diff --git a/CMakeLists.txt b/CMakeLists.txt
index 467d769a2d..83832e6258 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -5,7 +5,10 @@ list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_SOURCE_DIR}/externals/cmake-modul
 include(DownloadExternals)
 include(CMakeDependentOption)
 
-project(yuzu)
+set(CMAKE_OSX_ARCHITECTURES "x86_64")
+set(CMAKE_OSX_DEPLOYMENT_TARGET "10.15.0" CACHE STRING "")
+
+project(yuzu C CXX ASM)
 
 # Set bundled sdl2/qt as dependent options.
 # OFF by default, but if ENABLE_SDL2 and MSVC are true then ON
@@ -29,6 +32,10 @@ option(ENABLE_VULKAN "Enables Vulkan backend" ON)
 
 option(USE_DISCORD_PRESENCE "Enables Discord Rich Presence" OFF)
 
+if(APPLE)
+  option(OSX_USE_DEFAULT_SEARCH_PATH "Don't prioritize system library paths" OFF)
+endif()
+
 if(EXISTS ${PROJECT_SOURCE_DIR}/hooks/pre-commit AND NOT EXISTS ${PROJECT_SOURCE_DIR}/.git/hooks/pre-commit)
     message(STATUS "Copying pre-commit hook")
     file(COPY hooks/pre-commit
@@ -276,6 +283,31 @@ if (APPLE)
     # Umbrella framework for everything GUI-related
     find_library(COCOA_LIBRARY Cocoa)
     set(PLATFORM_LIBRARIES ${COCOA_LIBRARY} ${IOKIT_LIBRARY} ${COREVIDEO_LIBRARY})
+
+    if(NOT OSX_USE_DEFAULT_SEARCH_PATH)
+      # Hack up the path to prioritize the path to built-in OS libraries to
+      # increase the chance of not depending on a bunch of copies of them
+      # installed by MacPorts, Fink, Homebrew, etc, and ending up copying
+      # them into the bundle.  Since we optionally depend on libraries which
+      # are not part of OS X (ffmpeg, etc.), however, don't remove the default
+      # path entirely as was done in a previous version of this file.  This is
+      # still kinda evil, since it defeats the user's path settings...
+      # See http://www.cmake.org/cmake/help/v3.0/command/find_program.html
+      list(APPEND CMAKE_PREFIX_PATH "/usr")
+    endif()
+
+    # Linker flags.
+    # Drop unreachable code and data.
+    set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -Wl,-dead_strip,-dead_strip_dylibs")
+
+    find_library(APPKIT_LIBRARY AppKit)
+    find_library(APPSERV_LIBRARY ApplicationServices)
+    find_library(CARBON_LIBRARY Carbon)
+    find_library(COCOA_LIBRARY Cocoa)
+    find_library(COREFOUNDATION_LIBRARY CoreFoundation)
+    find_library(CORESERV_LIBRARY CoreServices)
+    find_library(FOUNDATION_LIBRARY Foundation)
+    find_library(IOK_LIBRARY IOKit)
 elseif (WIN32)
     # WSAPoll and SHGetKnownFolderPath (AppData/Roaming) didn't exist before WinNT 6.x (Vista)
     add_definitions(-D_WIN32_WINNT=0x0600 -DWINVER=0x0600)
diff --git a/CMakeModules/PostprocessBundle.cmake b/CMakeModules/PostprocessBundle.cmake
new file mode 100644
index 0000000000..1065de7dc6
--- /dev/null
+++ b/CMakeModules/PostprocessBundle.cmake
@@ -0,0 +1,46 @@
+# This module can be used in two different ways.
+#
+# When invoked as `cmake -P PostprocessBundle.cmake`, it fixes up an
+# application folder to be standalone. It bundles all required libraries from
+# the system and fixes up library IDs. Any additional shared libraries, like
+# plugins, that are found under Contents/MacOS/ will be made standalone as well.
+#
+# When called with `include(PostprocessBundle)`, it defines a helper
+# function `postprocess_bundle` that sets up the command form of the
+# module as a post-build step.
+
+if(CMAKE_GENERATOR)
+	# Being called as include(PostprocessBundle), so define a helper function.
+	set(_POSTPROCESS_BUNDLE_MODULE_LOCATION "${CMAKE_CURRENT_LIST_FILE}")
+	function(postprocess_bundle target)
+		add_custom_command(TARGET ${target} POST_BUILD
+			COMMAND ${CMAKE_COMMAND} -DBUNDLE_PATH="$<TARGET_FILE_DIR:${target}>/../.."
+				-P "${_POSTPROCESS_BUNDLE_MODULE_LOCATION}"
+		)
+	endfunction()
+	return()
+endif()
+
+get_filename_component(BUNDLE_PATH "${BUNDLE_PATH}" REALPATH)
+message(STATUS "Fixing up application bundle: ${BUNDLE_PATH}")
+
+# Make sure to fix up any additional shared libraries (like plugins) that are
+# needed.
+file(GLOB_RECURSE extra_libs "${BUNDLE_PATH}/Contents/MacOS/*.dylib")
+
+# BundleUtilities doesn't support DYLD_FALLBACK_LIBRARY_PATH behavior, which
+# makes it sometimes break on libraries that do weird things with @rpath. Specify
+# equivalent search directories until https://gitlab.kitware.com/cmake/cmake/issues/16625
+# is fixed and in our minimum CMake version.
+set(extra_dirs "/usr/local/lib" "/lib" "/usr/lib")
+
+# BundleUtilities is overly verbose, so disable most of its messages
+function(message)
+	if(NOT ARGV MATCHES "^STATUS;")
+		_message(${ARGV})
+	endif()
+endfunction()
+
+include(BundleUtilities)
+set(BU_CHMOD_BUNDLE_ITEMS ON)
+fixup_bundle("${BUNDLE_PATH}" "${extra_libs}" "${extra_dirs}")
\ No newline at end of file
diff --git a/externals/MoltenVK/Resources/vulkan/icd.d/MoltenVK_icd.json b/externals/MoltenVK/Resources/vulkan/icd.d/MoltenVK_icd.json
new file mode 100644
index 0000000000..a95edb05b4
--- /dev/null
+++ b/externals/MoltenVK/Resources/vulkan/icd.d/MoltenVK_icd.json
@@ -0,0 +1,7 @@
+{
+    "file_format_version" : "1.0.0",
+    "ICD": {
+        "library_path": "../../../Frameworks/libMoltenVK.dylib",
+        "api_version" : "1.0.0"
+    }
+}
\ No newline at end of file
diff --git a/externals/MoltenVK/include/MoltenVK/mvk_datatypes.h b/externals/MoltenVK/include/MoltenVK/mvk_datatypes.h
new file mode 100644
index 0000000000..55320a4ae2
--- /dev/null
+++ b/externals/MoltenVK/include/MoltenVK/mvk_datatypes.h
@@ -0,0 +1,481 @@
+/*
+ * mvk_datatypes.h
+ *
+ * Copyright (c) 2015-2020 The Brenwill Workshop Ltd. (http://www.brenwill.com)
+ *
+ * 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.
+ */
+
+
+/* 
+ * This file contains functions for converting between Vulkan and Metal data types.
+ *
+ * The functions here are used internally by MoltenVK, and are exposed here 
+ * as a convenience for use elsewhere within applications using MoltenVK.
+ */
+
+#ifndef __mvkDataTypes_h_
+#define __mvkDataTypes_h_ 1
+
+#ifdef __cplusplus
+extern "C" {
+#endif	//  __cplusplus
+	
+#include "mvk_vulkan.h"
+
+#import <Metal/Metal.h>
+#import <CoreGraphics/CoreGraphics.h>
+
+
+#pragma mark -
+#pragma mark Image properties
+
+#pragma mark Texture formats
+
+/** Enumerates the data type of a format. */
+typedef enum {
+    kMVKFormatNone,             /**< Format type is unknown. */
+	kMVKFormatColorHalf,		/**< A 16-bit floating point color. */
+    kMVKFormatColorFloat,		/**< A 32-bit floating point color. */
+	kMVKFormatColorInt8,        /**< A signed 8-bit integer color. */
+	kMVKFormatColorUInt8,		/**< An unsigned 8-bit integer color. */
+	kMVKFormatColorInt16,       /**< A signed 16-bit integer color. */
+	kMVKFormatColorUInt16,		/**< An unsigned 16-bit integer color. */
+    kMVKFormatColorInt32,       /**< A signed 32-bit integer color. */
+    kMVKFormatColorUInt32,		/**< An unsigned 32-bit integer color. */
+    kMVKFormatDepthStencil,     /**< A depth and stencil value. */
+    kMVKFormatCompressed,       /**< A block-compressed color. */
+} MVKFormatType;
+
+/** Returns whether the VkFormat is supported by this implementation. */
+bool mvkVkFormatIsSupported(VkFormat vkFormat);
+
+/** Returns whether the MTLPixelFormat is supported by this implementation. */
+bool mvkMTLPixelFormatIsSupported(MTLPixelFormat mtlFormat);
+
+/** Returns the format type corresponding to the specified Vulkan VkFormat, */
+MVKFormatType mvkFormatTypeFromVkFormat(VkFormat vkFormat);
+
+/** Returns the format type corresponding to the specified Metal MTLPixelFormat, */
+MVKFormatType mvkFormatTypeFromMTLPixelFormat(MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the Metal MTLPixelFormat corresponding to the specified Vulkan VkFormat,
+ * or returns MTLPixelFormatInvalid if no corresponding MTLPixelFormat exists.
+ *
+ * Not all MTLPixelFormats returned by this function are supported by all GPU's, 
+ * and, internally, MoltenVK may substitute and use a different MTLPixelFormat than
+ * is returned by this function for a particular Vulkan VkFormat value.
+ *
+ * Not all macOS GPU's support the MTLPixelFormatDepth24Unorm_Stencil8 pixel format.
+ * Even though this function will return that value when passed the corresponding 
+ * VkFormat value, internally, MoltenVK will use the MTLPixelFormatDepth32Float_Stencil8
+ * instead when a GPU does not support the MTLPixelFormatDepth24Unorm_Stencil8 pixel format.
+ * On an macOS device that has more than one GPU, one of the GPU's may support the
+ * MTLPixelFormatDepth24Unorm_Stencil8 pixel format while another may not.
+ */
+MTLPixelFormat mvkMTLPixelFormatFromVkFormat(VkFormat vkFormat);
+
+/**
+ * Returns the Vulkan VkFormat corresponding to the specified Metal MTLPixelFormat,
+ * or returns VK_FORMAT_UNDEFINED if no corresponding VkFormat exists.
+ */
+VkFormat mvkVkFormatFromMTLPixelFormat(MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the size, in bytes, of a texel block of the specified Vulkan format.
+ * For uncompressed formats, the returned value corresponds to the size in bytes of a single texel.
+ */
+uint32_t mvkVkFormatBytesPerBlock(VkFormat vkFormat);
+
+/** 
+ * Returns the size, in bytes, of a texel block of the specified Metal format.
+ * For uncompressed formats, the returned value corresponds to the size in bytes of a single texel.
+ */
+uint32_t mvkMTLPixelFormatBytesPerBlock(MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the size of the compression block, measured in texels for a Vulkan format.
+ * The returned value will be {1, 1} for non-compressed formats.
+ */
+VkExtent2D mvkVkFormatBlockTexelSize(VkFormat vkFormat);
+
+/**
+ * Returns the size of the compression block, measured in texels for a Metal format.
+ * The returned value will be {1, 1} for non-compressed formats.
+ */
+VkExtent2D mvkMTLPixelFormatBlockTexelSize(MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the size, in bytes, of a texel of the specified Vulkan format.
+ * The returned value may be fractional for certain compressed formats.
+ */
+float mvkVkFormatBytesPerTexel(VkFormat vkFormat);
+
+/**
+ * Returns the size, in bytes, of a texel of the specified Metal format.
+ * The returned value may be fractional for certain compressed formats.
+ */
+float mvkMTLPixelFormatBytesPerTexel(MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the size, in bytes, of a row of texels of the specified Vulkan format.
+ *
+ * For compressed formats, this takes into consideration the compression block size,
+ * and texelsPerRow should specify the width in texels, not blocks. The result is rounded
+ * up if texelsPerRow is not an integer multiple of the compression block width.
+ */
+size_t mvkVkFormatBytesPerRow(VkFormat vkFormat, uint32_t texelsPerRow);
+
+/**
+ * Returns the size, in bytes, of a row of texels of the specified Metal format.
+ *
+ * For compressed formats, this takes into consideration the compression block size,
+ * and texelsPerRow should specify the width in texels, not blocks. The result is rounded
+ * up if texelsPerRow is not an integer multiple of the compression block width.
+ */
+size_t mvkMTLPixelFormatBytesPerRow(MTLPixelFormat mtlFormat, uint32_t texelsPerRow);
+
+/**
+ * Returns the size, in bytes, of a texture layer of the specified Vulkan format.
+ *
+ * For compressed formats, this takes into consideration the compression block size,
+ * and texelRowsPerLayer should specify the height in texels, not blocks. The result is
+ * rounded up if texelRowsPerLayer is not an integer multiple of the compression block height.
+ */
+size_t mvkVkFormatBytesPerLayer(VkFormat vkFormat, size_t bytesPerRow, uint32_t texelRowsPerLayer);
+
+/**
+ * Returns the size, in bytes, of a texture layer of the specified Metal format.
+ * For compressed formats, this takes into consideration the compression block size,
+ * and texelRowsPerLayer should specify the height in texels, not blocks. The result is
+ * rounded up if texelRowsPerLayer is not an integer multiple of the compression block height.
+ */
+size_t mvkMTLPixelFormatBytesPerLayer(MTLPixelFormat mtlFormat, size_t bytesPerRow, uint32_t texelRowsPerLayer);
+
+/** 
+ * Returns the default properties for the specified Vulkan format.
+ *
+ * Not all MTLPixelFormats returned by this function are supported by all GPU's, and, as a
+ * result, MoltenVK may return a different value from the vkGetPhysicalDeviceFormatProperties()
+ * function than is returned here. Use the vkGetPhysicalDeviceFormatProperties() function to
+ * return the properties for a particular GPU.
+ *
+ * Setting assumeGPUSupportsDefault to true allows the default format properties to be returned.
+ * The assumeGPUSupportsDefault flag can be set to false if it is already known that the format
+ * is not supported by a particular GPU for images, in which case all of the returned properties
+ * will be disabled, except possibly VK_FORMAT_FEATURE_VERTEX_BUFFER_BIT, which may be supported
+ * for the format even without image support.
+ */
+VkFormatProperties mvkVkFormatProperties(VkFormat vkFormat, bool assumeGPUSupportsDefault = true);
+
+/** Returns the name of the specified Vulkan format. */
+const char* mvkVkFormatName(VkFormat vkFormat);
+
+/** Returns the name of the specified Metal pixel format. */
+const char* mvkMTLPixelFormatName(MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the MTLClearColor value corresponding to the color value in the VkClearValue,
+ * extracting the color value that is VkFormat for the VkFormat.
+ */
+MTLClearColor mvkMTLClearColorFromVkClearValue(VkClearValue vkClearValue,
+											   VkFormat vkFormat);
+
+/** Returns the Metal depth value corresponding to the depth value in the specified VkClearValue. */
+double mvkMTLClearDepthFromVkClearValue(VkClearValue vkClearValue);
+
+/** Returns the Metal stencil value corresponding to the stencil value in the specified VkClearValue. */
+uint32_t mvkMTLClearStencilFromVkClearValue(VkClearValue vkClearValue);
+
+/** Returns whether the specified Metal MTLPixelFormat can be used as a depth format. */
+bool mvkMTLPixelFormatIsDepthFormat(MTLPixelFormat mtlFormat);
+
+/** Returns whether the specified Metal MTLPixelFormat can be used as a stencil format. */
+bool mvkMTLPixelFormatIsStencilFormat(MTLPixelFormat mtlFormat);
+
+/** Returns whether the specified Metal MTLPixelFormat is a PVRTC format. */
+bool mvkMTLPixelFormatIsPVRTCFormat(MTLPixelFormat mtlFormat);
+
+/** Returns the Metal texture type from the specified Vulkan image properties. */
+MTLTextureType mvkMTLTextureTypeFromVkImageType(VkImageType vkImageType,
+												uint32_t arraySize,
+												bool isMultisample);
+
+/** Returns the Vulkan image type from the Metal texture type. */
+VkImageType mvkVkImageTypeFromMTLTextureType(MTLTextureType mtlTextureType);
+
+/** Returns the Metal MTLTextureType corresponding to the Vulkan VkImageViewType. */
+MTLTextureType mvkMTLTextureTypeFromVkImageViewType(VkImageViewType vkImageViewType, bool isMultisample);
