26 files changed, 1032 insertions, 200 deletions

diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt
index bdd885273..5639021d3 100644
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -47,6 +47,7 @@ add_custom_command(OUTPUT scm_rev.cpp
       "${VIDEO_CORE}/shader/decode/integer_set.cpp"
       "${VIDEO_CORE}/shader/decode/integer_set_predicate.cpp"
       "${VIDEO_CORE}/shader/decode/memory.cpp"
+      "${VIDEO_CORE}/shader/decode/texture.cpp"
       "${VIDEO_CORE}/shader/decode/other.cpp"
       "${VIDEO_CORE}/shader/decode/predicate_set_predicate.cpp"
       "${VIDEO_CORE}/shader/decode/predicate_set_register.cpp"
@@ -90,11 +91,18 @@ add_library(common STATIC
     logging/log.h
     logging/text_formatter.cpp
     logging/text_formatter.h
+    lz4_compression.cpp
+    lz4_compression.h
     math_util.h
+    memory_hook.cpp
+    memory_hook.h
     microprofile.cpp
     microprofile.h
     microprofileui.h
     misc.cpp
+    multi_level_queue.h
+    page_table.cpp
+    page_table.h
     param_package.cpp
     param_package.h
     quaternion.h
@@ -113,6 +121,8 @@ add_library(common STATIC
     threadsafe_queue.h
     timer.cpp
     timer.h
+    uint128.cpp
+    uint128.h
     vector_math.h
     web_result.h
 )
@@ -128,3 +138,4 @@ endif()
 create_target_directory_groups(common)
 
 target_link_libraries(common PUBLIC Boost::boost fmt microprofile)
+target_link_libraries(common PRIVATE lz4_static)
diff --git a/src/common/bit_field.h b/src/common/bit_field.h
index 21e07925d..902e668e3 100644
--- a/src/common/bit_field.h
+++ b/src/common/bit_field.h
@@ -34,6 +34,7 @@
 #include <limits>
 #include <type_traits>
 #include "common/common_funcs.h"
+#include "common/swap.h"
 
 /*
  * Abstract bitfield class
@@ -108,15 +109,9 @@
  * symptoms.
  */
 #pragma pack(1)
-template <std::size_t Position, std::size_t Bits, typename T>
+template <std::size_t Position, std::size_t Bits, typename T, typename EndianTag = LETag>
 struct BitField {
 private:
-    // We hide the copy assigment operator here, because the default copy
-    // assignment would copy the full storage value, rather than just the bits
-    // relevant to this particular bit field.
-    // We don't delete it because we want BitField to be trivially copyable.
-    constexpr BitField& operator=(const BitField&) = default;
-
     // UnderlyingType is T for non-enum types and the underlying type of T if
     // T is an enumeration. Note that T is wrapped within an enable_if in the
     // former case to workaround compile errors which arise when using
@@ -127,6 +122,8 @@ private:
     // We store the value as the unsigned type to avoid undefined behaviour on value shifting
     using StorageType = std::make_unsigned_t<UnderlyingType>;
 
+    using StorageTypeWithEndian = typename AddEndian<StorageType, EndianTag>::type;
+
 public:
     /// Constants to allow limited introspection of fields if needed
     static constexpr std::size_t position = Position;
@@ -163,16 +160,20 @@ public:
     BitField(T val) = delete;
     BitField& operator=(T val) = delete;
 
-    // Force default constructor to be created
-    // so that we can use this within unions
-    constexpr BitField() = default;
+    constexpr BitField() noexcept = default;
+
+    constexpr BitField(const BitField&) noexcept = default;
+    constexpr BitField& operator=(const BitField&) noexcept = default;
+
+    constexpr BitField(BitField&&) noexcept = default;
+    constexpr BitField& operator=(BitField&&) noexcept = default;
 
     constexpr FORCE_INLINE operator T() const {
         return Value();
     }
 
     constexpr FORCE_INLINE void Assign(const T& value) {
-        storage = (storage & ~mask) | FormatValue(value);
+        storage = (static_cast<StorageType>(storage) & ~mask) | FormatValue(value);
     }
 
     constexpr T Value() const {
@@ -184,7 +185,7 @@ public:
     }
 
 private:
-    StorageType storage;
+    StorageTypeWithEndian storage;
 
     static_assert(bits + position <= 8 * sizeof(T), "Bitfield out of range");
 
@@ -195,3 +196,6 @@ private:
     static_assert(std::is_trivially_copyable_v<T>, "T must be trivially copyable in a BitField");
 };
 #pragma pack()
+
+template <std::size_t Position, std::size_t Bits, typename T>
+using BitFieldBE = BitField<Position, Bits, T, BETag>;
diff --git a/src/common/bit_util.h b/src/common/bit_util.h
index 1eea17ba1..a4f9ed4aa 100644
--- a/src/common/bit_util.h
+++ b/src/common/bit_util.h
@@ -58,4 +58,43 @@ inline u64 CountLeadingZeroes64(u64 value) {
     return __builtin_clzll(value);
 }
 #endif
+
+#ifdef _MSC_VER
+inline u32 CountTrailingZeroes32(u32 value) {
+    unsigned long trailing_zero = 0;
+
+    if (_BitScanForward(&trailing_zero, value) != 0) {
+        return trailing_zero;
+    }
+
+    return 32;
+}
+
+inline u64 CountTrailingZeroes64(u64 value) {
+    unsigned long trailing_zero = 0;
+
+    if (_BitScanForward64(&trailing_zero, value) != 0) {
+        return trailing_zero;
+    }
+
+    return 64;
+}
+#else
+inline u32 CountTrailingZeroes32(u32 value) {
+    if (value == 0) {
+        return 32;
+    }
+
+    return __builtin_ctz(value);
+}
+
+inline u64 CountTrailingZeroes64(u64 value) {
+    if (value == 0) {
+        return 64;
+    }
+
+    return __builtin_ctzll(value);
+}
+#endif
+
 } // namespace Common
diff --git a/src/common/color.h b/src/common/color.h
index 0379040be..3a2222077 100644
--- a/src/common/color.h
+++ b/src/common/color.h
@@ -55,36 +55,36 @@ constexpr u8 Convert8To6(u8 value) {
 /**
  * Decode a color stored in RGBA8 format
  * @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
  */
-inline Math::Vec4<u8> DecodeRGBA8(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGBA8(const u8* bytes) {
     return {bytes[3], bytes[2], bytes[1], bytes[0]};
 }
 
 /**
  * Decode a color stored in RGB8 format
  * @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
  */
-inline Math::Vec4<u8> DecodeRGB8(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGB8(const u8* bytes) {
     return {bytes[2], bytes[1], bytes[0], 255};
 }
 
 /**
  * Decode a color stored in RG8 (aka HILO8) format
  * @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
  */
-inline Math::Vec4<u8> DecodeRG8(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRG8(const u8* bytes) {
     return {bytes[1], bytes[0], 0, 255};
 }
 
 /**
  * Decode a color stored in RGB565 format
  * @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
  */
-inline Math::Vec4<u8> DecodeRGB565(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGB565(const u8* bytes) {
     u16_le pixel;
     std::memcpy(&pixel, bytes, sizeof(pixel));
     return {Convert5To8((pixel >> 11) & 0x1F), Convert6To8((pixel >> 5) & 0x3F),
@@ -94,9 +94,9 @@ inline Math::Vec4<u8> DecodeRGB565(const u8* bytes) {
 /**
  * Decode a color stored in RGB5A1 format
  * @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
  */
-inline Math::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
     u16_le pixel;
     std::memcpy(&pixel, bytes, sizeof(pixel));
     return {Convert5To8((pixel >> 11) & 0x1F), Convert5To8((pixel >> 6) & 0x1F),
@@ -106,9 +106,9 @@ inline Math::Vec4<u8> DecodeRGB5A1(const u8* bytes) {
 /**
  * Decode a color stored in RGBA4 format
  * @param bytes Pointer to encoded source color
- * @return Result color decoded as Math::Vec4<u8>
+ * @return Result color decoded as Common::Vec4<u8>
  */
-inline Math::Vec4<u8> DecodeRGBA4(const u8* bytes) {
+inline Common::Vec4<u8> DecodeRGBA4(const u8* bytes) {
     u16_le pixel;
     std::memcpy(&pixel, bytes, sizeof(pixel));
     return {Convert4To8((pixel >> 12) & 0xF), Convert4To8((pixel >> 8) & 0xF),
@@ -138,9 +138,9 @@ inline u32 DecodeD24(const u8* bytes) {
 /**
  * Decode a depth value and a stencil value stored in D24S8 format
  * @param bytes Pointer to encoded source values
- * @return Resulting values stored as a Math::Vec2
+ * @return Resulting values stored as a Common::Vec2
  */
-inline Math::Vec2<u32> DecodeD24S8(const u8* bytes) {
+inline Common::Vec2<u32> DecodeD24S8(const u8* bytes) {
     return {static_cast<u32>((bytes[2] << 16) | (bytes[1] << 8) | bytes[0]), bytes[3]};
 }
 
