diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt
index f8a55c2a7f..55a5f9eba9 100644
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -23,9 +23,6 @@ set(SRCS
             )
 
 set(HEADERS
-            atomic.h
-            atomic_gcc.h
-            atomic_win32.h
             bit_field.h
             break_points.h
             chunk_file.h
diff --git a/src/common/atomic.h b/src/common/atomic.h
deleted file mode 100644
index 941f5ad5ea..0000000000
--- a/src/common/atomic.h
+++ /dev/null
@@ -1,16 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project
-// Licensed under GPLv2
-// Refer to the license.txt file included.
-
-#pragma once
-
-#ifdef _WIN32
-
-#include "common/atomic_win32.h"
-
-#else
-
-// GCC-compatible compiler assumed!
-#include "common/atomic_gcc.h"
-
-#endif
diff --git a/src/common/atomic_gcc.h b/src/common/atomic_gcc.h
deleted file mode 100644
index 117e342f6b..0000000000
--- a/src/common/atomic_gcc.h
+++ /dev/null
@@ -1,110 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project
-// Licensed under GPLv2
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include "common/common.h"
-
-// Atomic operations are performed in a single step by the CPU. It is
-// impossible for other threads to see the operation "half-done."
-//
-// Some atomic operations can be combined with different types of memory
-// barriers called "Acquire semantics" and "Release semantics", defined below.
-//
-// Acquire semantics: Future memory accesses cannot be relocated to before the
-//                    operation.
-//
-// Release semantics: Past memory accesses cannot be relocated to after the
-//                    operation.
-//
-// These barriers affect not only the compiler, but also the CPU.
-
-namespace Common
-{
-
-inline void AtomicAdd(volatile u32& target, u32 value) {
-    __sync_add_and_fetch(&target, value);
-}
-
-inline void AtomicAnd(volatile u32& target, u32 value) {
-    __sync_and_and_fetch(&target, value);
-}
-
-inline void AtomicDecrement(volatile u32& target) {
-    __sync_add_and_fetch(&target, -1);
-}
-
-inline void AtomicIncrement(volatile u32& target) {
-    __sync_add_and_fetch(&target, 1);
-}
-
-inline u32 AtomicLoad(volatile u32& src) {
-    return src; // 32-bit reads are always atomic.
-}
-inline u32 AtomicLoadAcquire(volatile u32& src) {
-    //keep the compiler from caching any memory references
-    u32 result = src; // 32-bit reads are always atomic.
-    //__sync_synchronize(); // TODO: May not be necessary.
-    // Compiler instruction only. x86 loads always have acquire semantics.
-    __asm__ __volatile__ ( "":::"memory" );
-    return result;
-}
-
-inline void AtomicOr(volatile u32& target, u32 value) {
-    __sync_or_and_fetch(&target, value);
-}
-
-inline void AtomicStore(volatile u32& dest, u32 value) {
-    dest = value; // 32-bit writes are always atomic.
-}
-inline void AtomicStoreRelease(volatile u32& dest, u32 value) {
-    __sync_lock_test_and_set(&dest, value); // TODO: Wrong! This function is has acquire semantics.
-}
-
-}
-
-// Old code kept here for reference in case we need the parts with __asm__ __volatile__.
-#if 0
-LONG SyncInterlockedIncrement(LONG *Dest)
-{
-#if defined(__GNUC__) && defined (__GNUC_MINOR__) && ((4 < __GNUC__) || (4 == __GNUC__ && 1 <= __GNUC_MINOR__))
-  return  __sync_add_and_fetch(Dest, 1);
-#else
-  register int result;
-  __asm__ __volatile__("lock; xadd %0,%1"
-                       : "=r" (result), "=m" (*Dest)
-                       : "0" (1), "m" (*Dest)
-                       : "memory");
-  return result;
-#endif
-}
-
-LONG SyncInterlockedExchangeAdd(LONG *Dest, LONG Val)
-{
-#if defined(__GNUC__) && defined (__GNUC_MINOR__) && ((4 < __GNUC__) || (4 == __GNUC__ && 1 <= __GNUC_MINOR__))
-  return  __sync_add_and_fetch(Dest, Val);
-#else
-  register int result;
-  __asm__ __volatile__("lock; xadd %0,%1"
-                       : "=r" (result), "=m" (*Dest)
-                       : "0" (Val), "m" (*Dest)
-                       : "memory");
-  return result;
-#endif
-}
-
-LONG SyncInterlockedExchange(LONG *Dest, LONG Val)
-{
-#if defined(__GNUC__) && defined (__GNUC_MINOR__) && ((4 < __GNUC__) || (4 == __GNUC__ && 1 <= __GNUC_MINOR__))
-  return  __sync_lock_test_and_set(Dest, Val);
-#else
-  register int result;
-  __asm__ __volatile__("lock; xchg %0,%1"
-                       : "=r" (result), "=m" (*Dest)
-                       : "0" (Val), "m" (*Dest)
-                       : "memory");
-  return result;
-#endif
-}
-#endif
diff --git a/src/common/atomic_win32.h b/src/common/atomic_win32.h
deleted file mode 100644
index 0808905f0d..0000000000
--- a/src/common/atomic_win32.h
+++ /dev/null
@@ -1,69 +0,0 @@
-// Copyright 2013 Dolphin Emulator Project
-// Licensed under GPLv2
-// Refer to the license.txt file included.
-
-#pragma once
-
-#include "common/common.h"
-#include <intrin.h>
-#include <Windows.h>
-
-// Atomic operations are performed in a single step by the CPU. It is
-// impossible for other threads to see the operation "half-done."
-//
-// Some atomic operations can be combined with different types of memory
-// barriers called "Acquire semantics" and "Release semantics", defined below.
-//
-// Acquire semantics: Future memory accesses cannot be relocated to before the
-//                    operation.
-//
-// Release semantics: Past memory accesses cannot be relocated to after the
-//                    operation.
-//
-// These barriers affect not only the compiler, but also the CPU.
-//
-// NOTE: Acquire and Release are not differentiated right now. They perform a
-// full memory barrier instead of a "one-way" memory barrier. The newest
-// Windows SDK has Acquire and Release versions of some Interlocked* functions.
-
-namespace Common
-{
-
-inline void AtomicAdd(volatile u32& target, u32 value) {
-    InterlockedExchangeAdd((volatile LONG*)&target, (LONG)value);
-}
-
-inline void AtomicAnd(volatile u32& target, u32 value) {
-    _InterlockedAnd((volatile LONG*)&target, (LONG)value);
-}
-
-inline void AtomicIncrement(volatile u32& target) {
-    InterlockedIncrement((volatile LONG*)&target);
-}
-
-inline void AtomicDecrement(volatile u32& target) {
-    InterlockedDecrement((volatile LONG*)&target);
-}
-
-inline u32 AtomicLoad(volatile u32& src) {
-    return src; // 32-bit reads are always atomic.
-}
-inline u32 AtomicLoadAcquire(volatile u32& src) {
-    u32 result = src; // 32-bit reads are always atomic.
-    _ReadBarrier(); // Compiler instruction only. x86 loads always have acquire semantics.
-    return result;
-}
-
-inline void AtomicOr(volatile u32& target, u32 value) {
-    _InterlockedOr((volatile LONG*)&target, (LONG)value);
-}
-
-inline void AtomicStore(volatile u32& dest, u32 value) {
-    dest = value; // 32-bit writes are always atomic.
-}
-inline void AtomicStoreRelease(volatile u32& dest, u32 value) {
-    _WriteBarrier(); // Compiler instruction only. x86 stores always have release semantics.
-    dest = value; // 32-bit writes are always atomic.
-}
-
-}
diff --git a/src/core/core_timing.cpp b/src/core/core_timing.cpp
index a4fc0aaa40..c30e367329 100644
--- a/src/core/core_timing.cpp
+++ b/src/core/core_timing.cpp
@@ -4,10 +4,10 @@
 
