Add RoundUpToMultiple and DivideAndRoundUp functions to MathUtils.h (#6989)

* Add RoundUpToMultiple and DivideAndRoundUp functions to MathUtils.h

Signed-off-by: Tommy Walton <waltont@amazon.com>

* Rename DivideByMultiple2 back to DivideByMultiple, now that I've confirmed it's not in use in the codebase. RHI::DivideByMultiple can be fully deprecated in favor of AZ::DivideAndRoundUp at a later date, once the deprecation strategy has been finalized.

Signed-off-by: Tommy Walton <waltont@amazon.com>

* Update based on PR feedback

Signed-off-by: Tommy Walton <waltont@amazon.com>

* Switched from std::numeric_limits to AZStd::numeric_limits and updated the header to indicate it works for non-power of two alignments, but that SizeAlignUp is more efficient if the alignment is a power of 2

Signed-off-by: Tommy Walton <waltont@amazon.com>

* Added missing arguments to the assert, and a missing namespace and include that failed to compile on non-unity builds

Signed-off-by: Tommy Walton <waltont@amazon.com>

Code/Framework/AzCore/AzCore/Math/MathUtils.h

@@ -9,6 +9,7 @@
 #pragma once
 
 #include <AzCore/base.h>
+#include <AzCore/std/limits.h>
 #include <AzCore/std/math.h>
 #include <AzCore/std/typetraits/conditional.h>
 #include <AzCore/std/typetraits/is_integral.h>
@@ -20,6 +21,7 @@
 #include <limits>
 #include <math.h>
 #include <utility>
+#include <inttypes.h>
 
 // We have a separate inline define for math functions.
 // The performance of these functions is very sensitive to inlining, and some compilers don't deal well with this.
@@ -256,13 +258,13 @@ namespace AZ
     struct ClampedIntegralLimits
     {
         //! If SourceType and ClampType are different, returns the greater value of 
-        //! std::numeric_limits<SourceType>::lowest() and std::numeric_limits<ClampType>::lowest(),
+        //! AZStd::numeric_limits<SourceType>::lowest() and AZStd::numeric_limits<ClampType>::lowest(),
-        //! otherwise returns std::numeric_limits<SourceType>::lowest().
+        //! otherwise returns AZStd::numeric_limits<SourceType>::lowest().
         static constexpr SourceType Min();
 
         //! If SourceType and ClampType are different, returns the lesser value of 
-        //! std::numeric_limits<SourceType>::max() and std::numeric_limits<ClampType>::max(),
+        //! AZStd::numeric_limits<SourceType>::max() and AZStd::numeric_limits<ClampType>::max(),
-        //! otherwise returns std::numeric_limits<SourceType>::max().
+        //! otherwise returns AZStd::numeric_limits<SourceType>::max().
         static constexpr SourceType Max();
 
         //! Safely clamps a value of type ValueType to the [Min(), Max()] range as determined by the
@@ -375,12 +377,12 @@ namespace AZ
     //! Returns a value t where Lerp(a, b, t) == value (or 0 if a == b).
     inline float LerpInverse(float a, float b, float value)
     {
-        return IsClose(a, b, std::numeric_limits<float>::epsilon()) ? 0.0f : (value - a) / (b - a);
+        return IsClose(a, b, AZStd::numeric_limits<float>::epsilon()) ? 0.0f : (value - a) / (b - a);
     }
 
     inline double LerpInverse(double a, double b, double value)
     {
-        return IsClose(a, b, std::numeric_limits<double>::epsilon()) ? 0.0 : (value - a) / (b - a);
+        return IsClose(a, b, AZStd::numeric_limits<double>::epsilon()) ? 0.0 : (value - a) / (b - a);
     }
 
     //! Returns true if the number provided is even.
@@ -431,19 +433,19 @@ namespace AZ
 
     AZ_MATH_INLINE float GetFloatQNaN()
     {
-        return std::numeric_limits<float>::quiet_NaN();
+        return AZStd::numeric_limits<float>::quiet_NaN();
     }
 