@@ -149,7 +149,7 @@ inline Math::Vec2<u32> DecodeD24S8(const u8* bytes) {
  * @param color Source color to encode
  * @param bytes Destination pointer to store encoded color
  */
-inline void EncodeRGBA8(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGBA8(const Common::Vec4<u8>& color, u8* bytes) {
     bytes[3] = color.r();
     bytes[2] = color.g();
     bytes[1] = color.b();
@@ -161,7 +161,7 @@ inline void EncodeRGBA8(const Math::Vec4<u8>& color, u8* bytes) {
  * @param color Source color to encode
  * @param bytes Destination pointer to store encoded color
  */
-inline void EncodeRGB8(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGB8(const Common::Vec4<u8>& color, u8* bytes) {
     bytes[2] = color.r();
     bytes[1] = color.g();
     bytes[0] = color.b();
@@ -172,7 +172,7 @@ inline void EncodeRGB8(const Math::Vec4<u8>& color, u8* bytes) {
  * @param color Source color to encode
  * @param bytes Destination pointer to store encoded color
  */
-inline void EncodeRG8(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRG8(const Common::Vec4<u8>& color, u8* bytes) {
     bytes[1] = color.r();
     bytes[0] = color.g();
 }
@@ -181,7 +181,7 @@ inline void EncodeRG8(const Math::Vec4<u8>& color, u8* bytes) {
  * @param color Source color to encode
  * @param bytes Destination pointer to store encoded color
  */
-inline void EncodeRGB565(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGB565(const Common::Vec4<u8>& color, u8* bytes) {
     const u16_le data =
         (Convert8To5(color.r()) << 11) | (Convert8To6(color.g()) << 5) | Convert8To5(color.b());
 
@@ -193,7 +193,7 @@ inline void EncodeRGB565(const Math::Vec4<u8>& color, u8* bytes) {
  * @param color Source color to encode
  * @param bytes Destination pointer to store encoded color
  */
-inline void EncodeRGB5A1(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGB5A1(const Common::Vec4<u8>& color, u8* bytes) {
     const u16_le data = (Convert8To5(color.r()) << 11) | (Convert8To5(color.g()) << 6) |
                         (Convert8To5(color.b()) << 1) | Convert8To1(color.a());
 
@@ -205,7 +205,7 @@ inline void EncodeRGB5A1(const Math::Vec4<u8>& color, u8* bytes) {
  * @param color Source color to encode
  * @param bytes Destination pointer to store encoded color
  */
-inline void EncodeRGBA4(const Math::Vec4<u8>& color, u8* bytes) {
+inline void EncodeRGBA4(const Common::Vec4<u8>& color, u8* bytes) {
     const u16 data = (Convert8To4(color.r()) << 12) | (Convert8To4(color.g()) << 8) |
                      (Convert8To4(color.b()) << 4) | Convert8To4(color.a());
 
diff --git a/src/common/common_types.h b/src/common/common_types.h
index 6b1766dca..4cec89fbd 100644
--- a/src/common/common_types.h
+++ b/src/common/common_types.h
@@ -40,10 +40,9 @@ using s64 = std::int64_t; ///< 64-bit signed int
 using f32 = float;  ///< 32-bit floating point
 using f64 = double; ///< 64-bit floating point
 
-// TODO: It would be nice to eventually replace these with strong types that prevent accidental
-// conversion between each other.
-using VAddr = u64; ///< Represents a pointer in the userspace virtual address space.
-using PAddr = u64; ///< Represents a pointer in the ARM11 physical address space.
+using VAddr = u64;    ///< Represents a pointer in the userspace virtual address space.
+using PAddr = u64;    ///< Represents a pointer in the ARM11 physical address space.
+using GPUVAddr = u64; ///< Represents a pointer in the GPU virtual address space.
 
 using u128 = std::array<std::uint64_t, 2>;
 static_assert(sizeof(u128) == 16, "u128 must be 128 bits wide");
diff --git a/src/common/detached_tasks.cpp b/src/common/detached_tasks.cpp
index a347d9e02..f268d6021 100644
--- a/src/common/detached_tasks.cpp
+++ b/src/common/detached_tasks.cpp
@@ -16,22 +16,22 @@ DetachedTasks::DetachedTasks() {
 }
 
 void DetachedTasks::WaitForAllTasks() {
-    std::unique_lock<std::mutex> lock(mutex);
+    std::unique_lock lock{mutex};
     cv.wait(lock, [this]() { return count == 0; });
 }
 
 DetachedTasks::~DetachedTasks() {
-    std::unique_lock<std::mutex> lock(mutex);
+    std::unique_lock lock{mutex};
     ASSERT(count == 0);
     instance = nullptr;
 }
 
 void DetachedTasks::AddTask(std::function<void()> task) {
-    std::unique_lock<std::mutex> lock(instance->mutex);
+    std::unique_lock lock{instance->mutex};
     ++instance->count;
     std::thread([task{std::move(task)}]() {
         task();
-        std::unique_lock<std::mutex> lock(instance->mutex);
+        std::unique_lock lock{instance->mutex};
         --instance->count;
         std::notify_all_at_thread_exit(instance->cv, std::move(lock));
     })
diff --git a/src/common/logging/backend.cpp b/src/common/logging/backend.cpp
index 12f6d0114..a03179520 100644
--- a/src/common/logging/backend.cpp
+++ b/src/common/logging/backend.cpp
@@ -39,19 +39,19 @@ public:
     Impl(Impl const&) = delete;
     const Impl& operator=(Impl const&) = delete;
 
-    void PushEntry(Entry e) {
-        std::lock_guard<std::mutex> lock(message_mutex);
-        message_queue.Push(std::move(e));
-        message_cv.notify_one();
+    void PushEntry(Class log_class, Level log_level, const char* filename, unsigned int line_num,
+                   const char* function, std::string message) {
+        message_queue.Push(
+            CreateEntry(log_class, log_level, filename, line_num, function, std::move(message)));
     }
 
     void AddBackend(std::unique_ptr<Backend> backend) {
-        std::lock_guard<std::mutex> lock(writing_mutex);
+        std::lock_guard lock{writing_mutex};
         backends.push_back(std::move(backend));
     }
 
     void RemoveBackend(std::string_view backend_name) {
-        std::lock_guard<std::mutex> lock(writing_mutex);
+        std::lock_guard lock{writing_mutex};
         const auto it =
             std::remove_if(backends.begin(), backends.end(),
                            [&backend_name](const auto& i) { return backend_name == i->GetName(); });
@@ -80,21 +80,19 @@ private:
         backend_thread = std::thread([&] {
             Entry entry;
             auto write_logs = [&](Entry& e) {
-                std::lock_guard<std::mutex> lock(writing_mutex);
+                std::lock_guard lock{writing_mutex};
                 for (const auto& backend : backends) {
                     backend->Write(e);
                 }
             };
             while (true) {
-                {
-                    std::unique_lock<std::mutex> lock(message_mutex);
-                    message_cv.wait(lock, [&] { return !running || message_queue.Pop(entry); });
-                }
-                if (!running) {
+                entry = message_queue.PopWait();
+                if (entry.final_entry) {
                     break;
                 }
                 write_logs(entry);
             }
+
             // Drain the logging queue. Only writes out up to MAX_LOGS_TO_WRITE to prevent a case
             // where a system is repeatedly spamming logs even on close.
             const int MAX_LOGS_TO_WRITE = filter.IsDebug() ? INT_MAX : 100;
@@ -106,18 +104,36 @@ private:
     }
 