+
+/** Returns the Metal texture usage from the Vulkan image usage. */
+MTLTextureUsage mvkMTLTextureUsageFromVkImageUsageFlags(VkImageUsageFlags vkImageUsageFlags);
+
+/** Returns the Vulkan image usage from the Metal texture usage and format. */
+VkImageUsageFlags mvkVkImageUsageFlagsFromMTLTextureUsage(MTLTextureUsage mtlUsage, MTLPixelFormat mtlFormat);
+
+/**
+ * Returns the numeric sample count corresponding to the specified Vulkan sample count flag.
+ *
+ * The specified flags value should have only one bit set, otherwise an invalid numeric value will be returned.
+ */
+uint32_t mvkSampleCountFromVkSampleCountFlagBits(VkSampleCountFlagBits vkSampleCountFlag);
+
+/** Returns the Vulkan bit flags corresponding to the numeric sample count, which must be a PoT value. */
+VkSampleCountFlagBits mvkVkSampleCountFlagBitsFromSampleCount(NSUInteger sampleCount);
+
+/** Returns the Metal texture swizzle from the Vulkan component swizzle. */
+MTLTextureSwizzle mvkMTLTextureSwizzleFromVkComponentSwizzle(VkComponentSwizzle vkSwizzle);
+
+/** Returns all four Metal texture swizzles from the Vulkan component mapping. */
+MTLTextureSwizzleChannels mvkMTLTextureSwizzleChannelsFromVkComponentMapping(VkComponentMapping vkMapping);
+
+
+#pragma mark Mipmaps
+
+/**
+ * Returns the number of mipmap levels available to an image with the specified side dimension.
+ * 
+ * If the specified dimension is a power-of-two, the value returned is (log2(dim) + 1).
+ * If the specified dimension is NOT a power-of-two, the value returned is 0, indicating
+ * that the image cannot support mipmaps.
+ */
+uint32_t mvkMipmapLevels(uint32_t dim);
+
+/**
+ * Returns the number of mipmap levels available to an image with the specified extent.
+ *
+ * If each dimension in the specified extent is a power-of-two, the value returned
+ * is MAX(log2(dim) + 1) across both dimensions. If either dimension in the specified 
+ * extent is NOT a power-of-two, the value returned is 1, indicating that the image 
+ * cannot support mipmaps, and that only the base mip level can be used.
+ */
+uint32_t mvkMipmapLevels2D(VkExtent2D extent);
+
+/**
+ * Returns the number of mipmap levels available to an image with the specified extent.
+ *
+ * If each dimension in the specified extent is a power-of-two, the value returned
+ * is MAX(log2(dim) + 1) across all dimensions. If either dimension in the specified
+ * extent is NOT a power-of-two, the value returned is 1, indicating that the image
+ * cannot support mipmaps, and that only the base mip level can be used.
+ */
+uint32_t mvkMipmapLevels3D(VkExtent3D extent);
+
+/** 
+ * Returns the size of the specified zero-based mipmap level, 
+ * when the size of the base level is the specified size. 
+ */
+VkExtent2D mvkMipmapLevelSizeFromBaseSize2D(VkExtent2D baseSize, uint32_t level);
+
+/**
+ * Returns the size of the specified zero-based mipmap level,
+ * when the size of the base level is the specified size.
+ */
+VkExtent3D mvkMipmapLevelSizeFromBaseSize3D(VkExtent3D baseSize, uint32_t level);
+
+/** 
+ * Returns the size of the mipmap base level, when the size of 
+ * the specified zero-based mipmap level is the specified size.
+ */
+VkExtent2D mvkMipmapBaseSizeFromLevelSize2D(VkExtent2D levelSize, uint32_t level);
+
+/**
+ * Returns the size of the mipmap base level, when the size of
+ * the specified zero-based mipmap level is the specified size.
+ */
+VkExtent3D mvkMipmapBaseSizeFromLevelSize3D(VkExtent3D levelSize, uint32_t level);
+
+
+#pragma mark Samplers
+
+/**
+ * Returns the Metal MTLSamplerAddressMode corresponding to the specified Vulkan VkSamplerAddressMode,
+ * or returns MTLSamplerAddressModeMirrorClampToEdge if no corresponding MTLSamplerAddressMode exists.
+ */
+MTLSamplerAddressMode mvkMTLSamplerAddressModeFromVkSamplerAddressMode(VkSamplerAddressMode vkMode);
+
+#ifdef __MAC_OS_X_VERSION_MAX_ALLOWED
+/**
+ * Returns the Metal MTLSamplerBorderColor corresponding to the specified Vulkan VkBorderColor,
+ * or returns MTLSamplerBorderColorTransparentBlack if no corresponding MTLSamplerBorderColor exists.
+ */
+MTLSamplerBorderColor mvkMTLSamplerBorderColorFromVkBorderColor(VkBorderColor vkColor);
+#endif
+
+/**
+ * Returns the Metal MTLSamplerMinMagFilter corresponding to the specified Vulkan VkFilter,
+ * or returns MTLSamplerMinMagFilterNearest if no corresponding MTLSamplerMinMagFilter exists.
+ */
+MTLSamplerMinMagFilter mvkMTLSamplerMinMagFilterFromVkFilter(VkFilter vkFilter);
+
+/**
+ * Returns the Metal MTLSamplerMipFilter corresponding to the specified Vulkan VkSamplerMipmapMode,
+ * or returns MTLSamplerMipFilterNotMipmapped if no corresponding MTLSamplerMipFilter exists.
+ */
+MTLSamplerMipFilter mvkMTLSamplerMipFilterFromVkSamplerMipmapMode(VkSamplerMipmapMode vkMode);
+
+
+#pragma mark -
+#pragma mark Render pipeline
+
+/** Identifies a particular shading stage in a pipeline. */
+typedef enum {
+	kMVKShaderStageVertex = 0,
+	kMVKShaderStageTessCtl,
+	kMVKShaderStageTessEval,
+	kMVKShaderStageFragment,
+	kMVKShaderStageCompute,
+	kMVKShaderStageMax
+} MVKShaderStage;
+
+/** Returns the Metal MTLColorWriteMask corresponding to the specified Vulkan VkColorComponentFlags. */
+MTLColorWriteMask mvkMTLColorWriteMaskFromVkChannelFlags(VkColorComponentFlags vkWriteFlags);
+
+/** Returns the Metal MTLBlendOperation corresponding to the specified Vulkan VkBlendOp. */
+MTLBlendOperation mvkMTLBlendOperationFromVkBlendOp(VkBlendOp vkBlendOp);
+
+/** Returns the Metal MTLBlendFactor corresponding to the specified Vulkan VkBlendFactor. */
+MTLBlendFactor mvkMTLBlendFactorFromVkBlendFactor(VkBlendFactor vkBlendFactor);
+
+/**
+ * Returns the Metal MTLVertexFormat corresponding to the specified
+ * Vulkan VkFormat as used as a vertex attribute format.
+ */
+MTLVertexFormat mvkMTLVertexFormatFromVkFormat(VkFormat vkFormat);
+
+/** Returns the Metal MTLVertexStepFunction corresponding to the specified Vulkan VkVertexInputRate. */
+MTLVertexStepFunction mvkMTLVertexStepFunctionFromVkVertexInputRate(VkVertexInputRate vkVtxStep);
+
+/** Returns the Metal MTLPrimitiveType corresponding to the specified Vulkan VkPrimitiveTopology. */
+MTLPrimitiveType mvkMTLPrimitiveTypeFromVkPrimitiveTopology(VkPrimitiveTopology vkTopology);
+
+/** Returns the Metal MTLPrimitiveTopologyClass corresponding to the specified Vulkan VkPrimitiveTopology. */
+MTLPrimitiveTopologyClass mvkMTLPrimitiveTopologyClassFromVkPrimitiveTopology(VkPrimitiveTopology vkTopology);
+
+/** Returns the Metal MTLTriangleFillMode corresponding to the specified Vulkan VkPolygonMode, */
+MTLTriangleFillMode mvkMTLTriangleFillModeFromVkPolygonMode(VkPolygonMode vkFillMode);
+
+/** Returns the Metal MTLLoadAction corresponding to the specified Vulkan VkAttachmentLoadOp. */
+MTLLoadAction mvkMTLLoadActionFromVkAttachmentLoadOp(VkAttachmentLoadOp vkLoadOp);
+
+/** Returns the Metal MTLStoreAction corresponding to the specified Vulkan VkAttachmentStoreOp. */
+MTLStoreAction mvkMTLStoreActionFromVkAttachmentStoreOp(VkAttachmentStoreOp vkStoreOp, bool hasResolveAttachment = false);
+
+/** Returns the Metal MTLViewport corresponding to the specified Vulkan VkViewport. */
+MTLViewport mvkMTLViewportFromVkViewport(VkViewport vkViewport);
+
+/** Returns the Metal MTLScissorRect corresponding to the specified Vulkan VkRect2D. */
+MTLScissorRect mvkMTLScissorRectFromVkRect2D(VkRect2D vkRect);
+
+/** Returns the Metal MTLCompareFunction corresponding to the specified Vulkan VkCompareOp, */
+MTLCompareFunction mvkMTLCompareFunctionFromVkCompareOp(VkCompareOp vkOp);
+
+/** Returns the Metal MTLStencilOperation corresponding to the specified Vulkan VkStencilOp, */
+MTLStencilOperation mvkMTLStencilOperationFromVkStencilOp(VkStencilOp vkOp);
+
+/** Returns the Metal MTLCullMode corresponding to the specified Vulkan VkCullModeFlags, */
+MTLCullMode mvkMTLCullModeFromVkCullModeFlags(VkCullModeFlags vkCull);
+
+/** Returns the Metal MTLWinding corresponding to the specified Vulkan VkFrontFace, */
+MTLWinding mvkMTLWindingFromVkFrontFace(VkFrontFace vkWinding);
+
+/** Returns the Metal MTLIndexType corresponding to the specified Vulkan VkIndexType, */
+MTLIndexType mvkMTLIndexTypeFromVkIndexType(VkIndexType vkIdxType);
+
+/** Returns the size, in bytes, of a vertex index of the specified type. */
+size_t mvkMTLIndexTypeSizeInBytes(MTLIndexType mtlIdxType);
+
+/** Returns the MoltenVK MVKShaderStage corresponding to the specified Vulkan VkShaderStageFlagBits. */
+MVKShaderStage mvkShaderStageFromVkShaderStageFlagBits(VkShaderStageFlagBits vkStage);
+
+/** Returns the Vulkan VkShaderStageFlagBits corresponding to the specified MoltenVK MVKShaderStage. */
+VkShaderStageFlagBits mvkVkShaderStageFlagBitsFromMVKShaderStage(MVKShaderStage mvkStage);
+
+/** Returns the Metal MTLWinding corresponding to the specified SPIR-V spv::ExecutionMode. */
+MTLWinding mvkMTLWindingFromSpvExecutionMode(uint32_t spvMode);
+
+/** Returns the Metal MTLTessellationPartitionMode corresponding to the specified SPIR-V spv::ExecutionMode. */
+MTLTessellationPartitionMode mvkMTLTessellationPartitionModeFromSpvExecutionMode(uint32_t spvMode);
+
+/**
+ * Returns the combination of Metal MTLRenderStage bits corresponding to the specified Vulkan VkPiplineStageFlags,
+ * taking into consideration whether the barrier is to be placed before or after the specified pipeline stages.
+ */
+MTLRenderStages mvkMTLRenderStagesFromVkPipelineStageFlags(VkPipelineStageFlags vkStages, bool placeBarrierBefore);
+
+/** Returns the combination of Metal MTLBarrierScope bits corresponding to the specified Vulkan VkAccessFlags. */
+MTLBarrierScope mvkMTLBarrierScopeFromVkAccessFlags(VkAccessFlags vkAccess);
+
+#pragma mark -
+#pragma mark Geometry conversions
+
+/** Returns a VkExtent2D that corresponds to the specified CGSize. */
+static inline VkExtent2D mvkVkExtent2DFromCGSize(CGSize cgSize) {
+	VkExtent2D vkExt;
+	vkExt.width = cgSize.width;
+	vkExt.height = cgSize.height;
+	return vkExt;
+}
+
+/** Returns a Metal MTLOrigin constructed from a VkOffset3D. */
+static inline MTLOrigin mvkMTLOriginFromVkOffset3D(VkOffset3D vkOffset) {
+	return MTLOriginMake(vkOffset.x, vkOffset.y, vkOffset.z);
+}
+
+/** Returns a Vulkan VkOffset3D constructed from a Metal MTLOrigin. */
+static inline VkOffset3D mvkVkOffset3DFromMTLSize(MTLOrigin mtlOrigin) {
+	return { (int32_t)mtlOrigin.x, (int32_t)mtlOrigin.y, (int32_t)mtlOrigin.z };
+}
+
+/** Returns a Metal MTLSize constructed from a VkExtent3D. */
+static inline MTLSize mvkMTLSizeFromVkExtent3D(VkExtent3D vkExtent) {
+	return MTLSizeMake(vkExtent.width, vkExtent.height, vkExtent.depth);
+}
+
+/** Returns a Vulkan VkExtent3D  constructed from a Metal MTLSize. */
+static inline VkExtent3D mvkVkExtent3DFromMTLSize(MTLSize mtlSize) {
+	return { (uint32_t)mtlSize.width, (uint32_t)mtlSize.height, (uint32_t)mtlSize.depth };
+}
+
+
+#pragma mark -
+#pragma mark Memory options
+
+/** Macro indicating the Vulkan memory type bits corresponding to Metal private memory (not host visible). */
+#define MVK_VK_MEMORY_TYPE_METAL_PRIVATE	(VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT)
+
+/** Macro indicating the Vulkan memory type bits corresponding to Metal shared memory (host visible and coherent). */
+#define MVK_VK_MEMORY_TYPE_METAL_SHARED		(VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_COHERENT_BIT)
+
+/** Macro indicating the Vulkan memory type bits corresponding to Metal managed memory (host visible and non-coherent). */
+#define MVK_VK_MEMORY_TYPE_METAL_MANAGED	(VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT | VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | VK_MEMORY_PROPERTY_HOST_CACHED_BIT)
+
+/** Macro indicating the Vulkan memory type bits corresponding to Metal memoryless memory (not host visible and lazily allocated). */
+#define MVK_VK_MEMORY_TYPE_METAL_MEMORYLESS	(VK_MEMORY_PROPERTY_DEVICE_LOCAL_BIT | VK_MEMORY_PROPERTY_LAZILY_ALLOCATED_BIT)
+
+/** Returns the Metal storage mode corresponding to the specified Vulkan memory flags. */
+MTLStorageMode mvkMTLStorageModeFromVkMemoryPropertyFlags(VkMemoryPropertyFlags vkFlags);
+
+/** Returns the Metal CPU cache mode corresponding to the specified Vulkan memory flags. */
+MTLCPUCacheMode mvkMTLCPUCacheModeFromVkMemoryPropertyFlags(VkMemoryPropertyFlags vkFlags);
+
+/** Returns the Metal resource option flags corresponding to the Metal storage mode and cache mode. */
+MTLResourceOptions mvkMTLResourceOptions(MTLStorageMode mtlStorageMode, MTLCPUCacheMode mtlCPUCacheMode);
+
+
+#ifdef __cplusplus
+}
+#endif	//  __cplusplus
+
+#endif
diff --git a/externals/MoltenVK/include/MoltenVK/mvk_vulkan.h b/externals/MoltenVK/include/MoltenVK/mvk_vulkan.h
new file mode 100644
index 0000000000..757bb7ed0a
--- /dev/null
+++ b/externals/MoltenVK/include/MoltenVK/mvk_vulkan.h
@@ -0,0 +1,49 @@
+/*
+ * mvk_vulkan.h
+ *
+ * Copyright (c) 2015-2020 The Brenwill Workshop Ltd. (http://www.brenwill.com)
+ *
+ * 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.
+ */
+
+
+/** 
+ * This is a convenience header file that loads vulkan.h with the appropriate Vulkan platform extensions.
+ *
+ * This header automatically enables the VK_EXT_metal_surface Vulkan extension.
+ *
+ * When building for iOS, this header also automatically enables the obsolete VK_MVK_ios_surface Vulkan extension.
+ * When building for macOS, this header also automatically enables the obsolete VK_MVK_macos_surface Vulkan extension.
+ * Both of these extensions are obsolete. Consider using the portable VK_EXT_metal_surface extension instead.
+ */
+
+#ifndef __mvk_vulkan_h_
+#define __mvk_vulkan_h_ 1
+
+
+#include <Availability.h>
+
+#define VK_USE_PLATFORM_METAL_EXT				1
+
+#ifdef __IPHONE_OS_VERSION_MAX_ALLOWED
+#	define VK_USE_PLATFORM_IOS_MVK				1
+#endif
+
+#ifdef __MAC_OS_X_VERSION_MAX_ALLOWED
+#	define VK_USE_PLATFORM_MACOS_MVK			1
+#endif
+
+#include <vulkan/vulkan.h>
+#include <vulkan-portability/vk_extx_portability_subset.h>
+
+#endif
diff --git a/externals/MoltenVK/include/MoltenVK/vk_mvk_moltenvk.h b/externals/MoltenVK/include/MoltenVK/vk_mvk_moltenvk.h
new file mode 100644
index 0000000000..078c134a4a
--- /dev/null
+++ b/externals/MoltenVK/include/MoltenVK/vk_mvk_moltenvk.h
@@ -0,0 +1,1050 @@
+/*
+ * vk_mvk_moltenvk.h
+ *
+ * Copyright (c) 2015-2020 The Brenwill Workshop Ltd. (http://www.brenwill.com)
+ *
+ * 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.
+ */
+
+
+/** Vulkan extension VK_MVK_moltenvk. */
+
+#ifndef __vk_mvk_moltenvk_h_
+#define __vk_mvk_moltenvk_h_ 1
+
+#ifdef __cplusplus
+extern "C" {
+#endif	//  __cplusplus
+	
+#include "mvk_vulkan.h"
+
+#ifdef __OBJC__
+#import <Metal/Metal.h>
+#import <IOSurface/IOSurfaceRef.h>
+#else
+typedef unsigned long MTLLanguageVersion;
+#endif
+
+
+/**
+ * The version number of MoltenVK is a single integer value, derived from the Major, Minor,
+ * and Patch version values, where each of the Major, Minor, and Patch components is allocated