     //! IsCloseMag(x, y, epsilon) returns true if y and x are sufficiently close, taking magnitude of x and y into account in the epsilon
     template<typename T>
-    AZ_MATH_INLINE bool IsCloseMag(T x, T y, T epsilonValue = std::numeric_limits<T>::epsilon())
+    AZ_MATH_INLINE bool IsCloseMag(T x, T y, T epsilonValue = AZStd::numeric_limits<T>::epsilon())
     {
         return (AZStd::abs(x - y) <= epsilonValue * GetMax<T>(GetMax<T>(T(1.0), AZStd::abs(x)), AZStd::abs(y)));
     }
 
     //! ClampIfCloseMag(x, y, epsilon) returns y when x and y are within epsilon of each other (taking magnitude into account).  Otherwise returns x.
     template<typename T>
-    AZ_MATH_INLINE T ClampIfCloseMag(T x, T y, T epsilonValue = std::numeric_limits<T>::epsilon())
+    AZ_MATH_INLINE T ClampIfCloseMag(T x, T y, T epsilonValue = AZStd::numeric_limits<T>::epsilon())
     {
         return IsCloseMag<T>(x, y, epsilonValue) ? y : x;
     }
@@ -461,6 +463,44 @@ namespace AZ
         return (azisfinite(x) != 0);
     }
 
+    //! Returns the value divided by alignment, where the result is rounded up if the remainder is non-zero.
+    //! Example: alignment: 4
+    //! Value:  0 1 2 3 4 5 6 7 8
+    //! Result: 0 1 1 1 1 2 2 2 2
+    constexpr uint32_t DivideAndRoundUp(uint32_t value, uint32_t alignment)
+    {
+        AZ_Assert(alignment != 0, "0 is an invalid multiple to round to.");
+        AZ_Assert(
+            AZStd::numeric_limits<uint32_t>::max() - value >= alignment,
+            "value '%" PRIu32 "' and alignment '%" PRIu32 "' will overflow when added together during DivideAndRoundUp.", value, alignment);
+        return (value + alignment - 1) / alignment;
+    }
+
+    constexpr uint64_t DivideAndRoundUp(uint64_t value, uint64_t alignment)
+    {
+        AZ_Assert(alignment != 0, "0 is an invalid multiple to round to.");
+        AZ_Assert(
+            AZStd::numeric_limits<uint64_t>::max() - value >= alignment,
+            "value '%" PRIu64 "' and alignment '%" PRIu64 "' will overflow when added together during DivideAndRoundUp.", value, alignment);
+        return (value + alignment - 1) / alignment;
+    }
+    
+    //! Returns the value rounded up to a multiple of alignment.
+    //! This function will work for non power of two alignments.
+    //! If your alignment is guaranteed to be a power of two, SizeAlignUp in base.h is a more efficient implementation.
+    //! Example: roundTo: 4
+    //! Value:  0 1 2 3 4 5 6 7 8
+    //! Result: 0 4 4 4 4 8 8 8 8
+    constexpr uint32_t RoundUpToMultiple(uint32_t value, uint32_t alignment)
+    {
+        return DivideAndRoundUp(value, alignment) * alignment;
+    }
+
+    constexpr uint64_t RoundUpToMultiple(uint64_t value, uint64_t alignment)
+    {
+        return DivideAndRoundUp(value, alignment) * alignment;
+    }
+
     //! Returns the maximum value for SourceType as constrained by the numerical range of ClampType.
     template <typename SourceType, typename ClampType>    
     constexpr SourceType ClampedIntegralLimits<SourceType, ClampType>::Min()
@@ -474,8 +514,8 @@ namespace AZ
         {
             // Both SourceType and ClampType are signed, take the greater of the lower limits of each type
             return sizeof(SourceType) < sizeof(ClampType) ? 
-                (std::numeric_limits<SourceType>::lowest)() : 
+                (AZStd::numeric_limits<SourceType>::lowest)() : 
-                static_cast<SourceType>((std::numeric_limits<ClampType>::lowest)());
+                static_cast<SourceType>((AZStd::numeric_limits<ClampType>::lowest)());
         }
     }
 