     ~Impl() {
-        running = false;
-        message_cv.notify_one();
+        Entry entry;
+        entry.final_entry = true;
+        message_queue.Push(entry);
         backend_thread.join();
     }
 
-    std::atomic_bool running{true};
-    std::mutex message_mutex, writing_mutex;
-    std::condition_variable message_cv;
+    Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
+                      const char* function, std::string message) const {
+        using std::chrono::duration_cast;
+        using std::chrono::steady_clock;
+
+        Entry entry;
+        entry.timestamp =
+            duration_cast<std::chrono::microseconds>(steady_clock::now() - time_origin);
+        entry.log_class = log_class;
+        entry.log_level = log_level;
+        entry.filename = Common::TrimSourcePath(filename);
+        entry.line_num = line_nr;
+        entry.function = function;
+        entry.message = std::move(message);
+
+        return entry;
+    }
+
+    std::mutex writing_mutex;
     std::thread backend_thread;
     std::vector<std::unique_ptr<Backend>> backends;
     Common::MPSCQueue<Log::Entry> message_queue;
     Filter filter;
+    std::chrono::steady_clock::time_point time_origin{std::chrono::steady_clock::now()};
 };
 
 void ConsoleBackend::Write(const Entry& entry) {
@@ -232,6 +248,7 @@ void DebuggerBackend::Write(const Entry& entry) {
     CLS(Render)                                                                                    \
     SUB(Render, Software)                                                                          \
     SUB(Render, OpenGL)                                                                            \
+    SUB(Render, Vulkan)                                                                            \
     CLS(Audio)                                                                                     \
     SUB(Audio, DSP)                                                                                \
     SUB(Audio, Sink)                                                                               \
@@ -275,25 +292,6 @@ const char* GetLevelName(Level log_level) {
 #undef LVL
 }
 
-Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
-                  const char* function, std::string message) {
-    using std::chrono::duration_cast;
-    using std::chrono::steady_clock;
-
-    static steady_clock::time_point time_origin = steady_clock::now();
-
-    Entry entry;
-    entry.timestamp = duration_cast<std::chrono::microseconds>(steady_clock::now() - time_origin);
-    entry.log_class = log_class;
-    entry.log_level = log_level;
-    entry.filename = Common::TrimSourcePath(filename);
-    entry.line_num = line_nr;
-    entry.function = function;
-    entry.message = std::move(message);
-
-    return entry;
-}
-
 void SetGlobalFilter(const Filter& filter) {
     Impl::Instance().SetGlobalFilter(filter);
 }
@@ -318,9 +316,7 @@ void FmtLogMessageImpl(Class log_class, Level log_level, const char* filename,
     if (!filter.CheckMessage(log_class, log_level))
         return;
 
-    Entry entry =
-        CreateEntry(log_class, log_level, filename, line_num, function, fmt::vformat(format, args));
-
-    instance.PushEntry(std::move(entry));
+    instance.PushEntry(log_class, log_level, filename, line_num, function,
+                       fmt::vformat(format, args));
 }
 } // namespace Log
diff --git a/src/common/logging/backend.h b/src/common/logging/backend.h
index 91bb0c309..fca0267a1 100644
--- a/src/common/logging/backend.h
+++ b/src/common/logging/backend.h
@@ -27,6 +27,7 @@ struct Entry {
     unsigned int line_num;
     std::string function;
     std::string message;
+    bool final_entry = false;
 
     Entry() = default;
     Entry(Entry&& o) = default;
@@ -134,10 +135,6 @@ const char* GetLogClassName(Class log_class);
  */
 const char* GetLevelName(Level log_level);
 
-/// Creates a log entry by formatting the given source location, and message.
-Entry CreateEntry(Class log_class, Level log_level, const char* filename, unsigned int line_nr,
-                  const char* function, std::string message);
-
 /**
  * The global filter will prevent any messages from even being processed if they are filtered. Each
  * backend can have a filter, but if the level is lower than the global filter, the backend will
diff --git a/src/common/logging/log.h b/src/common/logging/log.h
index d4ec31ec3..8ed6d5050 100644
--- a/src/common/logging/log.h
+++ b/src/common/logging/log.h
@@ -112,6 +112,7 @@ enum class Class : ClassType {
     Render,            ///< Emulator video output and hardware acceleration
     Render_Software,   ///< Software renderer backend
     Render_OpenGL,     ///< OpenGL backend
+    Render_Vulkan,     ///< Vulkan backend
     Audio,             ///< Audio emulation
     Audio_DSP,         ///< The HLE implementation of the DSP
     Audio_Sink,        ///< Emulator audio output backend
diff --git a/src/common/lz4_compression.cpp b/src/common/lz4_compression.cpp
new file mode 100644
index 000000000..dc9b4a916
--- /dev/null
+++ b/src/common/lz4_compression.cpp
@@ -0,0 +1,78 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <algorithm>
+#include <lz4hc.h>
+
+#include "common/assert.h"
+#include "common/lz4_compression.h"
+
+namespace Common::Compression {
+
+std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size) {
+    ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size");
+
+    const auto source_size_int = static_cast<int>(source_size);
+    const int max_compressed_size = LZ4_compressBound(source_size_int);
+    std::vector<u8> compressed(max_compressed_size);
+
+    const int compressed_size = LZ4_compress_default(reinterpret_cast<const char*>(source),
+                                                     reinterpret_cast<char*>(compressed.data()),
+                                                     source_size_int, max_compressed_size);
+
+    if (compressed_size <= 0) {
+        // Compression failed
+        return {};
+    }
+
+    compressed.resize(compressed_size);
+
+    return compressed;
+}
+
+std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size,
+                                  s32 compression_level) {
+    ASSERT_MSG(source_size <= LZ4_MAX_INPUT_SIZE, "Source size exceeds LZ4 maximum input size");
+
+    compression_level = std::clamp(compression_level, LZ4HC_CLEVEL_MIN, LZ4HC_CLEVEL_MAX);
+
+    const auto source_size_int = static_cast<int>(source_size);
+    const int max_compressed_size = LZ4_compressBound(source_size_int);
+    std::vector<u8> compressed(max_compressed_size);
+
+    const int compressed_size = LZ4_compress_HC(
+        reinterpret_cast<const char*>(source), reinterpret_cast<char*>(compressed.data()),
+        source_size_int, max_compressed_size, compression_level);
+
+    if (compressed_size <= 0) {
+        // Compression failed
+        return {};
+    }
+
+    compressed.resize(compressed_size);
+
+    return compressed;
+}
+
+std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size) {
+    return CompressDataLZ4HC(source, source_size, LZ4HC_CLEVEL_MAX);
+}
+
+std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed,
+                                  std::size_t uncompressed_size) {
+    std::vector<u8> uncompressed(uncompressed_size);
+    const int size_check = LZ4_decompress_safe(reinterpret_cast<const char*>(compressed.data()),
+                                               reinterpret_cast<char*>(uncompressed.data()),
+                                               static_cast<int>(compressed.size()),
+                                               static_cast<int>(uncompressed.size()));
+    if (static_cast<int>(uncompressed_size) != size_check) {
+        // Decompression failed
+        return {};
+    }
+    return uncompressed;
+}
+
+} // namespace Common::Compression
diff --git a/src/common/lz4_compression.h b/src/common/lz4_compression.h
new file mode 100644
index 000000000..fe2231a6c
--- /dev/null
+++ b/src/common/lz4_compression.h
@@ -0,0 +1,55 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <vector>
+
+#include "common/common_types.h"
+
+namespace Common::Compression {
+
+/**
+ * Compresses a source memory region with LZ4 and returns the compressed data in a vector.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataLZ4(const u8* source, std::size_t source_size);
+
+/**
+ * Utilizes the LZ4 subalgorithm LZ4HC with the specified compression level. Higher compression
+ * levels result in a smaller compressed size, but require more CPU time for compression. The
+ * compression level has almost no impact on decompression speed. Data compressed with LZ4HC can
+ * also be decompressed with the default LZ4 decompression.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ * @param compression_level the used compression level. Should be between 3 and 12.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataLZ4HC(const u8* source, std::size_t source_size, s32 compression_level);
+
+/**
+ * Utilizes the LZ4 subalgorithm LZ4HC with the highest possible compression level.
+ *
+ * @param source the uncompressed source memory region.
+ * @param source_size the size in bytes of the uncompressed source memory region.
+ *
+ * @return the compressed data.
+ */
+std::vector<u8> CompressDataLZ4HCMax(const u8* source, std::size_t source_size);
+
+/**
+ * Decompresses a source memory region with LZ4 and returns the uncompressed data in a vector.
+ *
+ * @param compressed the compressed source memory region.
+ * @param uncompressed_size the size in bytes of the uncompressed data.
+ *
+ * @return the decompressed data.
+ */
+std::vector<u8> DecompressDataLZ4(const std::vector<u8>& compressed, std::size_t uncompressed_size);
+
+} // namespace Common::Compression
\ No newline at end of file
diff --git a/src/common/math_util.h b/src/common/math_util.h
index 94b4394c5..cff3d48c5 100644
--- a/src/common/math_util.h
+++ b/src/common/math_util.h
@@ -7,7 +7,7 @@
 #include <cstdlib>
 #include <type_traits>
 