+ * two decimal digits, in the format MjMnPt. This creates a version number that is both human
+ * readable and allows efficient computational comparisons to a single integer number.
+ *
+ * The following examples illustrate how the MoltenVK version number is built from its components:
+ *   - 002000    (version 0.20.0)
+ *   - 010000    (version 1.0.0)
+ *   - 030104    (version 3.1.4)
+ *   - 401215    (version 4.12.15)
+ */
+#define MVK_VERSION_MAJOR   1
+#define MVK_VERSION_MINOR   0
+#define MVK_VERSION_PATCH   40
+
+#define MVK_MAKE_VERSION(major, minor, patch)    (((major) * 10000) + ((minor) * 100) + (patch))
+#define MVK_VERSION     MVK_MAKE_VERSION(MVK_VERSION_MAJOR, MVK_VERSION_MINOR, MVK_VERSION_PATCH)
+
+#define VK_MVK_MOLTENVK_SPEC_VERSION            24
+#define VK_MVK_MOLTENVK_EXTENSION_NAME          "VK_MVK_moltenvk"
+
+/**
+ * MoltenVK configuration settings.
+ *
+ * To be active, some configuration settings must be set before a VkDevice is created.
+ * See the description of the individual configuration structure members for more information.
+ *
+ * There are three mechanisms for setting the values of the MoltenVK configuration parameters:
+ *   - Runtime API via the vkGetMoltenVKConfigurationMVK()/vkSetMoltenVKConfigurationMVK() functions.
+ *   - Application runtime environment variables.
+ *   - Build settings at MoltenVK build time.
+ *
+ * To change the MoltenVK configuration settings at runtime using a programmatic API,
+ * use the vkGetMoltenVKConfigurationMVK() and vkSetMoltenVKConfigurationMVK() functions
+ * to retrieve, modify, and set a copy of the MVKConfiguration structure.
+ *
+ * The initial value of each of the configuration settings can established at runtime
+ * by a corresponding environment variable, or if the environment variable is not set,
+ * by a corresponding build setting at the time MoltenVK is compiled. The environment
+ * variable and build setting for each configuration parameter share the same name.
+ *
+ * For example, the initial value of the shaderConversionFlipVertexY configuration setting
+ * is set by the MVK_CONFIG_SHADER_CONVERSION_FLIP_VERTEX_Y at runtime, or by the
+ * MVK_CONFIG_SHADER_CONVERSION_FLIP_VERTEX_Y build setting when MoltenVK is compiled.
+ *
+ * This structure may be extended as new features are added to MoltenVK. If you are linking to
+ * an implementation of MoltenVK that was compiled from a different VK_MVK_MOLTENVK_SPEC_VERSION
+ * than your app was, the size of this structure in your app may be larger or smaller than the
+ * struct in MoltenVK. See the description of the vkGetMoltenVKConfigurationMVK() and
+ * vkSetMoltenVKConfigurationMVK() functions for information about how to handle this.
+ *
+ * TO SUPPORT DYNAMIC LINKING TO THIS STRUCTURE AS DESCRIBED ABOVE, THIS STRUCTURE SHOULD NOT
+ * BE CHANGED EXCEPT TO ADD ADDITIONAL MEMBERS ON THE END. EXISTING MEMBERS, AND THEIR ORDER,
+ * SHOULD NOT BE CHANGED.
+ *
+ * In addition to the configuration parmeters in this structure, there are several settings that
+ * can be configured through runtime environment variables or MoltenVK compile-time build settings:
+ *
+ * 1. The MVK_CONFIG_LOG_LEVEL runtime environment variable or MoltenVK compile-time build setting
+ *    controls the level of logging performned by MoltenVK using the following numeric values:
+ *      0: No logging.
+ *      1: Log errors only.
+ *      2: Log errors and informational messages.
+ *    If none of these is set, errors and informational messages are logged.
+ *
+ * 2. The MVK_CONFIG_TRACE_VULKAN_CALLS runtime environment variable or MoltenVK compile-time build
+ *    setting causes MoltenVK to log the name of each Vulkan call made by the application. The logging
+ *    format options can be controlled by setting the value of MVK_CONFIG_TRACE_VULKAN_CALLS as follows:
+ *        0: No Vulkan call logging.
+ *        1: Log the name of each Vulkan call when the call is entered.
+ *        2: Log the name of each Vulkan call when the call is entered and exited. This effectively
+ *           brackets any other logging activity within the scope of the Vulkan call.
+ *        3: Same as option 2, plus logs the time spent inside the Vulkan function.
+ *    If none of these is set, no Vulkan call logging will occur.
+ *
+ * 3. Setting the MVK_CONFIG_FORCE_LOW_POWER_GPU runtime environment variable or MoltenVK compile-time
+ *    build setting to 1 will force MoltenVK to use a low-power GPU, if one is availble on the device.
+ *    By default, this setting is disabled, allowing both low-power and high-power GPU's to be used.
+ *
+ * 4. Setting the MVK_ALLOW_METAL_FENCES or MVK_ALLOW_METAL_EVENTS runtime environment variable
+ *    or MoltenVK compile-time build setting to 1 will cause MoltenVK to use MTLFence or MTLEvent,
+ *    respectively, if it is available on the device, for VkSemaphore synchronization behaviour.
+ *    If both variables are set, MVK_ALLOW_METAL_FENCES takes priority over MVK_ALLOW_METAL_EVENTS.
+ *    If both are disabled, or if MTLFence or MTLEvent is not available on the device, MoltenVK
+ *    will use CPU synchronization to control VkSemaphore synchronization behaviour.
+ *    By default, MVK_ALLOW_METAL_FENCES is enabled and MVK_ALLOW_METAL_EVENTS is disabled,
+ *    meaning MoltenVK will use MTLFences, if they are available, to control VkSemaphore
+ *    synchronization behaviour, by default.
+ *
+ * 5. The MVK_CONFIG_AUTO_GPU_CAPTURE_SCOPE runtime environment variable or MoltenVK compile-time
+ *    build setting controls whether Xcode should run an automatic GPU capture without the user
+ *    having to trigger it manually via the Xcode user interface, and controls the scope under
+ *    which that GPU capture will occur. This is useful when trying to capture a one-shot GPU
+ *    trace, such as when running a Vulkan CTS test case. For the automatic GPU capture to occur,
+ *    the Xcode scheme under which the app is run must have the Metal GPU capture option turned on.
+ *    MVK_CONFIG_AUTO_GPU_CAPTURE_SCOPE should not be set to manually trigger a GPU capture via the
+ *    Xcode user interface.
+ *      0: No automatic GPU capture.
+ *      1: Capture all GPU commands issued during the lifetime of the VkDevice.
+ *    If none of these is set, no automatic GPU capture will occur.
+ *
+ * 6. The MVK_CONFIG_TEXTURE_1D_AS_2D runtime environment variable or MoltenVK compile-time build
+ *    setting controls whether MoltenVK should use a Metal 2D texture with a height of 1 for a
+ *    Vulkan 1D image, or use a native Metal 1D texture. Metal imposes significant restrictions
+ *    on native 1D textures, including not being renderable, clearable, or permitting mipmaps.
+ *    Using a Metal 2D texture allows Vulkan 1D textures to support this additional functionality.
+ *    This setting is enabled by default, and MoltenVK will use a Metal 2D texture for each Vulkan 1D image.
+ *
+ * 7. The MVK_CONFIG_PREALLOCATE_DESCRIPTORS runtime environment variable or MoltenVK compile-time
+ *    build setting controls whether MoltenVK should preallocate memory in each VkDescriptorPool
+ *    according to the values of the VkDescriptorPoolSize parameters. Doing so may improve
+ *    descriptor set allocation performance at a cost of preallocated application memory.
+ *    If this environment variable is disabled, the descriptors required for a descriptor set will
+ *    be dynamically allocated in application memory when the descriptor set itself is allocated.
+ *    This setting is disabled by default, and MoltenVK will dynamically allocate descriptors
+ *    when the containing descriptor set is allocated.
+ */
+typedef struct {
+
+	/**
+	 * If enabled, debugging capabilities will be enabled, including logging
+	 * shader code during runtime shader conversion.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_DEBUG
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter is false if MoltenVK was
+	 * built in Release mode, and true if MoltenVK was built in Debug mode.
+	 */
+    VkBool32 debugMode;
+
+	/**
+	 * If enabled, MSL vertex shader code created during runtime shader conversion will
+	 * flip the Y-axis of each vertex, as the Vulkan Y-axis is the inverse of OpenGL.
+	 *
+	 * An alternate way to reverse the Y-axis is to employ a negative Y-axis value on
+	 * the viewport, in which case this parameter can be disabled.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 * Specifically, this parameter can be enabled when compiling some pipelines,
+	 * and disabled when compiling others. Existing pipelines are not automatically
+	 * re-compiled when this parameter is changed.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_SHADER_CONVERSION_FLIP_VERTEX_Y
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to true.
+	 */
+    VkBool32 shaderConversionFlipVertexY;
+
+	/**
+	 * If enabled, queue command submissions (vkQueueSubmit() & vkQueuePresentKHR()) will be
+	 * processed on the thread that called the submission function. If disabled, processing
+	 * will be dispatched to a GCD dispatch_queue whose priority is determined by
+	 * VkDeviceQueueCreateInfo::pQueuePriorities during vkCreateDevice().
+	 *
+	 * The value of this parameter must be changed before creating a VkDevice,
+	 * for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_SYNCHRONOUS_QUEUE_SUBMITS
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to true for macOS 10.14
+	 * and above or iOS 12 and above, and false otherwise. The reason for this distinction
+	 * is that this feature should be disabled when emulation is required to support VkEvents
+	 * because native support for events (MTLEvent) is not available.
+	 */
+	VkBool32 synchronousQueueSubmits;
+
+	/**
+	 * If enabled, where possible, a Metal command buffer will be created and filled when each
+	 * Vulkan command buffer is filled. For applications that parallelize the filling of Vulkan
+	 * commmand buffers across multiple threads, this allows the Metal command buffers to also
+	 * be filled on the same parallel thread. Because each command buffer is filled separately,
+	 * this requires that each Vulkan command buffer requires a dedicated Metal command buffer.
+	 *
+	 * If disabled, a single Metal command buffer will be created and filled when the Vulkan
+	 * command buffers are submitted to the Vulkan queue. This allows a single Metal command
+	 * buffer to be used for all of the Vulkan command buffers in a queue submission. The
+	 * Metal command buffer is filled on the thread that processes the command queue submission.
+	 *
+	 * Depending on the nature of your application, you may find performance is improved by filling
+	 * the Metal command buffers on parallel threads, or you may find that performance is improved by
+	 * consolidating all Vulkan command buffers onto a single Metal command buffer during queue submission.
+	 *
+	 * Prefilling of a Metal command buffer will not occur during the filling of secondary command
+	 * buffers (VK_COMMAND_BUFFER_LEVEL_SECONDARY), or for primary command buffers that are intended
+	 * to be submitted to multiple queues concurrently (VK_COMMAND_BUFFER_USAGE_SIMULTANEOUS_USE_BIT).
+	 *
+	 * When enabling this feature, be aware that one Metal command buffer is required for each Vulkan
+	 * command buffer. Depending on the number of command buffers that you use, you may also need to
+	 * change the value of the maxActiveMetalCommandBuffersPerQueue setting.
+	 *
+	 * In addition, if this feature is enabled, be aware that if you have recorded commands to a
+	 * Vulkan command buffer, and then choose to reset that command buffer instead of submitting it,
+	 * the corresponding prefilled Metal command buffer will still be submitted. This is because Metal
+	 * command buffers do not support the concept of being reset after being filled. Depending on when
+	 * and how often you do this, it may cause unexpected visual artifacts and unnecessary GPU load.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 * Specifically, this parameter can be enabled when filling some command buffers,
+	 * and disabled when filling others.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_PREFILL_METAL_COMMAND_BUFFERS
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to false.
+	 */
+	VkBool32 prefillMetalCommandBuffers;
+
+	/**
+	 * The maximum number of Metal command buffers that can be concurrently active per Vulkan queue.
+	 * The number of active Metal command buffers required depends on the prefillMetalCommandBuffers
+	 * setting. If prefillMetalCommandBuffers is enabled, one Metal command buffer is required per
+	 * Vulkan command buffer. If prefillMetalCommandBuffers is disabled, one Metal command buffer
+	 * is required per command buffer queue submission, which may be significantly less than the
+	 * number of Vulkan command buffers.
+	 *
+	 * The value of this parameter must be changed before creating a VkDevice,
+	 * for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_MAX_ACTIVE_METAL_COMMAND_BUFFERS_PER_QUEUE
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to 64.
+	 */
+	uint32_t maxActiveMetalCommandBuffersPerQueue;
+
+	/**
+	 * Metal allows only 8192 occlusion queries per MTLBuffer. If enabled, MoltenVK
+	 * allocates a MTLBuffer for each query pool, allowing each query pool to support
+	 * 8192 queries, which may slow performance or cause unexpected behaviour if the query
+	 * pool is not established prior to a Metal renderpass, or if the query pool is changed
+	 * within a renderpass. If disabled, one MTLBuffer will be shared by all query pools,
+	 * which improves performance, but limits the total device queries to 8192.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 * Specifically, this parameter can be enabled when creating some query pools,
+	 * and disabled when creating others.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_SUPPORT_LARGE_QUERY_POOLS
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to true.
+	 */
+	VkBool32 supportLargeQueryPools;
+
+	/** Obsolete, ignored, and deprecated. All surface presentations are performed with a command buffer. */
+	VkBool32 presentWithCommandBuffer;
+
+	/**
+	 * If enabled, swapchain images will use simple Nearest sampling when magnifying the
+	 * swapchain image to fit a physical display surface. If disabled, swapchain images will
+	 * use Linear sampling when magnifying the swapchain image to fit a physical display surface.
+	 * Enabling this setting avoids smearing effects when swapchain images are simple interger
+	 * multiples of display pixels (eg- macOS Retina, and typical of graphics apps and games),
+	 * but may cause aliasing effects when using non-integer display scaling.
+	 *
+	 * The value of this parameter may be changed before creating a VkSwapchain,
+	 * for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_SWAPCHAIN_MAG_FILTER_USE_NEAREST