@@ -486,12 +526,12 @@ namespace AZ
         if constexpr (sizeof(SourceType) < sizeof(ClampType))
         {
             // If SourceType is narrower than ClampType, the upper limit will be SourceType's
-            return (std::numeric_limits<SourceType>::max)();
+            return (AZStd::numeric_limits<SourceType>::max)();
         }
         else if constexpr (sizeof(SourceType) > sizeof(ClampType))
         {
             // If SourceType is wider than ClampType, the upper limit will be ClampType's
-            return static_cast<SourceType>((std::numeric_limits<ClampType>::max)());
+            return static_cast<SourceType>((AZStd::numeric_limits<ClampType>::max)());
         }
         else
         {
@@ -499,13 +539,13 @@ namespace AZ
             {
                 // SourceType and ClampType are the same width, ClampType is signed
                 // so our upper limit will be ClampType
-                return static_cast<SourceType>((std::numeric_limits<ClampType>::max)());
+                return static_cast<SourceType>((AZStd::numeric_limits<ClampType>::max)());
             }       
             else
             {
                 // SourceType and ClampType are the same width, ClampType is unsigned
                 // then our upper limit will be SourceType
-                return (std::numeric_limits<SourceType>::max)();
+                return (AZStd::numeric_limits<SourceType>::max)();
             }
         }
     }
@@ -588,7 +628,7 @@ namespace AZ
             // LeftTypeSize <= RightTypeSize
             // LeftType is signed
             // RightType is unsigned
-            RightType max = static_cast<RightType>((std::numeric_limits<LeftType>::max)());
+            RightType max = static_cast<RightType>((AZStd::numeric_limits<LeftType>::max)());
 
             if (rhs > max)
             {
@@ -604,7 +644,7 @@ namespace AZ
             // LeftType < RightType
             // LeftType is unsigned
             // RightType is signed
-            RightType max = static_cast<RightType>((std::numeric_limits<LeftType>::max)());
+            RightType max = static_cast<RightType>((AZStd::numeric_limits<LeftType>::max)());
 
             if (rhs < 0)
             {

Code/Framework/AzCore/Tests/Math/MathUtilsTests.cpp

@@ -37,8 +37,8 @@ namespace UnitTest
         // min/max need to be substantially different to return a useful t value
 
         // Float
-        const float epsilonF = std::numeric_limits<float>::epsilon();
+        const float epsilonF = AZStd::numeric_limits<float>::epsilon();
-        const float doesntMatterF = std::numeric_limits<float>::signaling_NaN();
+        const float doesntMatterF = AZStd::numeric_limits<float>::signaling_NaN();
         float lowerF = 2.3f, upperF = 2.3f;
         EXPECT_EQ(0.0f, AZ::LerpInverse(lowerF, upperF, doesntMatterF));
         EXPECT_EQ(0.0f, AZ::LerpInverse(0.0f, 0.5f * epsilonF, doesntMatterF));
@@ -48,8 +48,8 @@ namespace UnitTest
         EXPECT_NEAR(1.0f, AZ::LerpInverse(1.0f, 1.0f + 5.0f * epsilonF, 1.0f + 5.0f * epsilonF), epsilonF);
 
