Merge branch 'dtm'

# Conflicts:
#	examples_tests

Author: Erfan-Ahmadi

include/nbl/builtin/hlsl/algorithm.hlsl

@@ -5,6 +5,7 @@
 #define _NBL_BUILTIN_HLSL_ALGORITHM_INCLUDED_
 
 #include "nbl/builtin/hlsl/functional.hlsl"
+#include "nbl/builtin/hlsl/math/intutil.hlsl"
 
 namespace nbl
 {
@@ -91,16 +92,10 @@ NBL_CONSTEXPR_INLINE_FUNC void swap(NBL_REF_ARG(T) lhs, NBL_REF_ARG(T) rhs)
 namespace impl
 {
 
-// TODO: move this to some other header
-bool isNPoT(const uint x)
-{
-    return x&(x-1);
-}
-
 template<class Accessor, class Comparator>
 struct bound_t
 {
-    static bound_t<Accessor,Comparator> setup(uint begin, const uint end, const typename Accessor::value_type _value, const Comparator _comp)
+    static bound_t<Accessor,Comparator> setup(uint32_t begin, const uint32_t end, const typename Accessor::value_type _value, const Comparator _comp)
     {
         bound_t<Accessor,Comparator> retval;
         retval.comp = _comp;
@@ -111,12 +106,12 @@ struct bound_t
     }
 
 
-    uint operator()(NBL_REF_ARG(Accessor) accessor)
+    uint32_t operator()(NBL_REF_ARG(Accessor) accessor)
     {
         if (isNPoT(len))
         {
-            const uint newLen = 0x1u<<firstbithigh(len);
-            const uint testPoint = it+(len-newLen);
+            const uint32_t newLen = 0x1u<<nbl::hlsl::findMSB(len);
+            const uint32_t testPoint = it+(len-newLen);
             len = newLen;
             comp_step(accessor,testPoint);
         }
@@ -133,24 +128,24 @@ struct bound_t
     void iteration(NBL_REF_ARG(Accessor) accessor)
     {
         len >>= 1;
-        const uint mid = it+len;
+        const uint32_t mid = it+len;
         comp_step(accessor,mid);
     }
 
-    void comp_step(NBL_REF_ARG(Accessor) accessor, const uint testPoint, const uint rightBegin)
+    void comp_step(NBL_REF_ARG(Accessor) accessor, const uint32_t testPoint, const uint32_t rightBegin)
     {
         if (comp(accessor[testPoint],value))
             it = rightBegin;
     }
-    void comp_step(NBL_REF_ARG(Accessor) accessor, const uint testPoint)
+    void comp_step(NBL_REF_ARG(Accessor) accessor, const uint32_t testPoint)
     {
         comp_step(accessor,testPoint,testPoint);
     }
 
     Comparator comp;
     typename Accessor::value_type value;
-    uint it;
-    uint len;
+    uint32_t it;
+    uint32_t len;
 };
 
 template<class Accessor, class Comparator>
@@ -168,14 +163,14 @@ struct lower_to_upper_comparator_transform_t
 
 
 template<class Accessor, class Comparator>
-uint lower_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint end, const typename Accessor::value_type value, const Comparator comp)
+uint32_t lower_bound(NBL_REF_ARG(Accessor) accessor, const uint32_t begin, const uint32_t end, const typename Accessor::value_type value, const Comparator comp)
 {
     impl::bound_t<Accessor,Comparator> implementation = impl::bound_t<Accessor,Comparator>::setup(begin,end,value,comp);
     return implementation(accessor);
 }
 
 template<class Accessor, class Comparator>
-uint upper_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint end, const typename Accessor::value_type value, const Comparator comp)
+uint32_t upper_bound(NBL_REF_ARG(Accessor) accessor, const uint32_t begin, const uint32_t end, const typename Accessor::value_type value, const Comparator comp)
 {
     //using TransformedComparator = impl::lower_to_upper_comparator_transform_t<Accessor,Comparator>;
     //TransformedComparator transformedComparator;
@@ -191,7 +186,7 @@ namespace impl
 