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to true.
+	 */
+	VkBool32 swapchainMagFilterUseNearest;
+
+	/**
+	 * The maximum amount of time, in nanoseconds, to wait for a Metal library, function, or
+	 * pipeline state object to be compiled and created by the Metal compiler. An internal error
+	 * within the Metal compiler can stall the thread for up to 30 seconds. Setting this value
+	 * limits that delay to a specified amount of time, allowing shader compilations to fail fast.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_METAL_COMPILE_TIMEOUT
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to infinite.
+	 */
+	uint64_t metalCompileTimeout;
+
+	/**
+	 * If enabled, performance statistics, as defined by the MVKPerformanceStatistics structure,
+	 * are collected, and can be retrieved via the vkGetPerformanceStatisticsMVK() function.
+	 *
+	 * You can also use the performanceLoggingFrameCount parameter to automatically log the
+	 * performance statistics collected by this parameter.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_PERFORMANCE_TRACKING
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to false.
+	 */
+	VkBool32 performanceTracking;
+
+	/**
+	 * If non-zero, performance statistics, as defined by the MVKPerformanceStatistics structure,
+	 * will be logged to the console. Frame-based statistics will be logged, on a repeating cycle,
+	 * once per this many frames. Non-frame-based statistics will be logged as they occur.
+	 *
+	 * The performanceTracking parameter must also be enabled. If this parameter is zero, or
+	 * the performanceTracking parameter is disabled, no performance statistics will be logged.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_PERFORMANCE_LOGGING_FRAME_COUNT
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to zero.
+	 */
+	uint32_t performanceLoggingFrameCount;
+
+	/**
+	 * If enabled, a MoltenVK logo watermark will be rendered on top of the scene.
+	 * This can be enabled for publicity during demos.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_DISPLAY_WATERMARK
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to false.
+	 */
+	VkBool32 displayWatermark;
+
+	/**
+	 * Metal does not distinguish functionality between queues, which would normally mean only
+	 * a single general-purpose queue family with multiple queues is needed. However, Vulkan
+	 * associates command buffers with a queue family, whereas Metal associates command buffers
+	 * with a specific Metal queue. In order to allow a Metal command buffer to be prefilled
+	 * before is is formally submitted to a Vulkan queue, each Vulkan queue family can support
+	 * only a single Metal queue. As a result, in order to provide parallel queue operations,
+	 * MoltenVK provides multiple queue families, each with a single queue.
+	 *
+	 * If this parameter is disabled, all queue families will be advertised as having general-purpose
+	 * graphics + compute + transfer functionality, which is how the actual Metal queues behave.
+	 *
+	 * If this parameter is enabled, one queue family will be advertised as having general-purpose
+	 * graphics + compute + transfer functionality, and the remaining queue families will be advertised
+	 * as having specialized graphics OR compute OR transfer functionality, to make it easier for some
+	 * apps to select a queue family with the appropriate requirements.
+	 *
+	 * The value of this parameter must be changed before creating a VkDevice, and before
+	 * querying a VkPhysicalDevice for queue family properties, for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_SPECIALIZED_QUEUE_FAMILIES
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to false.
+	 */
+	VkBool32 specializedQueueFamilies;
+
+	/**
+	 * If enabled, when the app creates a VkDevice from a VkPhysicalDevice (GPU) that is neither
+	 * headless nor low-power, and is different than the GPU used by the windowing system, the
+	 * windowing system will be forced to switch to use the GPU selected by the Vulkan app.
+	 * When the Vulkan app is ended, the windowing system will automatically switch back to
+	 * using the previous GPU, depending on the usage requirements of other running apps.
+	 *
+	 * If disabled, the Vulkan app will render using its selected GPU, and if the windowing
+	 * system uses a different GPU, the windowing system compositor will automatically copy
+	 * framebuffer content from the app GPU to the windowing system GPU.
+	 *
+	 * The value of this parmeter has no effect on systems with a single GPU, or when the
+	 * Vulkan app creates a VkDevice from a low-power or headless VkPhysicalDevice (GPU).
+	 *
+	 * Switching the windowing system GPU to match the Vulkan app GPU maximizes app performance,
+	 * because it avoids the windowing system compositor from having to copy framebuffer content
+	 * between GPUs on each rendered frame. However, doing so forces the entire system to
+	 * potentially switch to using a GPU that may consume more power while the app is running.
+	 *
+	 * Some Vulkan apps may want to render using a high-power GPU, but leave it up to the
+	 * system window compositor to determine how best to blend content with the windowing
+	 * system, and as a result, may want to disable this parameter.
+	 *
+	 * The value of this parameter must be changed before creating a VkDevice,
+	 * for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_SWITCH_SYSTEM_GPU
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to true.
+	 */
+	VkBool32 switchSystemGPU;
+
+	/**
+	 * If enabled, arbitrary VkImageView component swizzles are supported, as defined
+	 * in VkImageViewCreateInfo::components when creating a VkImageView.
+	 *
+	 * If disabled, a very limited set of VkImageView component swizzles are supported
+	 * via format substitutions.
+	 *
+	 * Metal does not natively support per-texture swizzling. If this parameter is enabled
+	 * both when a VkImageView is created, and when any pipeline that uses that VkImageView
+	 * is compiled, VkImageView swizzling is automatically performed in the converted Metal
+	 * shader code during all texture sampling and reading operations, regardless of whether
+	 * a swizzle is required for the VkImageView associated with the Metal texture.
+	 * This may result in reduced performance.
+	 *
+	 * The value of this parameter may be changed at any time during application runtime,
+	 * and the changed value will immediately effect subsequent MoltenVK behaviour.
+	 * Specifically, this parameter can be enabled when creating VkImageViews that need it,
+	 * and compiling pipelines that use those VkImageViews, and can be disabled when creating
+	 * VkImageViews that don't need it, and compiling pipelines that use those VkImageViews.
+	 *
+	 * Existing pipelines are not automatically re-compiled when this parameter is changed.
+	 *
+	 * An error is logged and returned during VkImageView creation if that VkImageView
+	 * requires full image view swizzling and this feature is not enabled. An error is
+	 * also logged when a pipeline that was not compiled with full image view swizzling
+	 * is presented with a VkImageView that is expecting it.
+	 *
+	 * An error is also retuned and logged when a VkPhysicalDeviceImageFormatInfo2KHR is passed
+	 * in a call to vkGetPhysicalDeviceImageFormatProperties2KHR() to query for an VkImageView
+	 * format that will require full swizzling to be enabled, and this feature is not enabled.
+	 *
+	 * If this parameter is disabled, the following limited set of VkImageView swizzles are
+	 * supported by MoltenVK, via automatic format substitution:
+	 *
+	 * Texture format			       Swizzle
+	 * --------------                  -------
+	 * VK_FORMAT_R8_UNORM              ZERO, ANY, ANY, RED
+	 * VK_FORMAT_A8_UNORM              ALPHA, ANY, ANY, ZERO
+	 * VK_FORMAT_R8G8B8A8_UNORM        BLUE, GREEN, RED, ALPHA
+	 * VK_FORMAT_R8G8B8A8_SRGB         BLUE, GREEN, RED, ALPHA
+	 * VK_FORMAT_B8G8R8A8_UNORM        BLUE, GREEN, RED, ALPHA
+	 * VK_FORMAT_B8G8R8A8_SRGB         BLUE, GREEN, RED, ALPHA
+	 * VK_FORMAT_D32_SFLOAT_S8_UINT    RED, ANY, ANY, ANY (stencil only)
+	 * VK_FORMAT_D24_UNORM_S8_UINT     RED, ANY, ANY, ANY (stencil only)
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_FULL_IMAGE_VIEW_SWIZZLE
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to false.
+	 */
+	VkBool32 fullImageViewSwizzle;
+
+	/**
+	 * The index of the queue family whose presentation submissions will
+	 * be used as the default GPU Capture Scope during debugging in Xcode.
+	 *
+	 * The value of this parameter must be changed before creating a VkDevice,
+	 * for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_DEFAULT_GPU_CAPTURE_SCOPE_QUEUE_FAMILY_INDEX
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to zero (the first queue family).
+	 */
+	uint32_t defaultGPUCaptureScopeQueueFamilyIndex;
+
+	/**
+	 * The index of the queue, within the queue family identified by the
+	 * defaultGPUCaptureScopeQueueFamilyIndex parameter, whose presentation submissions
+	 * will be used as the default GPU Capture Scope during debugging in Xcode.
+	 *
+	 * The value of this parameter must be changed before creating a VkDevice,
+	 * for the change to take effect.
+	 *
+	 * The initial value or this parameter is set by the
+	 * MVK_CONFIG_DEFAULT_GPU_CAPTURE_SCOPE_QUEUE_INDEX
+	 * runtime environment variable or MoltenVK compile-time build setting.
+	 * If neither is set, the value of this parameter defaults to zero (the first queue).
+	 */
+	uint32_t defaultGPUCaptureScopeQueueIndex;
+
+} MVKConfiguration;
+
+/**
+ * Features provided by the current implementation of Metal on the current device. You can
+ * retrieve a copy of this structure using the vkGetPhysicalDeviceMetalFeaturesMVK() function.
+ *
+ * This structure may be extended as new features are added to MoltenVK. If you are linking to
+ * an implementation of MoltenVK that was compiled from a different VK_MVK_MOLTENVK_SPEC_VERSION
+ * than your app was, the size of this structure in your app may be larger or smaller than the
+ * struct in MoltenVK. See the description of the vkGetPhysicalDeviceMetalFeaturesMVK() function
+ * for information about how to handle this.
+ *
+ * TO SUPPORT DYNAMIC LINKING TO THIS STRUCTURE AS DESCRIBED ABOVE, THIS STRUCTURE SHOULD NOT
+ * BE CHANGED EXCEPT TO ADD ADDITIONAL MEMBERS ON THE END. EXISTING MEMBERS, AND THEIR ORDER,
+ * SHOULD NOT BE CHANGED.
+ */
+typedef struct {
+    uint32_t mslVersion;                        /**< The version of the Metal Shading Language available on this device. The format of the integer is MMmmpp, with two decimal digts each for Major, minor, and patch version values (eg. MSL 1.2 would appear as 010200). */
+	VkBool32 indirectDrawing;                   /**< If true, draw calls support parameters held in a GPU buffer. */
+	VkBool32 baseVertexInstanceDrawing;         /**< If true, draw calls support specifiying the base vertex and instance. */
+    uint32_t dynamicMTLBufferSize;              /**< If greater than zero, dynamic MTLBuffers for setting vertex, fragment, and compute bytes are supported, and their content must be below this value. */
+    VkBool32 shaderSpecialization;              /**< If true, shader specialization (aka Metal function constants) is supported. */
+    VkBool32 ioSurfaces;                        /**< If true, VkImages can be underlaid by IOSurfaces via the vkUseIOSurfaceMVK() function, to support inter-process image transfers. */
+    VkBool32 texelBuffers;                      /**< If true, texel buffers are supported, allowing the contents of a buffer to be interpreted as an image via a VkBufferView. */
+	VkBool32 layeredRendering;                  /**< If true, layered rendering to multiple cube or texture array layers is supported. */
+	VkBool32 presentModeImmediate;              /**< If true, immediate surface present mode (VK_PRESENT_MODE_IMMEDIATE_KHR), allowing a swapchain image to be presented immediately, without waiting for the vertical sync period of the display, is supported. */
+	VkBool32 stencilViews;                      /**< If true, stencil aspect views are supported through the MTLPixelFormatX24_Stencil8 and MTLPixelFormatX32_Stencil8 formats. */
+	VkBool32 multisampleArrayTextures;          /**< If true, MTLTextureType2DMultisampleArray is supported. */
+	VkBool32 samplerClampToBorder;              /**< If true, the border color set when creating a sampler will be respected. */
+	uint32_t maxTextureDimension; 	     	  	/**< The maximum size of each texture dimension (width, height, or depth). */
+	uint32_t maxPerStageBufferCount;            /**< The total number of per-stage Metal buffers available for shader uniform content and attributes. */
+    uint32_t maxPerStageTextureCount;           /**< The total number of per-stage Metal textures available for shader uniform content. */
+    uint32_t maxPerStageSamplerCount;           /**< The total number of per-stage Metal samplers available for shader uniform content. */
+    VkDeviceSize maxMTLBufferSize;              /**< The max size of a MTLBuffer (in bytes). */
+    VkDeviceSize mtlBufferAlignment;            /**< The alignment used when allocating memory for MTLBuffers. Must be PoT. */
+    VkDeviceSize maxQueryBufferSize;            /**< The maximum size of an occlusion query buffer (in bytes). */
+	VkDeviceSize mtlCopyBufferAlignment;        /**< The alignment required during buffer copy operations (in bytes). */
+    VkSampleCountFlags supportedSampleCounts;   /**< A bitmask identifying the sample counts supported by the device. */
+	uint32_t minSwapchainImageCount;	 	  	/**< The minimum number of swapchain images that can be supported by a surface. */
+	uint32_t maxSwapchainImageCount;	 	  	/**< The maximum number of swapchain images that can be supported by a surface. */
+	VkBool32 combinedStoreResolveAction;		/**< If true, the device supports VK_ATTACHMENT_STORE_OP_STORE with a simultaneous resolve attachment. */
+	VkBool32 arrayOfTextures;			 	  	/**< If true, arrays of textures is supported. */
+	VkBool32 arrayOfSamplers;			 	  	/**< If true, arrays of texture samplers is supported. */
+	MTLLanguageVersion mslVersionEnum;			/**< The version of the Metal Shading Language available on this device, as a Metal enumeration. */
+	VkBool32 depthSampleCompare;				/**< If true, depth texture samplers support the comparison of the pixel value against a reference value. */
+	VkBool32 events;							/**< If true, Metal synchronization events (MTLEvent) are supported. */
+	VkBool32 memoryBarriers;					/**< If true, full memory barriers within Metal render passes are supported. */
+	VkBool32 multisampleLayeredRendering;       /**< If true, layered rendering to multiple multi-sampled cube or texture array layers is supported. */
+	VkBool32 stencilFeedback;					/**< If true, fragment shaders that write to [[stencil]] outputs are supported. */