         // Double
-        const double epsilonD = std::numeric_limits<double>::epsilon();
+        const double epsilonD = AZStd::numeric_limits<double>::epsilon();
-        const double doesntMatterD = std::numeric_limits<double>::signaling_NaN();
+        const double doesntMatterD = AZStd::numeric_limits<double>::signaling_NaN();
         double lowerD = 2.3, upperD = 2.3;
         EXPECT_EQ(0.0, AZ::LerpInverse(lowerD, upperD, doesntMatterD));
         EXPECT_EQ(0.0, AZ::LerpInverse(0.0, 0.5 * epsilonD, doesntMatterD));
@@ -58,4 +58,128 @@ namespace UnitTest
         EXPECT_NEAR(0.6, AZ::LerpInverse(1.0, 1.0 + 5.0 * epsilonD, 1.0 + 3.0 * epsilonD), epsilonD);
         EXPECT_NEAR(1.0, AZ::LerpInverse(1.0, 1.0 + 5.0 * epsilonD, 1.0 + 5.0 * epsilonD), epsilonD);
     }
+
+    template <typename T>
+    void TestRoundUpToMultipleIsCorrect()
+    {
+        // Example: alignment: 4
+        // inputValue:     0 1 2 3 4 5 6 7 8 ...
+        // expectedOutput: 0 4 4 4 4 8 8 8 8 ...
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(0) , static_cast<T>(1)) , 0);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(1) , static_cast<T>(1)) , 1);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(2) , static_cast<T>(1)) , 2);
+
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(0) , static_cast<T>(2)) , 0);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(1) , static_cast<T>(2)) , 2);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(2) , static_cast<T>(2)) , 2);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(3) , static_cast<T>(2)) , 4);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(4) , static_cast<T>(2)) , 4);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(5) , static_cast<T>(2)) , 6);
+
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(0) , static_cast<T>(8)) , 0);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(1) , static_cast<T>(8)) , 8);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(7) , static_cast<T>(8)) , 8);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(8) , static_cast<T>(8)) , 8);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(9) , static_cast<T>(8)) , 16);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(15), static_cast<T>(8)) , 16);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(16), static_cast<T>(8)) , 16);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(17), static_cast<T>(8)) , 24);
+
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(0) , static_cast<T>(13)), 0);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(1) , static_cast<T>(13)), 13);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(9) , static_cast<T>(13)), 13);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(12), static_cast<T>(13)), 13);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(13), static_cast<T>(13)), 13);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(14), static_cast<T>(13)), 26);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(25), static_cast<T>(13)), 26);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(26), static_cast<T>(13)), 26);
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(27), static_cast<T>(13)), 39);
+
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(0), AZStd::numeric_limits<T>::max()), 0);
+
+        T aVeryLargeNumberThatStillWontOverflow = AZStd::numeric_limits<T>::max() - 4;
+        EXPECT_EQ(RoundUpToMultiple(static_cast<T>(1), aVeryLargeNumberThatStillWontOverflow), aVeryLargeNumberThatStillWontOverflow);
+        EXPECT_EQ(RoundUpToMultiple(aVeryLargeNumberThatStillWontOverflow, static_cast<T>(1)), aVeryLargeNumberThatStillWontOverflow);
+    }
+
+    TEST(RoundUpToMultipleTest, RoundUpToMultipleUInt32_ValidInput_IsCorrect)
+    {
+        TestRoundUpToMultipleIsCorrect<uint32_t>();
+    }
+
+    TEST(RoundUpToMultipleTest, RoundUpToMultipleUInt64_ValidInput_IsCorrect)
+    {
+        TestRoundUpToMultipleIsCorrect<uint64_t>();
+    }
+
+    template<typename T>
+    void TestDivideAndRoundUpIsCorrect()
+    {
+        //! Example: alignment: 3
+        //! Value:  0 1 2 3 4 5 6 7 8
+        //! Result: 0 1 1 1 2 2 2 3 3
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(0), static_cast<T>(3)), 0);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(1), static_cast<T>(3)), 1);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(2), static_cast<T>(3)), 1);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(3), static_cast<T>(3)), 1);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(4), static_cast<T>(3)), 2);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(5), static_cast<T>(3)), 2);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(6), static_cast<T>(3)), 2);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(7), static_cast<T>(3)), 3);
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(8), static_cast<T>(3)), 3);
+
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(0), AZStd::numeric_limits<T>::max()), 0);
+
+        T aVeryLargeNumberThatStillWontOverflow = AZStd::numeric_limits<T>::max() - 4;
+        EXPECT_EQ(DivideAndRoundUp(static_cast<T>(1), aVeryLargeNumberThatStillWontOverflow), static_cast<T>(1));
+        EXPECT_EQ(DivideAndRoundUp(aVeryLargeNumberThatStillWontOverflow, static_cast<T>(1)), aVeryLargeNumberThatStillWontOverflow);
+
+    }
+
+    TEST(DivideAndRoundUpTest, DivideAndRoundUpUInt32_ValidInput_IsCorrect)
+    {
+        TestDivideAndRoundUpIsCorrect<uint32_t>();
+    }
+
+    TEST(DivideAndRoundUpTest, DivideAndRoundUpUInt64_ValidInput_IsCorrect)
+    {
+        TestDivideAndRoundUpIsCorrect<uint64_t>();
+    }
+
+    
+    class RoundUpInvalidInputTestsFixture : public ScopedAllocatorSetupFixture
+    {
+    };
+
+    TEST_F(RoundUpInvalidInputTestsFixture, DividAndRoundUp_AlignmentZeroUint32_Assert)
+    {
+        AZ_TEST_START_TRACE_SUPPRESSION;
+        DivideAndRoundUp(static_cast<uint32_t>(0), static_cast<uint32_t>(0));
+        AZ_TEST_STOP_TRACE_SUPPRESSION(1);
+    }
+
+    TEST_F(RoundUpInvalidInputTestsFixture, DividAndRoundUp_AlignmentZeroUint64_Assert)
+    {
+        AZ_TEST_START_TRACE_SUPPRESSION;
+        DivideAndRoundUp(static_cast<uint64_t>(0), static_cast<uint64_t>(0));
+        AZ_TEST_STOP_TRACE_SUPPRESSION(1);
+    }
+
+    TEST_F(RoundUpInvalidInputTestsFixture, DividAndRoundUp_OverflowUint32_Assert)
+    {
+        AZ_TEST_START_TRACE_SUPPRESSION;
+        DivideAndRoundUp(
+            static_cast<uint32_t>((AZStd::numeric_limits<uint32_t>::max() / 2) + 1),
+            static_cast<uint32_t>((AZStd::numeric_limits<uint32_t>::max() / 2) + 1));
+        AZ_TEST_STOP_TRACE_SUPPRESSION(1);
+    }
+
+    TEST_F(RoundUpInvalidInputTestsFixture, DividAndRoundUp_OverflowUint64_Assert)
+    {
+        AZ_TEST_START_TRACE_SUPPRESSION;
+        DivideAndRoundUp(
+            static_cast<uint64_t>((AZStd::numeric_limits<uint64_t>::max() / 2) + 1),
+            static_cast<uint64_t>((AZStd::numeric_limits<uint64_t>::max() / 2) + 1));
+        AZ_TEST_STOP_TRACE_SUPPRESSION(1);
+    }
 }

