + _GL_NV_shader_atomic_fp16_vector + fragment;
+alias TEST_ERROR_GEN_2 = GL_NV_shader_texture_footprint | GL_NV_ray_tracing_motion_blur;
+alias TEST_ERROR_GEN_3 = GL_ARB_shader_image_size | GL_NV_shader_texture_footprint + fragment;
+alias TEST_ERROR_GEN_4 = _sm_6_5 + fragment + vertex + cpp;
+
 ///
 */
 #undef CHECK_CAPS
diff --git a/source/slang/slang-capability.h b/source/slang/slang-capability.h
index 9e4bdb3a8..8bd03147e 100644
--- a/source/slang/slang-capability.h
+++ b/source/slang/slang-capability.h
@@ -73,6 +73,10 @@ struct CapabilityStageSet
         else
             atomSet->add(setToAdd);
     }
+
+    /// Join `this` with a compatble stage set of `CapabilityTargetSet other`.
+    /// Return false when `other` is fully incompatible.
+    /// incompatability is when `this->stage` is not a supported stage by `other.shaderStageSets`.
     bool tryJoin(const CapabilityTargetSet& other);
 };
 
@@ -85,6 +89,11 @@ struct CapabilityTargetSet
 
     CapabilityStageSets shaderStageSets{};
 
+    /// Join a compatable target set from `this` with `CapabilityTargetSet other`.
+    /// Return false when `other` is fully incompatible.
+    /// incompatability is when one of 2 senarios are true:
+    /// 1. `this->target` is not a supported target by `other.shaderStageSets`
+    /// 2. `this` has completly disjoint shader stages from other.
     bool tryJoin(const CapabilityTargetSets& other);
     void unionWith(const CapabilityTargetSet& other);
 };
@@ -157,6 +166,8 @@ public:
     void addCapability(List<List<CapabilityAtom>>& atomLists);
     /// Calculate a list of "compacted" atoms, which excludes any atoms from the expanded list that are implies by another item in the list.
 
+    /// returns true if 'this' is a better target for 'targetCaps' than 'that'
+    /// isEqual: is `this` and `that` equal
     bool isBetterForTarget(CapabilitySet const& that, CapabilitySet const& targetCaps, bool& isEqual) const;
 
     /// Find any capability sets which are in 'available' but not in 'required'. Return false if this situation occurs. 
diff --git a/source/slang/slang-serialize-ast-type-info.h b/source/slang/slang-serialize-ast-type-info.h
index 1d7628cd4..a8f459247 100644
--- a/source/slang/slang-serialize-ast-type-info.h
+++ b/source/slang/slang-serialize-ast-type-info.h
@@ -101,7 +101,9 @@ struct SerialTypeInfo<CapabilityAtomSet>
         List<CapabilityAtomSet::Element> UIntSetBuffer;
         reader->getArray(src, UIntSetBuffer);
 
-        dst = CapabilityAtomSet(UIntSetBuffer);
+        dst = CapabilityAtomSet();
+        for(Index i = 0; i < UIntSetBuffer.getCount(); i++)
+            dst.addRawElement(UIntSetBuffer[i], i);
     }
 };
 
@@ -244,11 +246,9 @@ struct SerialTypeInfo<CapabilitySet>
         auto& targetSets = dst.getCapabilityTargetSets();
         targetSets.clear();
         targetSets.reserve(items.getCount());
-        Index iter = 0;
         for (auto& i : items)
         {
             targetSets[i.target] = i;
-            iter++;
         }
     }
 };