+	VkBool32 textureBuffers;					/**< If true, textures of type MTLTextureTypeBuffer are supported. */
+	VkBool32 postDepthCoverage;					/**< If true, coverage masks in fragment shaders post-depth-test are supported. */
+	VkBool32 fences;							/**< If true, Metal synchronization fences (MTLFence) are supported. */
+	VkBool32 rasterOrderGroups;					/**< If true, Raster order groups in fragment shaders are supported. */
+	VkBool32 native3DCompressedTextures;		/**< If true, 3D compressed images are supported natively, without manual decompression. */
+	VkBool32 nativeTextureSwizzle;				/**< If true, component swizzle is supported natively, without manual swizzling in shaders. */
+	VkBool32 placementHeaps;					/**< If true, MTLHeap objects support placement of resources. */
+	VkDeviceSize pushConstantSizeAlignment;     /**< The alignment used internally when allocating memory for push constants. Must be PoT. */
+} MVKPhysicalDeviceMetalFeatures;
+
+/**
+ * MoltenVK swapchain performance statistics. You can retrieve a copy of this structure using
+ * the vkGetSwapchainPerformanceMVK() function.
+ *
+ * This structure may be extended as new features are added to MoltenVK. If you are linking to
+ * an implementation of MoltenVK that was compiled from a different VK_MVK_MOLTENVK_SPEC_VERSION
+ * than your app was, the size of this structure in your app may be larger or smaller than the
+ * struct in MoltenVK. See the description of the vkGetSwapchainPerformanceMVK() function for
+ * information about how to handle this.
+ *
+ * TO SUPPORT DYNAMIC LINKING TO THIS STRUCTURE AS DESCRIBED ABOVE, THIS STRUCTURE SHOULD NOT
+ * BE CHANGED EXCEPT TO ADD ADDITIONAL MEMBERS ON THE END. EXISTING MEMBERS, AND THEIR ORDER,
+ * SHOULD NOT BE CHANGED.
+ */
+typedef struct {
+    double lastFrameInterval;           /**< The time interval between this frame and the immediately previous frame, in milliseconds. */
+    double averageFrameInterval;        /**< The rolling average time interval between frames, in miliseconds. This value has less volatility than the lastFrameInterval value. */
+    double averageFramesPerSecond;      /**< The rolling average number of frames per second. This is simply the 1000 divided by the averageFrameInterval value. */
+} MVKSwapchainPerformance;
+
+/** MoltenVK performance of a particular type of activity. */
+typedef struct {
+    uint32_t count;             /**< The number of activities of this type. */
+    double averageDuration;     /**< The average duration of the activity, in milliseconds. */
+    double minimumDuration;     /**< The minimum duration of the activity, in milliseconds. */
+    double maximumDuration;     /**< The maximum duration of the activity, in milliseconds. */
+} MVKPerformanceTracker;
+
+/** MoltenVK performance of shader compilation activities. */
+typedef struct {
+	MVKPerformanceTracker hashShaderCode;				/** Create a hash from the incoming shader code. */
+    MVKPerformanceTracker spirvToMSL;					/** Convert SPIR-V to MSL source code. */
+    MVKPerformanceTracker mslCompile;					/** Compile MSL source code into a MTLLibrary. */
+    MVKPerformanceTracker mslLoad;						/** Load pre-compiled MSL code into a MTLLibrary. */
+	MVKPerformanceTracker shaderLibraryFromCache;		/** Retrieve a shader library from the cache, lazily creating it if needed. */
+    MVKPerformanceTracker functionRetrieval;			/** Retrieve a MTLFunction from a MTLLibrary. */
+    MVKPerformanceTracker functionSpecialization;		/** Specialize a retrieved MTLFunction. */
+    MVKPerformanceTracker pipelineCompile;				/** Compile MTLFunctions into a pipeline. */
+	MVKPerformanceTracker glslToSPRIV;					/** Convert GLSL to SPIR-V code. */
+} MVKShaderCompilationPerformance;
+
+/** MoltenVK performance of pipeline cache activities. */
+typedef struct {
+	MVKPerformanceTracker sizePipelineCache;			/** Calculate the size of cache data required to write MSL to pipeline cache data stream. */
+	MVKPerformanceTracker writePipelineCache;			/** Write MSL to pipeline cache data stream. */
+	MVKPerformanceTracker readPipelineCache;			/** Read MSL from pipeline cache data stream. */
+} MVKPipelineCachePerformance;
+
+/** MoltenVK performance of queue activities. */
+typedef struct {
+	MVKPerformanceTracker mtlQueueAccess;               /** Create an MTLCommmandQueue or access an existing cached instance. */
+	MVKPerformanceTracker mtlCommandBufferCompletion;   /** Completion of a MTLCommandBuffer on the GPU, from commit to completion callback. */
+	MVKPerformanceTracker nextCAMetalDrawable;			/** Retrieve next CAMetalDrawable from CAMetalLayer during presentation. */
+} MVKQueuePerformance;
+
+/**
+ * MoltenVK performance. You can retrieve a copy of this structure using the vkGetPerformanceStatisticsMVK() function.
+ *
+ * This structure may be extended as new features are added to MoltenVK. If you are linking to
+ * an implementation of MoltenVK that was compiled from a different VK_MVK_MOLTENVK_SPEC_VERSION
+ * than your app was, the size of this structure in your app may be larger or smaller than the
+ * struct in MoltenVK. See the description of the vkGetPerformanceStatisticsMVK() function for
+ * information about how to handle this.
+ *
+ * TO SUPPORT DYNAMIC LINKING TO THIS STRUCTURE AS DESCRIBED ABOVE, THIS STRUCTURE SHOULD NOT
+ * BE CHANGED EXCEPT TO ADD ADDITIONAL MEMBERS ON THE END. EXISTING MEMBERS, AND THEIR ORDER,
+ * SHOULD NOT BE CHANGED.
+ */
+typedef struct {
+	MVKShaderCompilationPerformance shaderCompilation;	/** Shader compilations activities. */
+	MVKPipelineCachePerformance pipelineCache;			/** Pipeline cache activities. */
+	MVKQueuePerformance queue;          				/** Queue activities. */
+} MVKPerformanceStatistics;
+
+
+#pragma mark -
+#pragma mark Function types
+
+typedef VkResult (VKAPI_PTR *PFN_vkGetMoltenVKConfigurationMVK)(VkInstance instance, MVKConfiguration* pConfiguration, size_t* pConfigurationSize);
+typedef VkResult (VKAPI_PTR *PFN_vkSetMoltenVKConfigurationMVK)(VkInstance instance, MVKConfiguration* pConfiguration, size_t* pConfigurationSize);
+typedef VkResult (VKAPI_PTR *PFN_vkGetPhysicalDeviceMetalFeaturesMVK)(VkPhysicalDevice physicalDevice, MVKPhysicalDeviceMetalFeatures* pMetalFeatures, size_t* pMetalFeaturesSize);
+typedef VkResult (VKAPI_PTR *PFN_vkGetSwapchainPerformanceMVK)(VkDevice device, VkSwapchainKHR swapchain, MVKSwapchainPerformance* pSwapchainPerf, size_t* pSwapchainPerfSize);
+typedef VkResult (VKAPI_PTR *PFN_vkGetPerformanceStatisticsMVK)(VkDevice device, MVKPerformanceStatistics* pPerf, size_t* pPerfSize);
+typedef void (VKAPI_PTR *PFN_vkGetVersionStringsMVK)(char* pMoltenVersionStringBuffer, uint32_t moltenVersionStringBufferLength, char* pVulkanVersionStringBuffer, uint32_t vulkanVersionStringBufferLength);
+
+#ifdef __OBJC__
+typedef void (VKAPI_PTR *PFN_vkGetMTLDeviceMVK)(VkPhysicalDevice physicalDevice, id<MTLDevice>* pMTLDevice);
+typedef VkResult (VKAPI_PTR *PFN_vkSetMTLTextureMVK)(VkImage image, id<MTLTexture> mtlTexture);
+typedef void (VKAPI_PTR *PFN_vkGetMTLTextureMVK)(VkImage image, id<MTLTexture>* pMTLTexture);
+typedef void (VKAPI_PTR *PFN_vkGetMTLBufferMVK)(VkBuffer buffer, id<MTLBuffer>* pMTLBuffer);
+typedef VkResult (VKAPI_PTR *PFN_vkUseIOSurfaceMVK)(VkImage image, IOSurfaceRef ioSurface);
+typedef void (VKAPI_PTR *PFN_vkGetIOSurfaceMVK)(VkImage image, IOSurfaceRef* pIOSurface);
+#endif // __OBJC__
+
+
+#pragma mark -
+#pragma mark Function prototypes
+
+#ifndef VK_NO_PROTOTYPES
+
+/** 
+ * Populates the pConfiguration structure with the current MoltenVK configuration settings.
+ *
+ * To change a specific configuration value, call vkGetMoltenVKConfigurationMVK() to retrieve
+ * the current configuration, make changes, and call  vkSetMoltenVKConfigurationMVK() to
+ * update all of the values.
+ *
+ * To be active, some configuration settings must be set before a VkDevice is created.
+ * See the description of the MVKConfiguration members for more information.
+ *
+ * If you are linking to an implementation of MoltenVK that was compiled from a different
+ * VK_MVK_MOLTENVK_SPEC_VERSION than your app was, the size of the MVKConfiguration structure
+ * in your app may be larger or smaller than the same struct as expected by MoltenVK.
+ *
+ * When calling this function, set the value of *pConfigurationSize to sizeof(MVKConfiguration),
+ * to tell MoltenVK the limit of the size of your MVKConfiguration structure. Upon return from
+ * this function, the value of *pConfigurationSize will hold the actual number of bytes copied
+ * into your passed MVKConfiguration structure, which will be the smaller of what your app
+ * thinks is the size of MVKConfiguration, and what MoltenVK thinks it is. This represents the
+ * safe access area within the structure for both MoltenVK and your app.
+ *
+ * If the size that MoltenVK expects for MVKConfiguration is different than the value passed in
+ * *pConfigurationSize, this function will return VK_INCOMPLETE, otherwise it will return VK_SUCCESS.
+ *
+ * Although it is not necessary, you can use this function to determine in advance the value
+ * that MoltenVK expects the size of MVKConfiguration to be by setting the value of pConfiguration
+ * to NULL. In that case, this function will set *pConfigurationSize to the size that MoltenVK
+ * expects MVKConfiguration to be.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkInstance object you provide here must have been retrieved directly from MoltenVK,
+ * and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan objects
+ * are often changed by layers, and passing them from one layer to another, or from
+ * a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkGetMoltenVKConfigurationMVK(
+	VkInstance                                  instance,
+	MVKConfiguration*                           pConfiguration,
+	size_t*                                     pConfigurationSize);
+
+/** 
+ * Sets the MoltenVK configuration settings to those found in the pConfiguration structure.
+ *
+ * To change a specific configuration value, call vkGetMoltenVKConfigurationMVK()
+ * to retrieve the current configuration, make changes, and call
+ * vkSetMoltenVKConfigurationMVK() to update all of the values.
+ *
+ * To be active, some configuration settings must be set before a VkDevice is created.
+ * See the description of the MVKConfiguration members for more information.
+ *
+ * If you are linking to an implementation of MoltenVK that was compiled from a different
+ * VK_MVK_MOLTENVK_SPEC_VERSION than your app was, the size of the MVKConfiguration structure
+ * in your app may be larger or smaller than the same struct as expected by MoltenVK.
+ *
+ * When calling this function, set the value of *pConfigurationSize to sizeof(MVKConfiguration),
+ * to tell MoltenVK the limit of the size of your MVKConfiguration structure. Upon return from
+ * this function, the value of *pConfigurationSize will hold the actual number of bytes copied
+ * out of your passed MVKConfiguration structure, which will be the smaller of what your app
+ * thinks is the size of MVKConfiguration, and what MoltenVK thinks it is. This represents the
+ * safe access area within the structure for both MoltenVK and your app.
+ *
+ * If the size that MoltenVK expects for MVKConfiguration is different than the value passed in
+ * *pConfigurationSize, this function will return VK_INCOMPLETE, otherwise it will return VK_SUCCESS.
+ *
+ * Although it is not necessary, you can use this function to determine in advance the value
+ * that MoltenVK expects the size of MVKConfiguration to be by setting the value of pConfiguration
+ * to NULL. In that case, this function will set *pConfigurationSize to the size that MoltenVK
+ * expects MVKConfiguration to be.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkInstance object you provide here must have been retrieved directly from MoltenVK,
+ * and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan objects
+ * are often changed by layers, and passing them from one layer to another, or from
+ * a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkSetMoltenVKConfigurationMVK(
+	VkInstance                                  instance,
+	const MVKConfiguration*                     pConfiguration,
+	size_t*                                     pConfigurationSize);
+
+/** 
+ * Populates the pMetalFeatures structure with the Metal-specific features
+ * supported by the specified physical device. 
+ *
+ * If you are linking to an implementation of MoltenVK that was compiled from a different
+ * VK_MVK_MOLTENVK_SPEC_VERSION than your app was, the size of the MVKPhysicalDeviceMetalFeatures
+ * structure in your app may be larger or smaller than the same struct as expected by MoltenVK.
+ *
+ * When calling this function, set the value of *pMetalFeaturesSize to sizeof(MVKPhysicalDeviceMetalFeatures),
+ * to tell MoltenVK the limit of the size of your MVKPhysicalDeviceMetalFeatures structure. Upon return from
+ * this function, the value of *pMetalFeaturesSize will hold the actual number of bytes copied into your
+ * passed MVKPhysicalDeviceMetalFeatures structure, which will be the smaller of what your app thinks is the
+ * size of MVKPhysicalDeviceMetalFeatures, and what MoltenVK thinks it is. This represents the safe access
+ * area within the structure for both MoltenVK and your app.
+ *
+ * If the size that MoltenVK expects for MVKPhysicalDeviceMetalFeatures is different than the value passed in
+ * *pMetalFeaturesSize, this function will return VK_INCOMPLETE, otherwise it will return VK_SUCCESS.
+ *
+ * Although it is not necessary, you can use this function to determine in advance the value that MoltenVK
+ * expects the size of MVKPhysicalDeviceMetalFeatures to be by setting the value of pMetalFeatures to NULL.
+ * In that case, this function will set *pMetalFeaturesSize to the size that MoltenVK expects
+ * MVKPhysicalDeviceMetalFeatures to be.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkPhysicalDevice object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkGetPhysicalDeviceMetalFeaturesMVK(
+	VkPhysicalDevice                            physicalDevice,
+	MVKPhysicalDeviceMetalFeatures*             pMetalFeatures,
+	size_t*                                     pMetalFeaturesSize);
+
+/**
+ * Populates the pSwapchainPerf structure with the current performance statistics for the swapchain.
+ *
+ * If you are linking to an implementation of MoltenVK that was compiled from a different
+ * VK_MVK_MOLTENVK_SPEC_VERSION than your app was, the size of the MVKSwapchainPerformance
+ * structure in your app may be larger or smaller than the same struct as expected by MoltenVK.
+ *
+ * When calling this function, set the value of *pSwapchainPerfSize to sizeof(MVKSwapchainPerformance),
+ * to tell MoltenVK the limit of the size of your MVKSwapchainPerformance structure. Upon return from
+ * this function, the value of *pSwapchainPerfSize will hold the actual number of bytes copied into