Code/Framework/AzNetworking/AzNetworking/UdpTransport/UdpNetworkInterface.cpp

@@ -17,6 +17,7 @@
 #include <AzNetworking/Utilities/NetworkCommon.h>
 #include <AzCore/Console/IConsole.h>
 #include <AzCore/Console/ILogger.h>
+#include <AzCore/Math/MathUtils.h>
 
 namespace AzNetworking
 {
@@ -539,7 +540,7 @@ namespace AzNetworking
             // Each fragmented packet we send adds an extra fragmented packet header, need to deduct that from our chunk size, otherwise we infinitely loop
             // SSL encryption can also inflate our payload so we pre-emptively deduct an estimated tax
             const uint32_t chunkSize = connection.GetConnectionMtu() - net_FragmentedHeaderOverhead - net_SslInflationOverhead;
-            const uint32_t numChunks = (packetSize + chunkSize - 1) / chunkSize; // We want to round up on the remainder
+            const uint32_t numChunks = AZ::DivideAndRoundUp(packetSize, chunkSize); // We want to round up on the remainder
             const uint8_t* chunkStart = packetData;
             const SequenceId fragmentedSequence = connection.m_fragmentQueue.GetNextFragmentedSequenceId();
             uint32_t bytesRemaining = packetSize;

Code/Framework/GridMate/GridMate/Carrier/SocketDriver.cpp

@@ -20,6 +20,7 @@
 #include <GridMate/Containers/unordered_set.h>
 #include <GridMate/Carrier/DriverEvents.h>
 
+#include <AzCore/Math/MathUtils.h>
 #include <AzCore/std/chrono/types.h>
 #include <AzCore/std/string/conversions.h>
 #include <AzCore/std/string/memorytoascii.h>
@@ -1951,7 +1952,7 @@ namespace GridMate
     char *SocketDriverCommon::RIOPlatformSocketDriver::AllocRIOBuffer(AZ::u64 bufferSize, AZ::u64 numBuffers, AZ::u64* amountAllocated /*=nullptr*/)
     {
         // calculate how much memory we are really asking for, and this must be page aligned.
-        AZ::u64 totalBufferSize = RoundUp(bufferSize * numBuffers, m_pageSize);
+        AZ::u64 totalBufferSize = AZ::RoundUpToMultiple(bufferSize * numBuffers, m_pageSize);
 
         if (amountAllocated != nullptr)
         {

Code/Framework/GridMate/GridMate/Carrier/SocketDriver.h

@@ -221,18 +221,6 @@ namespace GridMate
             void StopWaitForData() override;
 
         private:
-            AZ::u64 RoundUpAndDivide(AZ::u64 Value, AZ::u64 RoundTo) const
-            {
-                return ((Value + RoundTo - 1) / RoundTo);
-            }
-            AZ::u64 RoundUp(AZ::u64 Value, AZ::u64 RoundTo) const
-            {
-                // rounds value up to multiple of RoundTo
-                // Example: RoundTo: 4
-                // Value:  0 1 2 3 4 5 6 7 8
-                // Result: 0 4 4 4 4 8 8 8 8
-                return RoundUpAndDivide(Value, RoundTo) * RoundTo;
-            }
             char *AllocRIOBuffer(AZ::u64 bufferSize, AZ::u64 numBuffers, AZ::u64* amountAllocated=nullptr);
             bool FreeRIOBuffer(char *buffer);
 