-namespace MathUtil {
+namespace Common {
 
 constexpr float PI = 3.14159265f;
 
@@ -41,4 +41,4 @@ struct Rectangle {
     }
 };
 
-} // namespace MathUtil
+} // namespace Common
diff --git a/src/common/memory_hook.cpp b/src/common/memory_hook.cpp
new file mode 100644
index 000000000..3986986d6
--- /dev/null
+++ b/src/common/memory_hook.cpp
@@ -0,0 +1,11 @@
+// Copyright 2018 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/memory_hook.h"
+
+namespace Common {
+
+MemoryHook::~MemoryHook() = default;
+
+} // namespace Common
diff --git a/src/common/memory_hook.h b/src/common/memory_hook.h
new file mode 100644
index 000000000..adaa4c2c5
--- /dev/null
+++ b/src/common/memory_hook.h
@@ -0,0 +1,47 @@
+// Copyright 2016 Citra Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <memory>
+#include <optional>
+
+#include "common/common_types.h"
+
+namespace Common {
+
+/**
+ * Memory hooks have two purposes:
+ * 1. To allow reads and writes to a region of memory to be intercepted. This is used to implement
+ *    texture forwarding and memory breakpoints for debugging.
+ * 2. To allow for the implementation of MMIO devices.
+ *
+ * A hook may be mapped to multiple regions of memory.
+ *
+ * If a std::nullopt or false is returned from a function, the read/write request is passed through
+ * to the underlying memory region.
+ */
+class MemoryHook {
+public:
+    virtual ~MemoryHook();
+
+    virtual std::optional<bool> IsValidAddress(VAddr addr) = 0;
+
+    virtual std::optional<u8> Read8(VAddr addr) = 0;
+    virtual std::optional<u16> Read16(VAddr addr) = 0;
+    virtual std::optional<u32> Read32(VAddr addr) = 0;
+    virtual std::optional<u64> Read64(VAddr addr) = 0;
+
+    virtual bool ReadBlock(VAddr src_addr, void* dest_buffer, std::size_t size) = 0;
+
+    virtual bool Write8(VAddr addr, u8 data) = 0;
+    virtual bool Write16(VAddr addr, u16 data) = 0;
+    virtual bool Write32(VAddr addr, u32 data) = 0;
+    virtual bool Write64(VAddr addr, u64 data) = 0;
+
+    virtual bool WriteBlock(VAddr dest_addr, const void* src_buffer, std::size_t size) = 0;
+};
+
+using MemoryHookPointer = std::shared_ptr<MemoryHook>;
+} // namespace Common
diff --git a/src/common/multi_level_queue.h b/src/common/multi_level_queue.h
new file mode 100644
index 000000000..2b61b91e0
--- /dev/null
+++ b/src/common/multi_level_queue.h
@@ -0,0 +1,337 @@
+// Copyright 2019 TuxSH
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <array>
+#include <iterator>
+#include <list>
+#include <utility>
+
+#include "common/bit_util.h"
+#include "common/common_types.h"
+
+namespace Common {
+
+/**
+ * A MultiLevelQueue is a type of priority queue which has the following characteristics:
+ * - iteratable through each of its elements.
+ * - back can be obtained.
+ * - O(1) add, lookup (both front and back)
+ * - discrete priorities and a max of 64 priorities (limited domain)
+ * This type of priority queue is normaly used for managing threads within an scheduler
+ */
+template <typename T, std::size_t Depth>
+class MultiLevelQueue {
+public:
+    using value_type = T;
+    using reference = value_type&;
+    using const_reference = const value_type&;
+    using pointer = value_type*;
+    using const_pointer = const value_type*;
+
+    using difference_type = typename std::pointer_traits<pointer>::difference_type;
+    using size_type = std::size_t;
+
+    template <bool is_constant>
+    class iterator_impl {
+    public:
+        using iterator_category = std::bidirectional_iterator_tag;
+        using value_type = T;
+        using pointer = std::conditional_t<is_constant, T*, const T*>;
+        using reference = std::conditional_t<is_constant, const T&, T&>;
+        using difference_type = typename std::pointer_traits<pointer>::difference_type;
+
+        friend bool operator==(const iterator_impl& lhs, const iterator_impl& rhs) {
+            if (lhs.IsEnd() && rhs.IsEnd())
+                return true;
+            return std::tie(lhs.current_priority, lhs.it) == std::tie(rhs.current_priority, rhs.it);
+        }
+
+        friend bool operator!=(const iterator_impl& lhs, const iterator_impl& rhs) {
+            return !operator==(lhs, rhs);
+        }
+
+        reference operator*() const {
+            return *it;
+        }
+
+        pointer operator->() const {
+            return it.operator->();
+        }
+
+        iterator_impl& operator++() {
+            if (IsEnd()) {
+                return *this;
+            }
+
+            ++it;
+
+            if (it == GetEndItForPrio()) {
+                u64 prios = mlq.used_priorities;
+                prios &= ~((1ULL << (current_priority + 1)) - 1);
+                if (prios == 0) {
+                    current_priority = mlq.depth();
+                } else {
+                    current_priority = CountTrailingZeroes64(prios);
+                    it = GetBeginItForPrio();
+                }
+            }
+            return *this;
+        }
+
+        iterator_impl& operator--() {
+            if (IsEnd()) {
+                if (mlq.used_priorities != 0) {
+                    current_priority = 63 - CountLeadingZeroes64(mlq.used_priorities);
+                    it = GetEndItForPrio();
+                    --it;
+                }
+            } else if (it == GetBeginItForPrio()) {
+                u64 prios = mlq.used_priorities;
+                prios &= (1ULL << current_priority) - 1;
+                if (prios != 0) {
+                    current_priority = CountTrailingZeroes64(prios);
+                    it = GetEndItForPrio();
+                    --it;
+                }
+            } else {
+                --it;
+            }
+            return *this;
+        }
+
+        iterator_impl operator++(int) {
+            const iterator_impl v{*this};
+            ++(*this);
+            return v;
+        }
+
+        iterator_impl operator--(int) {
+            const iterator_impl v{*this};
+            --(*this);
+            return v;
+        }
+
+        // allow implicit const->non-const
+        iterator_impl(const iterator_impl<false>& other)
+            : mlq(other.mlq), it(other.it), current_priority(other.current_priority) {}
+
+        iterator_impl(const iterator_impl<true>& other)
+            : mlq(other.mlq), it(other.it), current_priority(other.current_priority) {}
+
+        iterator_impl& operator=(const iterator_impl<false>& other) {
+            mlq = other.mlq;
+            it = other.it;
+            current_priority = other.current_priority;
+            return *this;
+        }
+
+        friend class iterator_impl<true>;
+        iterator_impl() = default;
+
+    private:
+        friend class MultiLevelQueue;
+        using container_ref =
+            std::conditional_t<is_constant, const MultiLevelQueue&, MultiLevelQueue&>;
+        using list_iterator = std::conditional_t<is_constant, typename std::list<T>::const_iterator,
+                                                 typename std::list<T>::iterator>;
+
+        explicit iterator_impl(container_ref mlq, list_iterator it, u32 current_priority)
+            : mlq(mlq), it(it), current_priority(current_priority) {}
+        explicit iterator_impl(container_ref mlq, u32 current_priority)
+            : mlq(mlq), it(), current_priority(current_priority) {}
+
+        bool IsEnd() const {
+            return current_priority == mlq.depth();
+        }
+
+        list_iterator GetBeginItForPrio() const {
+            return mlq.levels[current_priority].begin();
+        }
+
+        list_iterator GetEndItForPrio() const {
+            return mlq.levels[current_priority].end();
+        }
+
+        container_ref mlq;
+        list_iterator it;
+        u32 current_priority;
+    };
+
+    using iterator = iterator_impl<false>;
+    using const_iterator = iterator_impl<true>;
+
+    void add(const T& element, u32 priority, bool send_back = true) {
+        if (send_back)
+            levels[priority].push_back(element);
+        else
+            levels[priority].push_front(element);
+        used_priorities |= 1ULL << priority;
+    }
+
+    void remove(const T& element, u32 priority) {
+        auto it = ListIterateTo(levels[priority], element);
+        if (it == levels[priority].end())
+            return;
+        levels[priority].erase(it);
+        if (levels[priority].empty()) {
+            used_priorities &= ~(1ULL << priority);
+        }
+    }
+
+    void adjust(const T& element, u32 old_priority, u32 new_priority, bool adjust_front = false) {
+        remove(element, old_priority);
+        add(element, new_priority, !adjust_front);
+    }
+    void adjust(const_iterator it, u32 old_priority, u32 new_priority, bool adjust_front = false) {
+        adjust(*it, old_priority, new_priority, adjust_front);
+    }
+
+    void transfer_to_front(const T& element, u32 priority, MultiLevelQueue& other) {
+        ListSplice(other.levels[priority], other.levels[priority].begin(), levels[priority],
+                   ListIterateTo(levels[priority], element));
+
+        other.used_priorities |= 1ULL << priority;
