diff --git a/CMakeLists.txt b/CMakeLists.txt
index 23d9310..70596df 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -1749,7 +1749,6 @@
     ${SOURCE_DIR}/System/Socket.cpp
     ${SOURCE_DIR}/System/Socket.hpp
     ${SOURCE_DIR}/System/Synchronization.hpp
-    ${SOURCE_DIR}/System/Thread.hpp
     ${SOURCE_DIR}/System/Timer.cpp
     ${SOURCE_DIR}/System/Timer.hpp
     ${SOURCE_DIR}/Device/*.cpp
diff --git a/src/Device/Renderer.cpp b/src/Device/Renderer.cpp
index d97c803..ddba431 100644
--- a/src/Device/Renderer.cpp
+++ b/src/Device/Renderer.cpp
@@ -42,9 +42,9 @@
 namespace sw
 {
 	static const int batchSize = 128;
-	AtomicInt threadCount(1);
-	AtomicInt Renderer::unitCount(1);
-	AtomicInt Renderer::clusterCount(1);
+	std::atomic<int> threadCount(1);
+	std::atomic<int> Renderer::unitCount(1);
+	std::atomic<int> Renderer::clusterCount(1);
 
 	template<typename T>
 	inline bool setBatchIndices(unsigned int batch[128][3], VkPrimitiveTopology topology, T indices, unsigned int start, unsigned int triangleCount)
@@ -572,7 +572,7 @@
 
 								// Commit to the task queue
 								qHead = (qHead + 1) & TASK_COUNT_BITS;
-								qSize++;
+								++qSize; // Atomic
 
 								break;
 							}
@@ -613,7 +613,7 @@
 				count = draw->count;
 				int batch = draw->batchSize;
 
-				primitiveProgress[unit].drawCall = currentDraw;
+				primitiveProgress[unit].drawCall = currentDraw.load();
 				primitiveProgress[unit].firstPrimitive = primitive;
 				primitiveProgress[unit].primitiveCount = count - primitive >= batch ? batch : count - primitive;
 
@@ -627,7 +627,7 @@
 
 				// Commit to the task queue
 				qHead = (qHead + 1) & TASK_COUNT_BITS;
-				qSize++;
+				++qSize; // Atomic
 			}
 		}
 	}
@@ -646,7 +646,7 @@
 		if(qSize != 0)
 		{
 			task[threadIndex] = taskQueue[(qHead - qSize) & TASK_COUNT_BITS];
-			qSize--;
+			--qSize; // Atomic
 
 			if(curThreadsAwake != threadCount)
 			{
@@ -678,7 +678,7 @@
 
 	void Renderer::executeTask(int threadIndex)
 	{
-		switch(task[threadIndex].type)
+		switch(task[threadIndex].type.load())
 		{
 		case Task::PRIMITIVES:
 			{
@@ -699,7 +699,7 @@
 				}
 
 				primitiveProgress[unit].visible = visible;
-				primitiveProgress[unit].references = clusterCount;
+				primitiveProgress[unit].references = clusterCount.load();
 			}
 			break;
 		case Task::PIXELS:
@@ -754,11 +754,11 @@
 			pixelProgress[cluster].processedPrimitives = 0;
 		}
 
-		int ref = primitiveProgress[unit].references--; // Atomic
+		int ref = --primitiveProgress[unit].references; // Atomic
 
 		if(ref == 0)
 		{
-			ref = draw.references--; // Atomic
+			ref = --draw.references; // Atomic
 
 			if(ref == 0)
 			{
@@ -839,7 +839,7 @@
 		}
 		else
 		{
-			switch(draw->indexType)
+			switch(draw->indexType.load())
 			{
 			case VK_INDEX_TYPE_UINT16:
 				if(!setBatchIndices(batch, topology, static_cast<const uint16_t*>(indices), start, triangleCount))
diff --git a/src/Device/Renderer.hpp b/src/Device/Renderer.hpp
index bfd4bb0..85367f1 100644
--- a/src/Device/Renderer.hpp
+++ b/src/Device/Renderer.hpp
@@ -22,9 +22,9 @@
 #include "Blitter.hpp"
 #include "Device/Config.hpp"
 #include "System/Synchronization.hpp"
-#include "System/Thread.hpp"
 #include "Vulkan/VkDescriptorSet.hpp"
 