 #include <vector>
 #include <cstdio>
+#include <atomic>
 
 #include "common/msg_handler.h"
 #include "common/std_mutex.h"
-#include "common/atomic.h"
 #include "common/chunk_file.h"
 
 #include "core/core_timing.h"
@@ -54,7 +54,7 @@ Event *eventPool = 0;
 Event *eventTsPool = 0;
 int allocatedTsEvents = 0;
 // Optimization to skip MoveEvents when possible.
-volatile u32 hasTsEvents = false;
+std::atomic<u32> hasTsEvents;
 
 // Downcount has been moved to currentMIPS, to save a couple of clocks in every ARM JIT block
 // as we can already reach that structure through a register.
@@ -202,7 +202,7 @@ void ScheduleEvent_Threadsafe(s64 cyclesIntoFuture, int event_type, u64 userdata
         tsLast->next = ne;
     tsLast = ne;
 
-    Common::AtomicStoreRelease(hasTsEvents, 1);
+    hasTsEvents.store(1, std::memory_order_release);
 }
 
 // Same as ScheduleEvent_Threadsafe(0, ...) EXCEPT if we are already on the CPU thread
@@ -484,7 +484,7 @@ void ProcessFifoWaitEvents()
 
 void MoveEvents()
 {
-    Common::AtomicStoreRelease(hasTsEvents, 0);
+    hasTsEvents.store(0, std::memory_order_release);
 
     std::lock_guard<std::recursive_mutex> lk(externalEventSection);
     // Move events from async queue into main queue