+
+        if (levels[priority].empty()) {
+            used_priorities &= ~(1ULL << priority);
+        }
+    }
+
+    void transfer_to_front(const_iterator it, u32 priority, MultiLevelQueue& other) {
+        transfer_to_front(*it, priority, other);
+    }
+
+    void transfer_to_back(const T& element, u32 priority, MultiLevelQueue& other) {
+        ListSplice(other.levels[priority], other.levels[priority].end(), levels[priority],
+                   ListIterateTo(levels[priority], element));
+
+        other.used_priorities |= 1ULL << priority;
+
+        if (levels[priority].empty()) {
+            used_priorities &= ~(1ULL << priority);
+        }
+    }
+
+    void transfer_to_back(const_iterator it, u32 priority, MultiLevelQueue& other) {
+        transfer_to_back(*it, priority, other);
+    }
+
+    void yield(u32 priority, std::size_t n = 1) {
+        ListShiftForward(levels[priority], n);
+    }
+
+    std::size_t depth() const {
+        return Depth;
+    }
+
+    std::size_t size(u32 priority) const {
+        return levels[priority].size();
+    }
+
+    std::size_t size() const {
+        u64 priorities = used_priorities;
+        std::size_t size = 0;
+        while (priorities != 0) {
+            const u64 current_priority = CountTrailingZeroes64(priorities);
+            size += levels[current_priority].size();
+            priorities &= ~(1ULL << current_priority);
+        }
+        return size;
+    }
+
+    bool empty() const {
+        return used_priorities == 0;
+    }
+
+    bool empty(u32 priority) const {
+        return (used_priorities & (1ULL << priority)) == 0;
+    }
+
+    u32 highest_priority_set(u32 max_priority = 0) const {
+        const u64 priorities =
+            max_priority == 0 ? used_priorities : (used_priorities & ~((1ULL << max_priority) - 1));
+        return priorities == 0 ? Depth : static_cast<u32>(CountTrailingZeroes64(priorities));
+    }
+
+    u32 lowest_priority_set(u32 min_priority = Depth - 1) const {
+        const u64 priorities = min_priority >= Depth - 1
+                                   ? used_priorities
+                                   : (used_priorities & ((1ULL << (min_priority + 1)) - 1));
+        return priorities == 0 ? Depth : 63 - CountLeadingZeroes64(priorities);
+    }
+
+    const_iterator cbegin(u32 max_prio = 0) const {
+        const u32 priority = highest_priority_set(max_prio);
+        return priority == Depth ? cend()
+                                 : const_iterator{*this, levels[priority].cbegin(), priority};
+    }
+    const_iterator begin(u32 max_prio = 0) const {
+        return cbegin(max_prio);
+    }
+    iterator begin(u32 max_prio = 0) {
+        const u32 priority = highest_priority_set(max_prio);
+        return priority == Depth ? end() : iterator{*this, levels[priority].begin(), priority};
+    }
+
+    const_iterator cend(u32 min_prio = Depth - 1) const {
+        return min_prio == Depth - 1 ? const_iterator{*this, Depth} : cbegin(min_prio + 1);
+    }
+    const_iterator end(u32 min_prio = Depth - 1) const {
+        return cend(min_prio);
+    }
+    iterator end(u32 min_prio = Depth - 1) {
+        return min_prio == Depth - 1 ? iterator{*this, Depth} : begin(min_prio + 1);
+    }
+
+    T& front(u32 max_priority = 0) {
+        const u32 priority = highest_priority_set(max_priority);
+        return levels[priority == Depth ? 0 : priority].front();
+    }
+    const T& front(u32 max_priority = 0) const {
+        const u32 priority = highest_priority_set(max_priority);
+        return levels[priority == Depth ? 0 : priority].front();
+    }
+
+    T back(u32 min_priority = Depth - 1) {
+        const u32 priority = lowest_priority_set(min_priority); // intended
+        return levels[priority == Depth ? 63 : priority].back();
+    }
+    const T& back(u32 min_priority = Depth - 1) const {
+        const u32 priority = lowest_priority_set(min_priority); // intended
+        return levels[priority == Depth ? 63 : priority].back();
+    }
+
+private:
+    using const_list_iterator = typename std::list<T>::const_iterator;
+
+    static void ListShiftForward(std::list<T>& list, const std::size_t shift = 1) {
+        if (shift >= list.size()) {
+            return;
+        }
+
+        const auto begin_range = list.begin();
+        const auto end_range = std::next(begin_range, shift);
+        list.splice(list.end(), list, begin_range, end_range);
+    }
+
+    static void ListSplice(std::list<T>& in_list, const_list_iterator position,
+                           std::list<T>& out_list, const_list_iterator element) {
+        in_list.splice(position, out_list, element);
+    }
+
+    static const_list_iterator ListIterateTo(const std::list<T>& list, const T& element) {
+        auto it = list.cbegin();
+        while (it != list.cend() && *it != element) {
+            ++it;
+        }
+        return it;
+    }
+
+    std::array<std::list<T>, Depth> levels;
+    u64 used_priorities = 0;
+};
+
+} // namespace Common
diff --git a/src/common/page_table.cpp b/src/common/page_table.cpp
new file mode 100644
index 000000000..69b7abc54
--- /dev/null
+++ b/src/common/page_table.cpp
@@ -0,0 +1,31 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include "common/page_table.h"
+
+namespace Common {
+
+PageTable::PageTable(std::size_t page_size_in_bits) : page_size_in_bits{page_size_in_bits} {}
+
+PageTable::~PageTable() = default;
+
+void PageTable::Resize(std::size_t address_space_width_in_bits) {
+    const std::size_t num_page_table_entries = 1ULL
+                                               << (address_space_width_in_bits - page_size_in_bits);
+
+    pointers.resize(num_page_table_entries);
+    attributes.resize(num_page_table_entries);
+    backing_addr.resize(num_page_table_entries);
+
+    // The default is a 39-bit address space, which causes an initial 1GB allocation size. If the
+    // vector size is subsequently decreased (via resize), the vector might not automatically
+    // actually reallocate/resize its underlying allocation, which wastes up to ~800 MB for
+    // 36-bit titles. Call shrink_to_fit to reduce capacity to what's actually in use.
+
+    pointers.shrink_to_fit();
+    attributes.shrink_to_fit();
+    backing_addr.shrink_to_fit();
+}
+
+} // namespace Common
diff --git a/src/common/page_table.h b/src/common/page_table.h
new file mode 100644
index 000000000..8b8ff0bb8
--- /dev/null
+++ b/src/common/page_table.h
@@ -0,0 +1,84 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <vector>
+#include <boost/icl/interval_map.hpp>
+#include "common/common_types.h"
+#include "common/memory_hook.h"
+
+namespace Common {
+
+enum class PageType : u8 {
+    /// Page is unmapped and should cause an access error.
+    Unmapped,
+    /// Page is mapped to regular memory. This is the only type you can get pointers to.
+    Memory,
+    /// Page is mapped to regular memory, but also needs to check for rasterizer cache flushing and
+    /// invalidation
+    RasterizerCachedMemory,
+    /// Page is mapped to a I/O region. Writing and reading to this page is handled by functions.
+    Special,
+    /// Page is allocated for use.
+    Allocated,
+};
+
+struct SpecialRegion {
+    enum class Type {
+        DebugHook,
+        IODevice,
+    } type;
+
+    MemoryHookPointer handler;
+
+    bool operator<(const SpecialRegion& other) const {
+        return std::tie(type, handler) < std::tie(other.type, other.handler);
+    }
+
+    bool operator==(const SpecialRegion& other) const {
+        return std::tie(type, handler) == std::tie(other.type, other.handler);
+    }
+};
+
+/**
+ * A (reasonably) fast way of allowing switchable and remappable process address spaces. It loosely
+ * mimics the way a real CPU page table works.
+ */
+struct PageTable {
+    explicit PageTable(std::size_t page_size_in_bits);
+    ~PageTable();
+
+    /**
+     * Resizes the page table to be able to accomodate enough pages within
+     * a given address space.
+     *
+     * @param address_space_width_in_bits The address size width in bits.
+     */
+    void Resize(std::size_t address_space_width_in_bits);
+
+    /**
+     * Vector of memory pointers backing each page. An entry can only be non-null if the
+     * corresponding entry in the `attributes` vector is of type `Memory`.
+     */
+    std::vector<u8*> pointers;
+
+    /**
+     * Contains MMIO handlers that back memory regions whose entries in the `attribute` vector is
+     * of type `Special`.
+     */
+    boost::icl::interval_map<u64, std::set<SpecialRegion>> special_regions;
+
+    /**
+     * Vector of fine grained page attributes. If it is set to any value other than `Memory`, then
+     * the corresponding entry in `pointers` MUST be set to null.
+     */
+    std::vector<PageType> attributes;
+
+    std::vector<u64> backing_addr;
+
+    const std::size_t page_size_in_bits{};
+};
+
+} // namespace Common
diff --git a/src/common/quaternion.h b/src/common/quaternion.h
index c528c0b68..370198ae0 100644
--- a/src/common/quaternion.h
+++ b/src/common/quaternion.h
@@ -6,12 +6,12 @@
 