+#include <atomic>
 #include <list>
 #include <mutex>
 #include <thread>
@@ -111,9 +111,16 @@
 				SUSPEND
 			};
 
-			AtomicInt type;
-			AtomicInt primitiveUnit;
-			AtomicInt pixelCluster;
+			void operator=(const Task& task)
+			{
+				type = task.type.load();
+				primitiveUnit = task.primitiveUnit.load();
+				pixelCluster = task.pixelCluster.load();
+			}
+
+			std::atomic<int> type;
+			std::atomic<int> primitiveUnit;
+			std::atomic<int> pixelCluster;
 		};
 
 		struct PrimitiveProgress
@@ -127,11 +134,11 @@
 				references = 0;
 			}
 
-			AtomicInt drawCall;
-			AtomicInt firstPrimitive;
-			AtomicInt primitiveCount;
-			AtomicInt visible;
-			AtomicInt references;
+			std::atomic<int> drawCall;
+			std::atomic<int> firstPrimitive;
+			std::atomic<int> primitiveCount;
+			std::atomic<int> visible;
+			std::atomic<int> references;
 		};
 
 		struct PixelProgress
@@ -143,9 +150,9 @@
 				executing = false;
 			}
 
-			AtomicInt drawCall;
-			AtomicInt processedPrimitives;
-			AtomicInt executing;
+			std::atomic<int> drawCall;
+			std::atomic<int> processedPrimitives;
+			std::atomic<int> executing;
 		};
 
 	public:
@@ -201,8 +208,8 @@
 		Triangle *triangleBatch[16];
 		Primitive *primitiveBatch[16];
 
-		AtomicInt exitThreads;
-		AtomicInt threadsAwake;
+		std::atomic<int> exitThreads;
+		std::atomic<int> threadsAwake;
 		std::thread *worker[16];
 		Event *resume[16];         // Events for resuming threads
 		Event *suspend[16];        // Events for suspending threads
@@ -219,19 +226,19 @@
 		DrawCall *drawCall[DRAW_COUNT];
 		DrawCall *drawList[DRAW_COUNT];
 
-		AtomicInt currentDraw;
-		AtomicInt nextDraw;
+		std::atomic<int> currentDraw;
+		std::atomic<int> nextDraw;
 
 		enum {
 			TASK_COUNT = 32,   // Size of the task queue (must be power of 2)
 			TASK_COUNT_BITS = TASK_COUNT - 1,
 		};
 		Task taskQueue[TASK_COUNT];
-		AtomicInt qHead;
-		AtomicInt qSize;
+		std::atomic<int> qHead;
+		std::atomic<int> qSize;
 
-		static AtomicInt unitCount;
-		static AtomicInt clusterCount;
+		static std::atomic<int> unitCount;
+		static std::atomic<int> clusterCount;
 
 		std::mutex schedulerMutex;
 
@@ -255,9 +262,9 @@
 
 		~DrawCall();
 
-		AtomicInt topology;
-		AtomicInt indexType;
-		AtomicInt batchSize;
+		std::atomic<int> topology;
+		std::atomic<int> indexType;
+		std::atomic<int> batchSize;
 
 		Routine *vertexRoutine;
 		Routine *setupRoutine;
@@ -277,9 +284,9 @@
 
 		std::list<vk::Query*> *queries;
 
-		AtomicInt primitive;    // Current primitive to enter pipeline
-		AtomicInt count;        // Number of primitives to render
-		AtomicInt references;   // Remaining references to this draw call, 0 when done drawing, -1 when resources unlocked and slot is free
+		std::atomic<int> primitive;    // Current primitive to enter pipeline
+		std::atomic<int> count;        // Number of primitives to render
+		std::atomic<int> references;   // Remaining references to this draw call, 0 when done drawing, -1 when resources unlocked and slot is free
 