 // extra indirection due to https://github.com/microsoft/DirectXShaderCompiler/issues/4771
 template<class Accessor, typename T>
-uint lower_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint end, const T value)
+uint32_t lower_bound(NBL_REF_ARG(Accessor) accessor, const uint32_t begin, const uint32_t end, const T value)
 {
     //using Comparator = nbl::hlsl::less<T>;
     //Comparator comp;
@@ -200,7 +195,7 @@ uint lower_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint en
     return nbl::hlsl::lower_bound<Accessor, nbl::hlsl::less<T> >(accessor,begin,end,value,comp);
 }
 template<class Accessor, typename T>
-uint upper_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint end, const T value)
+uint32_t upper_bound(NBL_REF_ARG(Accessor) accessor, const uint32_t begin, const uint32_t end, const T value)
 {
     //using Comparator = nbl::hlsl::less<T>;
     //Comparator comp;
@@ -212,12 +207,12 @@ uint upper_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint en
 }
 
 template<class Accessor>
-uint lower_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint end, const typename Accessor::value_type value)
+uint32_t lower_bound(NBL_REF_ARG(Accessor) accessor, const uint32_t begin, const uint32_t end, const typename Accessor::value_type value)
 {
     return impl::lower_bound<Accessor,typename Accessor::value_type>(accessor,begin,end,value);
 }
 template<class Accessor>
-uint upper_bound(NBL_REF_ARG(Accessor) accessor, const uint begin, const uint end, const typename Accessor::value_type value)
+uint32_t upper_bound(NBL_REF_ARG(Accessor) accessor, const uint32_t begin, const uint32_t end, const typename Accessor::value_type value)
 {
     return impl::upper_bound<Accessor,typename Accessor::value_type>(accessor,begin,end,value);
 }

include/nbl/builtin/hlsl/emulated/float64_t.hlsl

@@ -154,14 +154,20 @@ namespace hlsl
                     if ((data | ieee754::traits<float64_t>::signMask) == (rhs.data | ieee754::traits<float64_t>::signMask))
                         return _static_cast<this_t>(0ull);
 
-                    uint64_t rhsNormMantissaHigh = shiftAmount >= 64 ? 0ull : rhsNormMantissa >> shiftAmount;
+                    uint64_t rhsNormMantissaHigh = shiftAmount >= 64u ? 0ull : rhsNormMantissa >> shiftAmount;
                     uint64_t rhsNormMantissaLow = 0ull;
                     if (shiftAmount < 128)
                     {
                         if (shiftAmount >= 64)
+                        {
                             rhsNormMantissaLow = rhsNormMantissa >> (shiftAmount - 64);
+                        }
                         else
-                            rhsNormMantissaLow = rhsNormMantissa << (64 - shiftAmount);
+                        {
+                            const uint32_t lowMantissaShiftAmount = 64 - shiftAmount;
+                            if(lowMantissaShiftAmount < 64)
+                                rhsNormMantissaLow = rhsNormMantissa << lowMantissaShiftAmount;
+                        }
                     }
 
                     const int64_t mantissaDiff = int64_t(lhsNormMantissa) - int64_t(rhsNormMantissaHigh);
@@ -179,7 +185,7 @@ namespace hlsl
                 }
                 else
                 {
-                    rhsNormMantissa >>= shiftAmount;
+                    rhsNormMantissa = shiftAmount > 63 ? 0ull : rhsNormMantissa >> shiftAmount;
                     resultMantissa = lhsNormMantissa + rhsNormMantissa;
 
                     if (resultMantissa & 1ull << 53)

src/nbl/asset/utils/CHLSLCompiler.cpp

@@ -382,7 +382,7 @@ std::string CHLSLCompiler::preprocessShader(std::string&& code, IShader::E_SHADE
     );
 
     // for debugging cause MSVC doesn't like to show more than 21k LoC in TextVisualizer
-    if constexpr (true)
+    if constexpr (false)
     {
         system::ISystem::future_t<core::smart_refctd_ptr<system::IFile>> future;
         m_system->createFile(future,system::path(preprocessOptions.sourceIdentifier).parent_path()/"preprocessed.hlsl",system::IFileBase::ECF_WRITE);

src/nbl/video/CVulkanConnection.cpp

@@ -286,7 +286,7 @@ core::smart_refctd_ptr<CVulkanConnection> CVulkanConnection::create(core::smart_
     }
 
     VkDebugUtilsMessengerEXT vk_debugMessenger = VK_NULL_HANDLE;
-    if (debugCallback)
+    if (enabledFeatures.debugUtils && debugCallback)
     {
         if (vkCreateDebugUtilsMessengerEXT(vk_instance,&debugMessengerCreateInfo,nullptr,&vk_debugMessenger) != VK_SUCCESS)
         {