 #include "common/vector_math.h"
 
-namespace Math {
+namespace Common {
 
 template <typename T>
 class Quaternion {
 public:
-    Math::Vec3<T> xyz;
+    Vec3<T> xyz;
     T w{};
 
     Quaternion<decltype(-T{})> Inverse() const {
@@ -38,12 +38,12 @@ public:
 };
 
 template <typename T>
-auto QuaternionRotate(const Quaternion<T>& q, const Math::Vec3<T>& v) {
+auto QuaternionRotate(const Quaternion<T>& q, const Vec3<T>& v) {
     return v + 2 * Cross(q.xyz, Cross(q.xyz, v) + v * q.w);
 }
 
-inline Quaternion<float> MakeQuaternion(const Math::Vec3<float>& axis, float angle) {
+inline Quaternion<float> MakeQuaternion(const Vec3<float>& axis, float angle) {
     return {axis * std::sin(angle / 2), std::cos(angle / 2)};
 }
 
-} // namespace Math
+} // namespace Common
diff --git a/src/common/swap.h b/src/common/swap.h
index 32af0b6ac..b3eab1324 100644
--- a/src/common/swap.h
+++ b/src/common/swap.h
@@ -17,6 +17,8 @@
 
 #pragma once
 
+#include <type_traits>
+
 #if defined(_MSC_VER)
 #include <cstdlib>
 #elif defined(__linux__)
@@ -28,8 +30,8 @@
 #include <cstring>
 #include "common/common_types.h"
 
-// GCC 4.6+
-#if __GNUC__ >= 5 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 6)
+// GCC
+#ifdef __GNUC__
 
 #if __BYTE_ORDER__ && (__BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) && !defined(COMMON_LITTLE_ENDIAN)
 #define COMMON_LITTLE_ENDIAN 1
@@ -38,7 +40,7 @@
 #endif
 
 // LLVM/clang
-#elif __clang__
+#elif defined(__clang__)
 
 #if __LITTLE_ENDIAN__ && !defined(COMMON_LITTLE_ENDIAN)
 #define COMMON_LITTLE_ENDIAN 1
@@ -170,7 +172,7 @@ struct swap_struct_t {
     using swapped_t = swap_struct_t;
 
 protected:
-    T value = T();
+    T value;
 
     static T swap(T v) {
         return F::swap(v);
@@ -605,52 +607,154 @@ struct swap_double_t {
     }
 };
 
-#if COMMON_LITTLE_ENDIAN
-using u16_le = u16;
-using u32_le = u32;
-using u64_le = u64;
+template <typename T>
+struct swap_enum_t {
+    static_assert(std::is_enum_v<T>);
+    using base = std::underlying_type_t<T>;
+
+public:
+    swap_enum_t() = default;
+    swap_enum_t(const T& v) : value(swap(v)) {}
+
+    swap_enum_t& operator=(const T& v) {
+        value = swap(v);
+        return *this;
+    }
+
+    operator T() const {
+        return swap(value);
+    }
+
+    explicit operator base() const {
+        return static_cast<base>(swap(value));
+    }
 
-using s16_le = s16;
-using s32_le = s32;
-using s64_le = s64;
+protected:
+    T value{};
+    // clang-format off
+    using swap_t = std::conditional_t<
+        std::is_same_v<base, u16>, swap_16_t<u16>, std::conditional_t<
+        std::is_same_v<base, s16>, swap_16_t<s16>, std::conditional_t<
+        std::is_same_v<base, u32>, swap_32_t<u32>, std::conditional_t<
+        std::is_same_v<base, s32>, swap_32_t<s32>, std::conditional_t<
+        std::is_same_v<base, u64>, swap_64_t<u64>, std::conditional_t<
+        std::is_same_v<base, s64>, swap_64_t<s64>, void>>>>>>;
+    // clang-format on
+    static T swap(T x) {
+        return static_cast<T>(swap_t::swap(static_cast<base>(x)));
+    }
+};
 
-using float_le = float;
-using double_le = double;
+struct SwapTag {}; // Use the different endianness from the system
+struct KeepTag {}; // Use the same endianness as the system
 
-using u64_be = swap_struct_t<u64, swap_64_t<u64>>;
-using s64_be = swap_struct_t<s64, swap_64_t<s64>>;
+template <typename T, typename Tag>
+struct AddEndian;
 
-using u32_be = swap_struct_t<u32, swap_32_t<u32>>;
-using s32_be = swap_struct_t<s32, swap_32_t<s32>>;
+// KeepTag specializations
 
