diff --git a/changelog/atomic_cas.dd b/changelog/atomic_cas.dd
index 4a18661ddb..efecff0ec8 100644
--- a/changelog/atomic_cas.dd
+++ b/changelog/atomic_cas.dd
@@ -1,4 +1,4 @@
-Added overloads for `core.atomic.cas` which don't discard their result.
+Added overloads for `core.atomic : cas` which don't discard their result
 
-Existing `core.atomic.cas` functions discard the result, which make it impossible to implement certain operations.
+Existing `core.atomic : cas` functions discard the result, which make it impossible to implement certain operations.
 A new set of overloads was added which take `ifThis` by pointer and write the result back to the argument.
diff --git a/changelog/atomic_exchange.dd b/changelog/atomic_exchange.dd
index 12beb3b1db..cb69ca6484 100644
--- a/changelog/atomic_exchange.dd
+++ b/changelog/atomic_exchange.dd
@@ -1,3 +1,3 @@
-Added `core.atomic.atomicExchange`.
+Added `core.atomic : atomicExchange`
 
-Added missing `core.atomic.atomicExchange` function to the atomic suite.
+Added missing `atomicExchange` function to the atomic suite.
diff --git a/changelog/std_string.dd b/changelog/std_string.dd
index efae530954..f3a7b38ef3 100644
--- a/changelog/std_string.dd
+++ b/changelog/std_string.dd
@@ -1,6 +1,6 @@
-Added `core.stdcpp.string`.
+Added `core.stdcpp : string`
 
-Added `core.stdcpp.string`, which links against C++ `std::basic_string`
+Added `core.stdcpp : string`, which links against C++ `std::basic_string`.
 
 Known issues:
 
diff --git a/changelog/std_vector.dd b/changelog/std_vector.dd
index b8c940d147..e9399d621d 100644
--- a/changelog/std_vector.dd
+++ b/changelog/std_vector.dd
@@ -1,3 +1,3 @@
-Added `core.stdcpp.vector`.
+Added `core.stdcpp : vector`
 
-Added `core.stdcpp.vector`, which links against C++ `std::vector`
+Added `core.stdcpp : vector`, which links against C++ `std::vector`.
diff --git a/src/core/atomic.d b/src/core/atomic.d
index d47b876aa7..eec6e86f70 100644
--- a/src/core/atomic.d
+++ b/src/core/atomic.d
@@ -466,7 +466,7 @@ in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
 *  true if the store occurred, false if not.
 */
 bool casWeak(MemoryOrder succ = MemoryOrder.seq,MemoryOrder fail = MemoryOrder.seq,T,V)(T* here, T* ifThis, V writeThis) pure nothrow @nogc @trusted
-if (!is(T == shared S, S) && !is(V == shared U, U))
+    if (!is(T == shared S, S) && !is(V == shared U, U))
 in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
 {
     // resolve implicit conversions
@@ -483,7 +483,7 @@ in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
 
 /// Ditto
 bool casWeak(MemoryOrder succ = MemoryOrder.seq,MemoryOrder fail = MemoryOrder.seq,T,V1,V2)(shared(T)* here, V1* ifThis, V2 writeThis) pure nothrow @nogc @trusted
-if (!is(T == class) && (is(T : V1) || is(shared T : V1)))
+    if (!is(T == class) && (is(T : V1) || is(shared T : V1)))
 in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
 {
     static if (is (V1 == shared U1, U1))
@@ -508,7 +508,7 @@ in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
 
 /// Ditto
 bool casWeak(MemoryOrder succ = MemoryOrder.seq,MemoryOrder fail = MemoryOrder.seq,T,V)(shared(T)* here, shared(T)* ifThis, shared(V) writeThis) pure nothrow @nogc @trusted
-if (is(T == class))
+    if (is(T == class))
 in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
 {
     return atomicCompareExchangeWeak!(succ, fail)(cast(T*)here, cast(T*)ifThis, cast(V)writeThis);
@@ -519,11 +519,20 @@ in (atomicPtrIsProperlyAligned(here), "Argument `here` is not properly aligned")
  * that all loads and stores before a call to this function are executed before any
  * loads and stores after the call.
  */
-void atomicFence(MemoryOrder order = MemoryOrder.seq)() nothrow @nogc @safe
+void atomicFence(MemoryOrder order = MemoryOrder.seq)() pure nothrow @nogc @safe
 {
     core.internal.atomic.atomicFence!order();
 }
 
+/**
+ * Gives a hint to the processor that the calling thread is in a 'spin-wait' loop,
+ * allowing to more efficiently allocate resources.
+ */
+void pause() pure nothrow @nogc @safe
+{
+    core.internal.atomic.pause();
+}
+
 /**
  * Performs the binary operation 'op' on val using 'mod' as the modifier.
  *
diff --git a/src/core/internal/atomic.d b/src/core/internal/atomic.d
index 215e992c6a..5e9207b57a 100644
--- a/src/core/internal/atomic.d
+++ b/src/core/internal/atomic.d
@@ -589,7 +589,7 @@ bool atomicCompareExchangeStrongNoResult(MemoryOrder succ = MemoryOrder.seq, Mem
         static assert (false, "Unsupported architecture.");
 }
 
-void atomicFence(MemoryOrder order = MemoryOrder.seq)() nothrow @nogc @safe
+void atomicFence(MemoryOrder order = MemoryOrder.seq)() pure nothrow @nogc @safe
 {
     // TODO: `mfence` should only be required for seq_cst operations, but this depends on
     //       the compiler's backend knowledge to not reorder code inappropriately,
@@ -633,7 +633,7 @@ void atomicFence(MemoryOrder order = MemoryOrder.seq)() nothrow @nogc @safe
         }
         else version (D_InlineAsm_X86_64)
         {
-            asm nothrow @nogc @trusted
+            asm pure nothrow @nogc @trusted
             {
                 naked;
                 mfence;
@@ -645,6 +645,33 @@ void atomicFence(MemoryOrder order = MemoryOrder.seq)() nothrow @nogc @safe
     }
 }
 
+void pause() pure nothrow @nogc @safe
+{
+    version (D_InlineAsm_X86)
+    {
+        asm pure nothrow @nogc @trusted
+        {
+            naked;
+            rep; nop;
+            ret;
+        }
+    }
+    else version (D_InlineAsm_X86_64)
+    {
+        asm pure nothrow @nogc @trusted
+        {
+            naked;
+//            pause; // TODO: DMD should add this opcode to its inline asm
+            rep; nop;
+            ret;
+        }
+    }
+    else
+    {
+        // ARM should `yield`
+        // other architectures? otherwise some sort of nop...
+    }
+}
 
 private:
 
diff --git a/src/core/internal/spinlock.d b/src/core/internal/spinlock.d
index cad029d0c2..36d806ad01 100644
--- a/src/core/internal/spinlock.d
+++ b/src/core/internal/spinlock.d
@@ -52,7 +52,7 @@ shared struct SpinLock
     {
         import core.time;
         if (k < pauseThresh)
-            return pause();
+            return core.atomic.pause();
         else if (k < 32)
             return Thread.yield();
         Thread.sleep(1.msecs);
@@ -67,25 +67,9 @@ private:
         enum X86 = false;
 
     static if (X86)
-    {
         enum pauseThresh = 16;
-        void pause()
-        {
-            asm @trusted @nogc nothrow
-            {
-                // pause instruction
-                rep;
-                nop;
-            }
-        }
-    }
     else
-    {
         enum pauseThresh = 4;
-        void pause()
-        {
-        }
-    }
 
     size_t val;
     Contention contention;