Gems/Atom/Asset/ImageProcessingAtom/Code/Source/Compressors/ASTCCompressor.cpp

@@ -179,7 +179,7 @@ namespace ImageProcessingAtom
 
         // Create a context based on the configuration
         astcenc_context* context;
-        AZ::u32 blockCount = ((srcImage->GetWidth(0)+ dstFormatInfo->blockWidth-1)/dstFormatInfo->blockWidth) * ((srcImage->GetHeight(0) + dstFormatInfo->blockHeight-1)/dstFormatInfo->blockHeight);
+        AZ::u32 blockCount = AZ::DivideAndRoundUp(srcImage->GetWidth(0), dstFormatInfo->blockWidth) * AZ::DivideAndRoundUp(srcImage->GetHeight(0), dstFormatInfo->blockHeight);
         AZ::u32 threadCount = AZStd::min(AZStd::thread::hardware_concurrency()/2, blockCount);
         status = astcenc_context_alloc(&config, threadCount, &context);
         AZ_Assert( status == ASTCENC_SUCCESS, "ERROR: Codec context alloc failed: %s\n", astcenc_get_error_string(status));

Gems/Atom/Asset/ImageProcessingAtom/Code/Source/Processing/PixelFormatInfo.cpp

@@ -10,6 +10,8 @@
 #include <Processing/PixelFormatInfo.h>
 #include <Processing/DDSHeader.h>
 
+#include <AzCore/Math/MathUtils.h>
+
 namespace ImageProcessingAtom
 {
     CPixelFormats* CPixelFormats::s_instance = nullptr;
@@ -370,11 +372,11 @@ namespace ImageProcessingAtom
         {
             if (outWidth % pFormatInfo->blockWidth != 0)
             {
-                outWidth = ((outWidth + pFormatInfo->blockWidth - 1) / pFormatInfo->blockWidth) * pFormatInfo->blockWidth;
+                outWidth = AZ::RoundUpToMultiple(outWidth, pFormatInfo->blockWidth);
             }
             if (outHeight % pFormatInfo->blockHeight != 0)
             {
-                outHeight = ((outHeight + pFormatInfo->blockHeight - 1) / pFormatInfo->blockHeight) * pFormatInfo->blockHeight;
+                outHeight = AZ::RoundUpToMultiple(outHeight, pFormatInfo->blockHeight);
             }
         }
     }
@@ -390,10 +392,11 @@ namespace ImageProcessingAtom
             return 0;
         }
 
-        // get number of blocks (ceiling round up for block count) and multiply with bits per block. Divided by 8 to get
+        // get number of blocks and multiply with bits per block. Divided by 8 to get
         // final byte size
-        return (((imageWidth + pFormatInfo->blockWidth - 1) / pFormatInfo->blockWidth) *
+        return (AZ::DivideAndRoundUp(imageWidth, pFormatInfo->blockWidth) *
-                ((imageHeight + pFormatInfo->blockHeight - 1) / pFormatInfo->blockHeight) * pFormatInfo->bitsPerBlock) / 8;
+                AZ::DivideAndRoundUp(imageHeight, pFormatInfo->blockHeight) *
+                pFormatInfo->bitsPerBlock) / 8;
     }
 
     bool CPixelFormats::IsFormatSingleChannel(EPixelFormat fmt)

Gems/Atom/RHI/Code/Include/Atom/RHI.Reflect/Bits.h

@@ -8,6 +8,7 @@
 #pragma once
 
 #include <AzCore/base.h>
+#include <AzCore/Math/MathUtils.h>
 
 #ifndef AZ_BIT
 #define AZ_BIT(x) (1u << x)
@@ -237,6 +238,7 @@ namespace AZ
         }
 