-using u16_be = swap_struct_t<u16, swap_16_t<u16>>;
-using s16_be = swap_struct_t<s16, swap_16_t<s16>>;
+template <typename T>
+struct AddEndian<T, KeepTag> {
+    using type = T;
+};
 
-using float_be = swap_struct_t<float, swap_float_t<float>>;
-using double_be = swap_struct_t<double, swap_double_t<double>>;
-#else
+// SwapTag specializations
+
+template <>
+struct AddEndian<u8, SwapTag> {
+    using type = u8;
+};
+
+template <>
+struct AddEndian<u16, SwapTag> {
+    using type = swap_struct_t<u16, swap_16_t<u16>>;
+};
+
+template <>
+struct AddEndian<u32, SwapTag> {
+    using type = swap_struct_t<u32, swap_32_t<u32>>;
+};
 
-using u64_le = swap_struct_t<u64, swap_64_t<u64>>;
-using s64_le = swap_struct_t<s64, swap_64_t<s64>>;
+template <>
+struct AddEndian<u64, SwapTag> {
+    using type = swap_struct_t<u64, swap_64_t<u64>>;
+};
+
+template <>
+struct AddEndian<s8, SwapTag> {
+    using type = s8;
+};
 
-using u32_le = swap_struct_t<u32, swap_32_t<u32>>;
-using s32_le = swap_struct_t<s32, swap_32_t<s32>>;
+template <>
+struct AddEndian<s16, SwapTag> {
+    using type = swap_struct_t<s16, swap_16_t<s16>>;
+};
 
-using u16_le = swap_struct_t<u16, swap_16_t<u16>>;
-using s16_le = swap_struct_t<s16, swap_16_t<s16>>;
+template <>
+struct AddEndian<s32, SwapTag> {
+    using type = swap_struct_t<s32, swap_32_t<s32>>;
+};
+
+template <>
+struct AddEndian<s64, SwapTag> {
+    using type = swap_struct_t<s64, swap_64_t<s64>>;
+};
+
+template <>
+struct AddEndian<float, SwapTag> {
+    using type = swap_struct_t<float, swap_float_t<float>>;
+};
+
+template <>
+struct AddEndian<double, SwapTag> {
+    using type = swap_struct_t<double, swap_double_t<double>>;
+};
+
+template <typename T>
+struct AddEndian<T, SwapTag> {
+    static_assert(std::is_enum_v<T>);
+    using type = swap_enum_t<T>;
+};
 
-using float_le = swap_struct_t<float, swap_float_t<float>>;
-using double_le = swap_struct_t<double, swap_double_t<double>>;
+// Alias LETag/BETag as KeepTag/SwapTag depending on the system
+#if COMMON_LITTLE_ENDIAN
 
-using u16_be = u16;
-using u32_be = u32;
-using u64_be = u64;
+using LETag = KeepTag;
+using BETag = SwapTag;
 
-using s16_be = s16;
-using s32_be = s32;
-using s64_be = s64;
+#else
 
-using float_be = float;
-using double_be = double;
+using BETag = KeepTag;
+using LETag = SwapTag;
 
 #endif
+
+// Aliases for LE types
+using u16_le = AddEndian<u16, LETag>::type;
+using u32_le = AddEndian<u32, LETag>::type;
+using u64_le = AddEndian<u64, LETag>::type;
+
+using s16_le = AddEndian<s16, LETag>::type;
+using s32_le = AddEndian<s32, LETag>::type;
+using s64_le = AddEndian<s64, LETag>::type;
+
+template <typename T>
+using enum_le = std::enable_if_t<std::is_enum_v<T>, typename AddEndian<T, LETag>::type>;
+
+using float_le = AddEndian<float, LETag>::type;
+using double_le = AddEndian<double, LETag>::type;
+
+// Aliases for BE types
+using u16_be = AddEndian<u16, BETag>::type;
+using u32_be = AddEndian<u32, BETag>::type;
+using u64_be = AddEndian<u64, BETag>::type;
+
+using s16_be = AddEndian<s16, BETag>::type;
+using s32_be = AddEndian<s32, BETag>::type;
+using s64_be = AddEndian<s64, BETag>::type;
+
+template <typename T>
+using enum_be = std::enable_if_t<std::is_enum_v<T>, typename AddEndian<T, BETag>::type>;
+
+using float_be = AddEndian<float, BETag>::type;
+using double_be = AddEndian<double, BETag>::type;
diff --git a/src/common/thread.cpp b/src/common/thread.cpp
index 5144c0d9f..fe7a420cc 100644
--- a/src/common/thread.cpp
+++ b/src/common/thread.cpp
@@ -27,18 +27,6 @@ namespace Common {
 
 #ifdef _MSC_VER
 
-void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask) {
-    SetThreadAffinityMask(thread, mask);
-}
-
-void SetCurrentThreadAffinity(u32 mask) {
-    SetThreadAffinityMask(GetCurrentThread(), mask);
-}
-
-void SwitchCurrentThread() {
-    SwitchToThread();
-}
-
 // Sets the debugger-visible name of the current thread.
 // Uses undocumented (actually, it is now documented) trick.
 // http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vsdebug/html/vxtsksettingthreadname.asp
@@ -70,31 +58,6 @@ void SetCurrentThreadName(const char* name) {
 
 #else // !MSVC_VER, so must be POSIX threads
 
-void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask) {
-#ifdef __APPLE__
-    thread_policy_set(pthread_mach_thread_np(thread), THREAD_AFFINITY_POLICY, (integer_t*)&mask, 1);
-#elif (defined __linux__ || defined __FreeBSD__) && !(defined ANDROID)
-    cpu_set_t cpu_set;
-    CPU_ZERO(&cpu_set);
-
-    for (int i = 0; i != sizeof(mask) * 8; ++i)
-        if ((mask >> i) & 1)
-            CPU_SET(i, &cpu_set);
-
-    pthread_setaffinity_np(thread, sizeof(cpu_set), &cpu_set);
-#endif
-}
-
-void SetCurrentThreadAffinity(u32 mask) {
-    SetThreadAffinity(pthread_self(), mask);
-}
-
-#ifndef _WIN32
-void SwitchCurrentThread() {
-    usleep(1000 * 1);
-}
-#endif
-
 // MinGW with the POSIX threading model does not support pthread_setname_np
 #if !defined(_WIN32) || defined(_MSC_VER)
 void SetCurrentThreadName(const char* name) {
diff --git a/src/common/thread.h b/src/common/thread.h
index 2cf74452d..0cfd98be6 100644
--- a/src/common/thread.h
+++ b/src/common/thread.h
@@ -9,14 +9,13 @@
 #include <cstddef>
 #include <mutex>
 #include <thread>
-#include "common/common_types.h"
 
 namespace Common {
 
 class Event {
 public:
     void Set() {
-        std::lock_guard<std::mutex> lk(mutex);
+        std::lock_guard lk{mutex};
         if (!is_set) {
             is_set = true;
             condvar.notify_one();
@@ -24,14 +23,14 @@ public:
     }
 
     void Wait() {
-        std::unique_lock<std::mutex> lk(mutex);
+        std::unique_lock lk{mutex};
         condvar.wait(lk, [&] { return is_set; });
         is_set = false;
     }
 
     template <class Clock, class Duration>
     bool WaitUntil(const std::chrono::time_point<Clock, Duration>& time) {
-        std::unique_lock<std::mutex> lk(mutex);
+        std::unique_lock lk{mutex};
         if (!condvar.wait_until(lk, time, [this] { return is_set; }))
             return false;
         is_set = false;
@@ -39,7 +38,7 @@ public:
     }
 
     void Reset() {
-        std::unique_lock<std::mutex> lk(mutex);
+        std::unique_lock lk{mutex};
         // no other action required, since wait loops on the predicate and any lingering signal will
         // get cleared on the first iteration
         is_set = false;
@@ -57,7 +56,7 @@ public:
 