+ * your passed MVKSwapchainPerformance structure, which will be the smaller of what your app thinks
+ * is the size of MVKSwapchainPerformance, and what MoltenVK thinks it is. This represents the safe
+ * access area within the structure for both MoltenVK and your app.
+ *
+ * If the size that MoltenVK expects for MVKSwapchainPerformance is different than the value passed in
+ * *pSwapchainPerfSize, this function will return VK_INCOMPLETE, otherwise it will return VK_SUCCESS.
+ *
+ * Although it is not necessary, you can use this function to determine in advance the value
+ * that MoltenVK expects the size of MVKSwapchainPerformance to be by setting the value of
+ * pSwapchainPerf to NULL. In that case, this function will set *pSwapchainPerfSize to the
+ * size that MoltenVK expects MVKSwapchainPerformance to be.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkDevice and VkSwapchainKHR objects you provide here must have been retrieved directly
+ * from MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkGetSwapchainPerformanceMVK(
+	VkDevice                                    device,
+	VkSwapchainKHR                              swapchain,
+	MVKSwapchainPerformance*                    pSwapchainPerf,
+	size_t*                                     pSwapchainPerfSize);
+
+/**
+ * Populates the pPerf structure with the current performance statistics for the device.
+ *
+ * If you are linking to an implementation of MoltenVK that was compiled from a different
+ * VK_MVK_MOLTENVK_SPEC_VERSION than your app was, the size of the MVKPerformanceStatistics
+ * structure in your app may be larger or smaller than the same struct as expected by MoltenVK.
+ *
+ * When calling this function, set the value of *pPerfSize to sizeof(MVKPerformanceStatistics),
+ * to tell MoltenVK the limit of the size of your MVKPerformanceStatistics structure. Upon return
+ * from this function, the value of *pPerfSize will hold the actual number of bytes copied into
+ * your passed MVKPerformanceStatistics structure, which will be the smaller of what your app
+ * thinks is the size of MVKPerformanceStatistics, and what MoltenVK thinks it is. This
+ * represents the safe access area within the structure for both MoltenVK and your app.
+ *
+ * If the size that MoltenVK expects for MVKPerformanceStatistics is different than the value passed
+ * in *pPerfSize, this function will return VK_INCOMPLETE, otherwise it will return VK_SUCCESS.
+ *
+ * Although it is not necessary, you can use this function to determine in advance the value
+ * that MoltenVK expects the size of MVKPerformanceStatistics to be by setting the value of
+ * pPerf to NULL. In that case, this function will set *pPerfSize to the size that MoltenVK
+ * expects MVKPerformanceStatistics to be.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkDevice object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkGetPerformanceStatisticsMVK(
+	VkDevice                                    device,
+	MVKPerformanceStatistics*            		pPerf,
+	size_t*                                     pPerfSize);
+
+/**
+ * Returns a human readable version of the MoltenVK and Vulkan versions.
+ *
+ * This function is provided as a convenience for reporting. Use the MVK_VERSION, 
+ * VK_API_VERSION_1_0, and VK_HEADER_VERSION macros for programmatically accessing
+ * the corresponding version numbers.
+ */
+VKAPI_ATTR void VKAPI_CALL vkGetVersionStringsMVK(
+    char*                                       pMoltenVersionStringBuffer,
+    uint32_t                                    moltenVersionStringBufferLength,
+    char*                                       pVulkanVersionStringBuffer,
+    uint32_t                                    vulkanVersionStringBufferLength);
+
+/**
+ * Sets the number of threads in a workgroup for a compute kernel.
+ *
+ * This needs to be called if you are creating compute shader modules from MSL
+ * source code or MSL compiled code. Workgroup size is determined automatically
+ * if you're using SPIR-V.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkShaderModule object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR void VKAPI_CALL vkSetWorkgroupSizeMVK(
+    VkShaderModule                              shaderModule,
+    uint32_t                                    x,
+    uint32_t                                    y,
+    uint32_t                                    z);
+
+#ifdef __OBJC__
+
+/**
+ * Returns, in the pMTLDevice pointer, the MTLDevice used by the VkPhysicalDevice.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkPhysicalDevice object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR void VKAPI_CALL vkGetMTLDeviceMVK(
+    VkPhysicalDevice                           physicalDevice,
+    id<MTLDevice>*                             pMTLDevice);
+
+/**
+ * Sets the VkImage to use the specified MTLTexture.
+ *
+ * Any differences in the properties of mtlTexture and this image will modify the
+ * properties of this image.
+ *
+ * If a MTLTexture has already been created for this image, it will be destroyed.
+ *
+ * Returns VK_SUCCESS.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkImage object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkSetMTLTextureMVK(
+    VkImage                                     image,
+    id<MTLTexture>                              mtlTexture);
+
+/**
+ * Returns, in the pMTLTexture pointer, the MTLTexture currently underlaying the VkImage.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkImage object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR void VKAPI_CALL vkGetMTLTextureMVK(
+    VkImage                                     image,
+    id<MTLTexture>*                             pMTLTexture);
+
+/**
+* Returns, in the pMTLBuffer pointer, the MTLBuffer currently underlaying the VkBuffer.
+*
+* This function is not supported by the Vulkan SDK Loader and Layers framework.
+* The VkBuffer object you provide here must have been retrieved directly from
+* MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+* objects are often changed by layers, and passing them from one layer to another,
+* or from a layer directly to MoltenVK, will result in undefined behaviour.
+*/
+VKAPI_ATTR void VKAPI_CALL vkGetMTLBufferMVK(
+    VkBuffer                                    buffer,
+    id<MTLBuffer>*                              pMTLBuffer);
+
+/**
+ * Indicates that a VkImage should use an IOSurface to underlay the Metal texture.
+ *
+ * If ioSurface is not null, it will be used as the IOSurface, and any differences
+ * in the properties of that IOSurface will modify the properties of this image.
+ *
+ * If ioSurface is null, this image will create and use an IOSurface
+ * whose properties are compatible with the properties of this image.
+ *
+ * If a MTLTexture has already been created for this image, it will be destroyed.
+ *
+ * IOSurfaces are supported on the following platforms:
+ *   -  macOS 10.11 and above
+ *   -  iOS 11.0 and above
+ *
+ * To enable IOSurface support, ensure the Deployment Target build setting
+ * (MACOSX_DEPLOYMENT_TARGET or IPHONEOS_DEPLOYMENT_TARGET) is set to at least
+ * one of the values above when compiling MoltenVK, and any app that uses MoltenVK.
+ *
+ * Returns:
+ *   - VK_SUCCESS.
+ *   - VK_ERROR_FEATURE_NOT_PRESENT if IOSurfaces are not supported on the platform.
+ *   - VK_ERROR_INITIALIZATION_FAILED if ioSurface is specified and is not compatible with this VkImage.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkImage object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR VkResult VKAPI_CALL vkUseIOSurfaceMVK(
+    VkImage                                     image,
+    IOSurfaceRef                                ioSurface);
+
+/**
+ * Returns, in the pIOSurface pointer, the IOSurface currently underlaying the VkImage,
+ * as set by the useIOSurfaceMVK() function, or returns null if the VkImage is not using
+ * an IOSurface, or if the platform does not support IOSurfaces.
+ *
+ * This function is not supported by the Vulkan SDK Loader and Layers framework.
+ * The VkImage object you provide here must have been retrieved directly from
+ * MoltenVK, and not through the Vulkan SDK Loader and Layers framework. Opaque Vulkan
+ * objects are often changed by layers, and passing them from one layer to another,
+ * or from a layer directly to MoltenVK, will result in undefined behaviour.
+ */
+VKAPI_ATTR void VKAPI_CALL vkGetIOSurfaceMVK(
+    VkImage                                     image,
+    IOSurfaceRef*                               pIOSurface);
+
+#endif // __OBJC__
+
+
+#pragma mark -
+#pragma mark Shaders
+
+/**
+ * Enumerates the magic number values to set in the MVKMSLSPIRVHeader when
+ * submitting a SPIR-V stream that contains either Metal Shading Language source
+ * code or Metal Shading Language compiled binary code in place of SPIR-V code.
+ */
+typedef enum {
+    kMVKMagicNumberSPIRVCode        = 0x07230203,    /**< SPIR-V stream contains standard SPIR-V code. */
+    kMVKMagicNumberMSLSourceCode    = 0x19960412,    /**< SPIR-V stream contains Metal Shading Language source code. */
+    kMVKMagicNumberMSLCompiledCode  = 0x19981215,    /**< SPIR-V stream contains Metal Shading Language compiled binary code. */
+} MVKMSLMagicNumber;
+
+/**
+ * Describes the header at the start of an SPIR-V stream, when it contains either
+ * Metal Shading Language source code or Metal Shading Language compiled binary code.
+ *
+ * To submit MSL source code to the vkCreateShaderModule() function in place of SPIR-V
+ * code, prepend a MVKMSLSPIRVHeader containing the kMVKMagicNumberMSLSourceCode magic
+ * number to the MSL source code. The MSL source code must be null-terminated.
+ *
+ * To submit MSL compiled binary code to the vkCreateShaderModule() function in place of
+ * SPIR-V code, prepend a MVKMSLSPIRVHeader containing the kMVKMagicNumberMSLCompiledCode
+ * magic number to the MSL compiled binary code.
+ *
+ * In both cases, the pCode element of VkShaderModuleCreateInfo should pointer to the
+ * location of the MVKMSLSPIRVHeader, and the MSL code should start at the byte immediately
+ * after the MVKMSLSPIRVHeader.
+ *
+ * The codeSize element of VkShaderModuleCreateInfo should be set to the entire size of
+ * the submitted code memory, including the additional sizeof(MVKMSLSPIRVHeader) bytes
+ * taken up by the MVKMSLSPIRVHeader, and, in the case of MSL source code, including
+ * the null-terminator byte.
+ */
+typedef uint32_t MVKMSLSPIRVHeader;
+
+
+#endif // VK_NO_PROTOTYPES
+
+
+#ifdef __cplusplus
+}
+#endif	//  __cplusplus
+
+#endif
diff --git a/externals/MoltenVK/include/vulkan-portability/vk_extx_portability_subset.h b/externals/MoltenVK/include/vulkan-portability/vk_extx_portability_subset.h
new file mode 100644
index 0000000000..b6f7f449d5
--- /dev/null
+++ b/externals/MoltenVK/include/vulkan-portability/vk_extx_portability_subset.h
@@ -0,0 +1,76 @@
+#ifndef VK_EXTX_PORTABILITY_SUBSET_H_
+#define VK_EXTX_PORTABILITY_SUBSET_H_ 1
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+** Copyright (c) 2018 The Khronos Group Inc.
+**
+** 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.
+*/
+
+/*
+ Please Note:  This extension is currently defined as "EXTX", meaning "multivendor experimental".
+ That means the definition of this extension is in active development, and may break compatibility
+ between point releases (defined as any increment of VK_EXTX_PORTABILITY_SUBSET_SPEC_VERSION).
+ You are free to explore the extension and provide feedback, but it is not recommended to use this
+ extension for shipping applications, particularly applications that require the driver implementing this
+ extension to be linked dynamically and potentially "dropped-in" to the application execution environment.
+ */
+
+#include "vulkan/vulkan.h"
+
+#define VK_EXTX_PORTABILITY_SUBSET_SPEC_VERSION       1
+#define VK_EXTX_PORTABILITY_SUBSET_EXTENSION_NAME     "VK_EXTX_portability_subset"
+
+#define VK_EXTX_PORTABILITY_SUBSET_EXTENSION_ID 164
+// See enum_offset() from https://www.khronos.org/registry/vulkan/specs/1.1/styleguide.html#_assigning_extension_token_values
+#define VK_EXTX_PORTABILITY_SUBSET_STYPE_ID(x) \
+    ((VkStructureType)(1000000000 + 1000 * (VK_EXTX_PORTABILITY_SUBSET_EXTENSION_ID - 1) + x))
+#define VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PORTABILITY_SUBSET_FEATURES_EXTX      VK_EXTX_PORTABILITY_SUBSET_STYPE_ID(0)
+#define VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PORTABILITY_SUBSET_PROPERTIES_EXTX    VK_EXTX_PORTABILITY_SUBSET_STYPE_ID(1)
+#define VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGE_VIEW_SUPPORT_EXTX               VK_EXTX_PORTABILITY_SUBSET_STYPE_ID(2)
+
+typedef struct VkPhysicalDevicePortabilitySubsetFeaturesEXTX {
+     VkStructureType    sType;
+     void*              pNext;
+     VkBool32           triangleFans;
+     VkBool32           separateStencilMaskRef;
+     VkBool32           events;
+     VkBool32           standardImageViews;
+     VkBool32           samplerMipLodBias;
+} VkPhysicalDevicePortabilitySubsetFeaturesEXTX;
+
+typedef struct VkPhysicalDevicePortabilitySubsetPropertiesEXTX {
+     VkStructureType    sType;
+     void*              pNext;
+     uint32_t           minVertexInputBindingStrideAlignment;
+} VkPhysicalDevicePortabilitySubsetPropertiesEXTX;
+
+typedef struct VkPhysicalDeviceImageViewSupportEXTX {
+     VkStructureType        sType;
+     void*                  pNext;
+     VkImageViewCreateFlags flags;
+     VkImageViewType        viewType;
+     VkFormat               format;
+     VkComponentMapping     components;
+     VkImageAspectFlags     aspectMask;
+} VkPhysicalDeviceImageViewSupportEXTX;
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif // VK_EXTX_PORTABILITY_SUBSET_H_
diff --git a/externals/MoltenVK/lib/libMoltenVK.dylib b/externals/MoltenVK/lib/libMoltenVK.dylib
new file mode 100755
index 0000000000..4860478731
Binary files /dev/null and b/externals/MoltenVK/lib/libMoltenVK.dylib differ
diff --git a/externals/MoltenVK/lib/libVkLayer_khronos_validation.dylib b/externals/MoltenVK/lib/libVkLayer_khronos_validation.dylib
new file mode 100755
index 0000000000..3f7c74ba38
Binary files /dev/null and b/externals/MoltenVK/lib/libVkLayer_khronos_validation.dylib differ
diff --git a/externals/MoltenVK/lib/libvulkan.dylib b/externals/MoltenVK/lib/libvulkan.dylib
new file mode 100755
index 0000000000..3eb6e1d694
Binary files /dev/null and b/externals/MoltenVK/lib/libvulkan.dylib differ
diff --git a/src/core/file_sys/system_archive/shared_font.cpp b/src/core/file_sys/system_archive/shared_font.cpp
index 2c05eb42e2..2d698bbd55 100644
--- a/src/core/file_sys/system_archive/shared_font.cpp
+++ b/src/core/file_sys/system_archive/shared_font.cpp
@@ -23,7 +23,7 @@ VirtualFile PackBFTTF(const std::array<u8, Size>& data, const std::string& name)
 