 		DrawData *data;
 	};
diff --git a/src/System/BUILD.gn b/src/System/BUILD.gn
index 7599617..38256f7 100644
--- a/src/System/BUILD.gn
+++ b/src/System/BUILD.gn
@@ -30,7 +30,6 @@
     "Memory.hpp",
     "Socket.cpp",
     "Socket.hpp",
-    "Thread.hpp",
     "Timer.cpp",
     "Timer.hpp",
   ]
diff --git a/src/System/Thread.hpp b/src/System/Thread.hpp
deleted file mode 100644
index 3b6770d..0000000
--- a/src/System/Thread.hpp
+++ /dev/null
@@ -1,42 +0,0 @@
-// Copyright 2016 The SwiftShader Authors. All Rights Reserved.
-//
-// Licensed under the Apache License, Version 2.0 (the "License");
-// you may not use this file except in compliance with the License.
-// You may obtain a copy of the License at
-//
-//    http://www.apache.org/licenses/LICENSE-2.0
-//
-// Unless required by applicable law or agreed to in writing, software
-// distributed under the License is distributed on an "AS IS" BASIS,
-// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-// See the License for the specific language governing permissions and
-// limitations under the License.
-
-#ifndef sw_Thread_hpp
-#define sw_Thread_hpp
-
-#include <atomic>
-
-namespace sw
-{
-	class AtomicInt
-	{
-	public:
-		AtomicInt() : ai() {}
-		AtomicInt(int i) : ai(i) {}
-
-		inline operator int() const { return ai.load(std::memory_order_acquire); }
-		inline void operator=(const AtomicInt& i) { ai.store(i.ai.load(std::memory_order_acquire), std::memory_order_release); }
-		inline void operator=(int i) { ai.store(i, std::memory_order_release); }
-		inline void operator--() { ai.fetch_sub(1, std::memory_order_acq_rel); }
-		inline void operator++() { ai.fetch_add(1, std::memory_order_acq_rel); }
-		inline int operator--(int) { return ai.fetch_sub(1, std::memory_order_acq_rel) - 1; }
-		inline int operator++(int) { return ai.fetch_add(1, std::memory_order_acq_rel) + 1; }
-		inline void operator-=(int i) { ai.fetch_sub(i, std::memory_order_acq_rel); }
-		inline void operator+=(int i) { ai.fetch_add(i, std::memory_order_acq_rel); }
-	private:
-		std::atomic<int> ai;
-	};
-}
-
-#endif   // sw_Thread_hpp
diff --git a/src/Vulkan/vulkan.vcxproj b/src/Vulkan/vulkan.vcxproj
index fcc808d..9da32ff 100644
--- a/src/Vulkan/vulkan.vcxproj
+++ b/src/Vulkan/vulkan.vcxproj
@@ -274,7 +274,6 @@
     <ClInclude Include="..\System\SharedLibrary.hpp" />
     <ClInclude Include="..\System\Socket.hpp" />
     <ClInclude Include="..\System\Synchronization.hpp" />
-    <ClInclude Include="..\System\Thread.hpp" />
     <ClInclude Include="..\System\Timer.hpp" />
     <ClInclude Include="..\System\Types.hpp" />
     <ClInclude Include="..\WSI\VkSurfaceKHR.hpp" />
diff --git a/src/Vulkan/vulkan.vcxproj.filters b/src/Vulkan/vulkan.vcxproj.filters
index 3181b58..4f325dc 100644
--- a/src/Vulkan/vulkan.vcxproj.filters
+++ b/src/Vulkan/vulkan.vcxproj.filters
@@ -488,9 +488,6 @@
     <ClInclude Include="..\System\Synchronization.hpp">
       <Filter>Header Files\System</Filter>
     </ClInclude>
-    <ClInclude Include="..\System\Thread.hpp">
-      <Filter>Header Files\System</Filter>
-    </ClInclude>
     <ClInclude Include="..\System\Timer.hpp">
       <Filter>Header Files\System</Filter>
     </ClInclude>