         /**
+        * O3DE_DEPRECATION_NOTICE(GHI-7407)
         * Returns the value divided by alignment, where the result is rounded up if the remainder is non-zero.
         */
         template <typename T> inline T DivideByMultiple(T value, size_t alignment)

Gems/Atom/RHI/Code/Include/Atom/RHI/ConstantsData.h

@@ -228,7 +228,7 @@ namespace AZ
         {
             AZStd::span<const uint8_t> constantBytes = GetConstantRaw(inputIndex);
             const size_t elementSize = sizeof(T);
-            const size_t elementCount = DivideByMultiple(constantBytes.size(), elementSize);
+            const size_t elementCount = AZ::DivideAndRoundUp(constantBytes.size(), elementSize);
             const size_t sizeInBytes = elementCount * elementSize;
             if (ValidateConstantAccess(inputIndex, ValidateConstantAccessExpect::Complete, 0, sizeInBytes))
             {

Gems/Atom/RHI/Code/Include/Atom/RHI/DispatchItem.h

@@ -41,17 +41,17 @@ namespace AZ
 
             uint16_t GetNumberOfGroupsX() const
             {
-                return aznumeric_cast<uint16_t>((m_totalNumberOfThreadsX + m_threadsPerGroupX - 1) / m_threadsPerGroupX);
+                return aznumeric_cast<uint16_t>(DivideAndRoundUp(m_totalNumberOfThreadsX, aznumeric_caster(m_threadsPerGroupX)));
             }
 
             uint16_t GetNumberOfGroupsY() const
             {
-                return aznumeric_cast<uint16_t>((m_totalNumberOfThreadsY + m_threadsPerGroupY - 1) / m_threadsPerGroupY);
+                return aznumeric_cast<uint16_t>(DivideAndRoundUp(m_totalNumberOfThreadsY, aznumeric_caster(m_threadsPerGroupY)));
             }
 
             uint16_t GetNumberOfGroupsZ() const
             {
-                return aznumeric_cast<uint16_t>((m_totalNumberOfThreadsZ + m_threadsPerGroupZ - 1) / m_threadsPerGroupZ);
+                return aznumeric_cast<uint16_t>(DivideAndRoundUp(m_totalNumberOfThreadsZ, aznumeric_caster(m_threadsPerGroupZ)));
             }
 
             // Different platforms require number of groups or number of threads or both in their Dispatch() call

Gems/Atom/RHI/Code/Source/RHI.Reflect/ImageSubresource.cpp

@@ -358,8 +358,7 @@ namespace AZ
             }
             else
             {
-                const uint32_t bitsPerPixel = RHI::GetFormatSize(imageFormat) * 8;
+                subresourceLayout.m_bytesPerRow = imageSize.m_width * RHI::GetFormatSize(imageFormat);
-                subresourceLayout.m_bytesPerRow = (imageSize.m_width * bitsPerPixel + 7) / 8; // round up to nearest byte
                 subresourceLayout.m_rowCount = imageSize.m_height;
                 subresourceLayout.m_size.m_width = imageSize.m_width;
                 subresourceLayout.m_size.m_height = imageSize.m_height;

Gems/Atom/RHI/Code/Source/RHI/DrawList.cpp

@@ -20,7 +20,7 @@ namespace AZ
                 return DrawListView{};
             }
 
-            const size_t itemsPerPartition = DivideByMultiple(drawList.size(), partitionCount);
+            const size_t itemsPerPartition = AZ::DivideAndRoundUp(drawList.size(), partitionCount);
             const size_t itemOffset = partitionIndex * itemsPerPartition;
             const size_t itemCount = AZStd::min(drawList.size() - itemOffset, itemsPerPartition);
             return DrawListView(&drawList[itemOffset], itemCount);

Gems/Atom/RHI/Code/Source/RHI/FrameScheduler.cpp

@@ -281,7 +281,7 @@ namespace AZ
                     {
                         srgPool->CompileGroupsBegin();
                         const uint32_t compilesInPool = srgPool->GetGroupsToCompileCount();
-                        const uint32_t jobCount = DivideByMultiple(compilesInPool, compilesPerJob);
+                        const uint32_t jobCount = AZ::DivideAndRoundUp(compilesInPool, compilesPerJob);
                         AZ::TaskDescriptor srgCompileDesc{"SrgCompile", "Graphics"};
                         AZ::TaskDescriptor srgCompileEndDesc{"SrgCompileEnd", "Graphics"};
 