     std::vector<u8> bfttf(Size + sizeof(u64));
 
-    u64 offset = 0;
+    std::size_t offset = 0;
     Service::NS::EncryptSharedFont(vec, bfttf, offset);
     return std::make_shared<VectorVfsFile>(std::move(bfttf), name);
 }
diff --git a/src/core/frontend/emu_window.h b/src/core/frontend/emu_window.h
index 8749ac95f2..331d26e306 100644
--- a/src/core/frontend/emu_window.h
+++ b/src/core/frontend/emu_window.h
@@ -19,6 +19,7 @@ enum class WindowSystemType {
     Windows,
     X11,
     Wayland,
+    MacOS,
 };
 
 /**
@@ -113,7 +114,7 @@ public:
     /**
      * Returns system information about the drawing area.
      */
-    const WindowSystemInfo& GetWindowInfo() const {
+    WindowSystemInfo& GetWindowInfo() {
         return window_info;
     }
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
index bdae8b8875..0dde16a6ed 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
@@ -24,7 +24,7 @@ u32 nvhost_nvdec::ioctl(Ioctl command, const std::vector<u8>& input, const std::
         return SetNVMAPfd(input, output);
     }
 
-    UNIMPLEMENTED_MSG("Unimplemented ioctl");
+    //UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 }
 