     /// Blocks until all "count" threads have called Sync()
     void Sync() {
-        std::unique_lock<std::mutex> lk(mutex);
+        std::unique_lock lk{mutex};
         const std::size_t current_generation = generation;
 
         if (++waiting == count) {
@@ -78,9 +77,6 @@ private:
     std::size_t generation = 0; // Incremented once each time the barrier is used
 };
 
-void SetThreadAffinity(std::thread::native_handle_type thread, u32 mask);
-void SetCurrentThreadAffinity(u32 mask);
-void SwitchCurrentThread(); // On Linux, this is equal to sleep 1ms
 void SetCurrentThreadName(const char* name);
 
 } // namespace Common
diff --git a/src/common/thread_queue_list.h b/src/common/thread_queue_list.h
index e7594db68..791f99a8c 100644
--- a/src/common/thread_queue_list.h
+++ b/src/common/thread_queue_list.h
@@ -6,7 +6,6 @@
 
 #include <array>
 #include <deque>
-#include <boost/range/algorithm_ext/erase.hpp>
 
 namespace Common {
 
@@ -111,8 +110,9 @@ struct ThreadQueueList {
     }
 
     void remove(Priority priority, const T& thread_id) {
-        Queue* cur = &queues[priority];
-        boost::remove_erase(cur->data, thread_id);
+        Queue* const cur = &queues[priority];
+        const auto iter = std::remove(cur->data.begin(), cur->data.end(), thread_id);
+        cur->data.erase(iter, cur->data.end());
     }
 
     void rotate(Priority priority) {
diff --git a/src/common/threadsafe_queue.h b/src/common/threadsafe_queue.h
index edf13bc49..e714ba5b3 100644
--- a/src/common/threadsafe_queue.h
+++ b/src/common/threadsafe_queue.h
@@ -7,17 +7,17 @@
 // a simple lockless thread-safe,
 // single reader, single writer queue
 
-#include <algorithm>
 #include <atomic>
+#include <condition_variable>
 #include <cstddef>
 #include <mutex>
-#include "common/common_types.h"
+#include <utility>
 
 namespace Common {
-template <typename T, bool NeedSize = true>
+template <typename T>
 class SPSCQueue {
 public:
-    SPSCQueue() : size(0) {
+    SPSCQueue() {
         write_ptr = read_ptr = new ElementPtr();
     }
     ~SPSCQueue() {
@@ -25,13 +25,12 @@ public:
         delete read_ptr;
     }
 
-    u32 Size() const {
-        static_assert(NeedSize, "using Size() on FifoQueue without NeedSize");
+    std::size_t Size() const {
         return size.load();
     }
 
     bool Empty() const {
-        return !read_ptr->next.load();
+        return Size() == 0;
     }
 
     T& Front() const {
@@ -47,13 +46,14 @@ public:
         ElementPtr* new_ptr = new ElementPtr();
         write_ptr->next.store(new_ptr, std::memory_order_release);
         write_ptr = new_ptr;
-        if (NeedSize)
-            size++;
+        cv.notify_one();
+
+        ++size;
     }
 
     void Pop() {
-        if (NeedSize)
-            size--;
+        --size;
+
         ElementPtr* tmpptr = read_ptr;
         // advance the read pointer
         read_ptr = tmpptr->next.load();
@@ -66,8 +66,7 @@ public:
         if (Empty())
             return false;
 
-        if (NeedSize)
-            size--;
+        --size;
 
         ElementPtr* tmpptr = read_ptr;
         read_ptr = tmpptr->next.load(std::memory_order_acquire);
@@ -77,6 +76,16 @@ public:
         return true;
     }
 
+    T PopWait() {
+        if (Empty()) {
+            std::unique_lock lock{cv_mutex};
+            cv.wait(lock, [this]() { return !Empty(); });
+        }
+        T t;
+        Pop(t);
+        return t;
+    }
+
     // not thread-safe
     void Clear() {
         size.store(0);
@@ -89,7 +98,7 @@ private:
     // and a pointer to the next ElementPtr
     class ElementPtr {
     public:
-        ElementPtr() : next(nullptr) {}
+        ElementPtr() {}
         ~ElementPtr() {
             ElementPtr* next_ptr = next.load();
 
@@ -98,21 +107,23 @@ private:
         }
 
         T current;
-        std::atomic<ElementPtr*> next;
+        std::atomic<ElementPtr*> next{nullptr};
     };
 
     ElementPtr* write_ptr;
     ElementPtr* read_ptr;
-    std::atomic<u32> size;
+    std::atomic_size_t size{0};
+    std::mutex cv_mutex;
+    std::condition_variable cv;
 };
 
 // a simple thread-safe,
 // single reader, multiple writer queue
 
-template <typename T, bool NeedSize = true>
+template <typename T>
 class MPSCQueue {
 public:
-    u32 Size() const {
+    std::size_t Size() const {
         return spsc_queue.Size();
     }
 
@@ -126,7 +137,7 @@ public:
 
     template <typename Arg>
     void Push(Arg&& t) {
-        std::lock_guard<std::mutex> lock(write_lock);
+        std::lock_guard lock{write_lock};
         spsc_queue.Push(t);
     }
 
@@ -138,13 +149,17 @@ public:
         return spsc_queue.Pop(t);
     }
 
+    T PopWait() {
+        return spsc_queue.PopWait();
+    }
+
     // not thread-safe
     void Clear() {
         spsc_queue.Clear();
     }
 
 private:
-    SPSCQueue<T, NeedSize> spsc_queue;
+    SPSCQueue<T> spsc_queue;
     std::mutex write_lock;
 };
 } // namespace Common
diff --git a/src/common/uint128.cpp b/src/common/uint128.cpp
new file mode 100644
index 000000000..32bf56730
--- /dev/null
+++ b/src/common/uint128.cpp
@@ -0,0 +1,45 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#ifdef _MSC_VER
+#include <intrin.h>
+
+#pragma intrinsic(_umul128)
+#endif
+#include <cstring>
+#include "common/uint128.h"
+
+namespace Common {
+
+u128 Multiply64Into128(u64 a, u64 b) {
+    u128 result;
+#ifdef _MSC_VER
+    result[0] = _umul128(a, b, &result[1]);
+#else
+    unsigned __int128 tmp = a;
+    tmp *= b;
+    std::memcpy(&result, &tmp, sizeof(u128));
+#endif
+    return result;
+}
+
+std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor) {
+    u64 remainder = dividend[0] % divisor;
+    u64 accum = dividend[0] / divisor;
+    if (dividend[1] == 0)
+        return {accum, remainder};
+    // We ignore dividend[1] / divisor as that overflows
+    const u64 first_segment = (dividend[1] % divisor) << 32;
+    accum += (first_segment / divisor) << 32;
+    const u64 second_segment = (first_segment % divisor) << 32;
+    accum += (second_segment / divisor);
+    remainder += second_segment % divisor;
+    if (remainder >= divisor) {
+        accum++;
+        remainder -= divisor;
+    }
+    return {accum, remainder};
+}
+
+} // namespace Common
diff --git a/src/common/uint128.h b/src/common/uint128.h
new file mode 100644
index 000000000..a3be2a2cb
--- /dev/null
+++ b/src/common/uint128.h
@@ -0,0 +1,19 @@
+// Copyright 2019 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#pragma once
+
+#include <utility>
+#include "common/common_types.h"
+
+namespace Common {
+
+// This function multiplies 2 u64 values and produces a u128 value;
+u128 Multiply64Into128(u64 a, u64 b);
+
+// This function divides a u128 by a u32 value and produces two u64 values:
+// the result of division and the remainder
+std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor);
+
+} // namespace Common
diff --git a/src/common/vector_math.h b/src/common/vector_math.h
index 8feb49941..429485329 100644
--- a/src/common/vector_math.h
+++ b/src/common/vector_math.h
@@ -33,7 +33,7 @@
 #include <cmath>
 #include <type_traits>
 
-namespace Math {
+namespace Common {
 
 template <typename T>
 class Vec2;
@@ -690,4 +690,4 @@ constexpr Vec4<T> MakeVec(const T& x, const Vec3<T>& yzw) {
     return MakeVec(x, yzw[0], yzw[1], yzw[2]);
 }
 
-} // namespace Math
+} // namespace Common