diff --git a/LICENSE b/LICENSE
new file mode 100644
index 0000000..da382fb
--- /dev/null
+++ b/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2020 Pavel Kirienko
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/o1heap.cpp b/o1heap.cpp
index 93e2ab3..03ccbba 100644
--- a/o1heap.cpp
+++ b/o1heap.cpp
@@ -34,8 +34,11 @@
 /// The assertion macro defaults to the standard assert().
 /// It can be overridden to manually suppress assertion checks or use a different error handling policy.
 #ifndef O1HEAP_ASSERT
-// Intentional violation of MISRA: the assertion check macro cannot be replaced with a function definition.
-#    define O1HEAP_ASSERT(x) assert(x)  // NOSONAR
+    #if defined(AU_CFG_ID_DEBUG) || defined(AU_CFG_ID_INTERNAL)
+        #define O1HEAP_ASSERT(x) SysAssert(x)
+    #else
+        #define O1HEAP_ASSERT(x) 
+    #endif
 #endif
 
 /// Branch probability annotations are used to improve the worst case execution time (WCET). They are entirely optional.
@@ -43,7 +46,6 @@
 /// If you are using a different compiler, consider overriding this value.
 #ifndef O1HEAP_LIKELY
 #    if defined(__GNUC__) || defined(__clang__) || defined(__CC_ARM)
-// Intentional violation of MISRA: branch hinting macro cannot be replaced with a function definition.
 #        define O1HEAP_LIKELY(x) __builtin_expect((x), 1)  // NOSONAR
 #    else
 #        define O1HEAP_LIKELY(x) x
@@ -134,26 +136,27 @@ O1HEAP_PRIVATE bool isPowerOf2(const size_t x)
 }
 
 /// Special case: if the argument is zero, returns zero.
-O1HEAP_PRIVATE uint8_t log2Floor(const size_t x);
 O1HEAP_PRIVATE uint8_t log2Floor(const size_t x)
 {
-    size_t  tmp = x;
-    uint8_t y = 0;
-    // This is currently the only exception to the statement "routines contain neither loops nor recursion".
-    // It is unclear if there is a better way to compute the binary logarithm than this.
-    while (tmp > 1U)
-    {
-        tmp >>= 1U;
-        y++;
-    }
-    return y;
+    O1HEAP_ASSERT(x > 0);
+    AuUInt8 index {};
+    AuBitScanReverse(index, x);
+    return (uint8_t)(((sizeof(x) * CHAR_BIT) - 1U) - index);
 }
 
-/// Special case: if the argument is zero, returns zero.
-O1HEAP_PRIVATE uint8_t log2Ceil(const size_t x);
-O1HEAP_PRIVATE uint8_t log2Ceil(const size_t x)
+O1HEAP_PRIVATE uint_fast8_t log2Ceil(const size_t x)
 {
-    return (uint8_t)(log2Floor(x) + (isPowerOf2(x) ? 0U : 1U));
+    AuUInt8 index {};
+    AuBitScanReverse(index, x - 1);
+    return (x <= 1U) ? 0U : (uint_fast8_t)((sizeof(x) * CHAR_BIT) - (index));
+}
+
+O1HEAP_PRIVATE size_t roundUpToPowerOf2(const size_t x)
+{
+    O1HEAP_ASSERT(x >= 2U);
+    AuUInt8 index {};
+    AuBitScanReverse(index, x - 1);
+    return ((size_t)1U) << ((sizeof(x) * CHAR_BIT) - (index));
 }
 
 /// Raise 2 into the specified power.
@@ -316,11 +319,11 @@ void *o1heapAllocate(O1HeapInstance *const handle, const size_t amount)
     {
         // Add the header size and align the allocation size to the power of 2.
         // See "Timing-Predictable Memory Allocation In Hard Real-Time Systems", Herter, page 27.
-        const size_t fragment_size = pow2(log2Ceil(amount + O1HEAP_ALIGNMENT));
+        const size_t fragment_size = roundUpToPowerOf2(amount + O1HEAP_ALIGNMENT);
         O1HEAP_ASSERT(fragment_size <= FRAGMENT_SIZE_MAX);
         O1HEAP_ASSERT(fragment_size >= FRAGMENT_SIZE_MIN);
         O1HEAP_ASSERT(fragment_size >= amount + O1HEAP_ALIGNMENT);
-        O1HEAP_ASSERT(isPowerOf2(fragment_size));
+        O1HEAP_ASSERT((fragment_size & (fragment_size - 1U)) == 0U); // Is power of 2.
 
         const uint8_t optimal_bin_index = log2Ceil(fragment_size / FRAGMENT_SIZE_MIN);  // Use CEIL when fetching.
         O1HEAP_ASSERT(optimal_bin_index < NUM_BINS_MAX);
@@ -333,8 +336,8 @@ void *o1heapAllocate(O1HeapInstance *const handle, const size_t amount)
         const size_t smallest_bin_mask = suitable_bins & ~(suitable_bins - 1U);  // Clear all bits but the lowest.
         if (O1HEAP_LIKELY(smallest_bin_mask != 0))
         {
-            O1HEAP_ASSERT(isPowerOf2(smallest_bin_mask));
-            const uint8_t bin_index = log2Floor(smallest_bin_mask);
+            O1HEAP_ASSERT((smallest_bin_mask & (smallest_bin_mask - 1U)) == 0U);  // Is power of 2.
+            const uint_fast8_t bin_index = log2Floor(smallest_bin_mask);
             O1HEAP_ASSERT(bin_index >= optimal_bin_index);
             O1HEAP_ASSERT(bin_index < NUM_BINS_MAX);
 
@@ -353,7 +356,7 @@ void *o1heapAllocate(O1HeapInstance *const handle, const size_t amount)
             O1HEAP_ASSERT(leftover % FRAGMENT_SIZE_MIN == 0U);       // Alignment check.
             if (O1HEAP_LIKELY(leftover >= FRAGMENT_SIZE_MIN))
             {
-                Fragment *const new_frag = (Fragment *)(void *)(((uint8_t *)frag) + fragment_size);
+                Fragment *const new_frag = (Fragment *)(void *)(((char *)frag) + fragment_size);
                 O1HEAP_ASSERT(((size_t)new_frag) % O1HEAP_ALIGNMENT == 0U);
                 new_frag->header.size = leftover;
                 new_frag->header.used = false;