@@ -332,7 +332,7 @@ namespace AZ
                     const auto compileIntervalsFunction = [compilesPerJob, &jobCompletion](ShaderResourceGroupPool* srgPool)
                     {
                         const uint32_t compilesInPool = srgPool->GetGroupsToCompileCount();
-                        const uint32_t jobCount = DivideByMultiple(compilesInPool, compilesPerJob);
+                        const uint32_t jobCount = AZ::DivideAndRoundUp(compilesInPool, compilesPerJob);
 
                         for (uint32_t i = 0; i < jobCount; ++i)
                         {

Gems/Atom/RHI/DX12/Code/Source/RHI/FrameGraphExecuter.cpp

@@ -92,7 +92,7 @@ namespace AZ
                 const uint32_t CommandListCostThreshold =
                     AZStd::max(
                         m_frameGraphExecuterData.m_commandListCostThresholdMin,
-                        RHI::DivideByMultiple(estimatedItemCount, m_frameGraphExecuterData.m_commandListsPerScopeMax));
+                        AZ::DivideAndRoundUp(estimatedItemCount, m_frameGraphExecuterData.m_commandListsPerScopeMax));
 
                 /**
                  * Computes a cost heuristic based on the number of items and number of attachments in
@@ -145,7 +145,7 @@ namespace AZ
                 else
                 {
                     // And then create a new group for the current scope with dedicated [1, N] command lists.
-                    const uint32_t commandListCount = AZStd::max(RHI::DivideByMultiple(totalScopeCost, CommandListCostThreshold), 1u);
+                    const uint32_t commandListCount = AZStd::max(AZ::DivideAndRoundUp(totalScopeCost, CommandListCostThreshold), 1u);
 
                     FrameGraphExecuteGroup* scopeContextGroup = AddGroup<FrameGraphExecuteGroup>();
                     scopeContextGroup->Init(device, scope, commandListCount, GetJobPolicy());

Gems/Atom/RHI/Vulkan/Code/Source/RHI/FrameGraphExecuter.cpp

@@ -81,7 +81,7 @@ namespace AZ
                 const uint32_t CommandListCostThreshold =
                     AZStd::max(
                         m_frameGraphExecuterData.m_commandListCostThresholdMin,
-                        RHI::DivideByMultiple(estimatedItemCount, m_frameGraphExecuterData.m_commandListsPerScopeMax));
+                        AZ::DivideAndRoundUp(estimatedItemCount, m_frameGraphExecuterData.m_commandListsPerScopeMax));
 
                 /**
                     * Computes a cost heuristic based on the number of items and number of attachments in
@@ -136,7 +136,7 @@ namespace AZ
                 else
                 {
                     // And then create a new group for the current scope with dedicated [1, N] command lists.
-                    const uint32_t commandListCount = AZStd::max(RHI::DivideByMultiple(totalScopeCost, CommandListCostThreshold), 1u);
+                    const uint32_t commandListCount = AZStd::max(AZ::DivideAndRoundUp(totalScopeCost, CommandListCostThreshold), 1u);
 
                     FrameGraphExecuteGroup* scopeContextGroup = AddGroup<FrameGraphExecuteGroup>();
                     scopeContextGroup->Init(device, scope, commandListCount, GetJobPolicy());

Gems/Vegetation/Code/Source/Debugger/DebugComponent.cpp

@@ -334,16 +334,11 @@ void DebugComponent::FillSectorEnd([[maybe_unused]] int sectorX, [[maybe_unused]
 
 namespace DebugComponentUtilities
 {
-    constexpr uint32 RoundUpAndDivide(uint32 value, uint32 divide)
-    {
-        return (value + divide - 1) / divide;
-    }
-
     template <typename ValueType>
     union LocalAliasingUnion
     {
         ValueType aliasedValue;
-        AZStd::size_t aliasedValueArray[RoundUpAndDivide(sizeof(ValueType), sizeof(AZStd::size_t))] = {};
+        AZStd::size_t aliasedValueArray[AZ::DivideAndRoundUp(sizeof(ValueType), sizeof(AZStd::size_t))] = {};
     };
 
     template <typename ValueType>