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
index c695b88632..8b4ded6383 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
@@ -24,7 +24,7 @@ u32 nvhost_vic::ioctl(Ioctl command, const std::vector<u8>& input, const std::ve
         return SetNVMAPfd(input, output);
     }
 
-    UNIMPLEMENTED_MSG("Unimplemented ioctl");
+    //UNIMPLEMENTED_MSG("Unimplemented ioctl");
     return 0;
 }
 
diff --git a/src/video_core/CMakeLists.txt b/src/video_core/CMakeLists.txt
index 44344117f7..f1b6f15307 100644
--- a/src/video_core/CMakeLists.txt
+++ b/src/video_core/CMakeLists.txt
@@ -217,6 +217,10 @@ if (ENABLE_VULKAN)
     target_compile_definitions(video_core PRIVATE HAS_VULKAN)
 endif()
 
+if (APPLE)
+  target_include_directories(video_core PRIVATE ../../externals/MoltenVK/include)
+endif()
+
 create_target_directory_groups(video_core)
 
 target_link_libraries(video_core PUBLIC common core)
diff --git a/src/video_core/query_cache.h b/src/video_core/query_cache.h
index e66054ed06..d144c53a51 100644
--- a/src/video_core/query_cache.h
+++ b/src/video_core/query_cache.h
@@ -21,6 +21,13 @@
 #include "video_core/memory_manager.h"
 #include "video_core/rasterizer_interface.h"
 
+#if defined(PAGE_SIZE)
+#undef PAGE_SIZE
+#endif
+#if defined(PAGE_SHIFT)
+#undef PAGE_SHIFT
+#endif
+
 namespace VideoCommon {
 
 template <class QueryCache, class HostCounter>
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index e3d31c3eb9..133c815985 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -117,7 +117,7 @@ constexpr GLenum GetGLShaderType(ShaderType shader_type) {
 /// Hashes one (or two) program streams
 u64 GetUniqueIdentifier(ShaderType shader_type, bool is_a, const ProgramCode& code,
                         const ProgramCode& code_b = {}) {
-    u64 unique_identifier = boost::hash_value(code);
+    std::size_t unique_identifier = boost::hash_value(code);
     if (is_a) {
         // VertexA programs include two programs
         boost::hash_combine(unique_identifier, boost::hash_value(code_b));
diff --git a/src/video_core/renderer_vulkan/renderer_vulkan.cpp b/src/video_core/renderer_vulkan/renderer_vulkan.cpp
index 599bb561ad..e3248101e4 100644
--- a/src/video_core/renderer_vulkan/renderer_vulkan.cpp
+++ b/src/video_core/renderer_vulkan/renderer_vulkan.cpp
@@ -14,6 +14,7 @@
 
 #include "common/assert.h"
 #include "common/dynamic_library.h"
+#include "common/file_util.h"
 #include "common/logging/log.h"
 #include "common/telemetry.h"
 #include "core/core.h"
@@ -47,6 +48,12 @@
 #include <vulkan/vulkan_xlib.h>
 #endif
 
+#ifdef __APPLE__
+#define VK_USE_PLATFORM_METAL_EXT
+#include <objc/message.h>
+#include <vulkan/vulkan_metal.h>
+#endif
+
 namespace Vulkan {
 
 namespace {
@@ -78,7 +85,7 @@ Common::DynamicLibrary OpenVulkanLibrary() {
     char* libvulkan_env = getenv("LIBVULKAN_PATH");
     if (!libvulkan_env || !library.Open(libvulkan_env)) {
         // Use the libvulkan.dylib from the application bundle.
-        std::string filename = File::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib";
+        std::string filename = FileUtil::GetBundleDirectory() + "/Contents/Frameworks/libvulkan.dylib";
         library.Open(filename.c_str());
     }
 #else
@@ -125,6 +132,11 @@ vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatc
     case Core::Frontend::WindowSystemType::Wayland:
         extensions.push_back(VK_KHR_WAYLAND_SURFACE_EXTENSION_NAME);
         break;
+#endif
+#ifdef __APPLE__
+    case Core::Frontend::WindowSystemType::MacOS:
+        extensions.push_back(VK_EXT_METAL_SURFACE_EXTENSION_NAME);
+        break;
 #endif
     default:
         LOG_ERROR(Render_Vulkan, "Presentation not supported on this platform");
@@ -137,6 +149,7 @@ vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatc
         extensions.push_back(VK_EXT_DEBUG_UTILS_EXTENSION_NAME);
     }
     extensions.push_back(VK_KHR_GET_PHYSICAL_DEVICE_PROPERTIES_2_EXTENSION_NAME);
+    extensions.push_back(VK_KHR_GET_SURFACE_CAPABILITIES_2_EXTENSION_NAME);
 
     const std::optional properties = vk::EnumerateInstanceExtensionProperties(dld);
     if (!properties) {
@@ -155,7 +168,29 @@ vk::Instance CreateInstance(Common::DynamicLibrary& library, vk::InstanceDispatc
         }
     }
 
-    static constexpr std::array layers_data{"VK_LAYER_LUNARG_standard_validation"};
+    // TODO reintroduce vk 1.0 loader?
+    // vk::ApplicationInfo application_info;
+    // application_info.pApplicationName = "yuzu Emulator";
+    // application_info.applicationVersion = VK_MAKE_VERSION(0, 1, 0);
+    // application_info.pEngineName = "yuzu Emulator";
+    // application_info.engineVersion = VK_MAKE_VERSION(0, 1, 0);
+
+    // // Try for Vulkan 1.1 if the loader supports it.
+    // if (dld.vkEnumerateInstanceVersion && vk::enumerateInstanceVersion(dld) >= VK_MAKE_VERSION(1, 1, 0)) {
+    //     application_info.apiVersion = VK_MAKE_VERSION(1, 1, 0);
+    // } else {
+    //     LOG_ERROR(Render_Vulkan, "Vulkan 1.1 is not susupported! Try updating your drivers");
+    //     application_info.apiVersion = VK_MAKE_VERSION(1, 0, 0);
+    // }
+    // const std::array layers = {"VK_LAYER_LUNARG_standard_validation"};
+    // const vk::InstanceCreateInfo instance_ci(
+    //     {}, &application_info, enable_layers ? static_cast<u32>(layers.size()) : 0, layers.data(),
+    //     static_cast<u32>(extensions.size()), extensions.data());
+    // vk::Instance unsafe_instance;
+    // if (vk::createInstance(&instance_ci, nullptr, &unsafe_instance, dld) != vk::Result::eSuccess) {
+
+    //static constexpr std::array layers_data{"VK_LAYER_LUNARG_standard_validation"};
+    static constexpr std::array layers_data{"VK_LAYER_KHRONOS_validation"};
     vk::Span<const char*> layers = layers_data;
     if (!enable_layers) {
         layers = {};
@@ -215,6 +250,47 @@ std::string BuildCommaSeparatedExtensions(std::vector<std::string> available_ext
     return separated_extensions;
 }
 
+void PrepareWindow(Core::Frontend::EmuWindow::WindowSystemInfo& wsi) {
+#if defined(VK_USE_PLATFORM_METAL_EXT)
+    // This is kinda messy, but it avoids having to write Objective C++ just to create a metal layer.
+    id view = reinterpret_cast<id>(wsi.render_surface);
+    Class clsCAMetalLayer = objc_getClass("CAMetalLayer");
+    if (!clsCAMetalLayer) {
+        LOG_ERROR(Render_Vulkan, "Failed to get CAMetalLayer class.");
+        return;
+    }
+
+    // [CAMetalLayer layer]
+    id layer = reinterpret_cast<id (*)(Class, SEL)>(objc_msgSend)(objc_getClass("CAMetalLayer"),
+                                                        sel_getUid("layer"));
+    if (!layer) {
+        LOG_ERROR(Render_Vulkan, "Failed to create Metal layer.");
+        return;
+    }
+
+    // [view setWantsLayer:YES]
+    reinterpret_cast<void (*)(id, SEL, BOOL)>(objc_msgSend)(view, sel_getUid("setWantsLayer:"), YES);
+
+    // [view setLayer:layer]
+    reinterpret_cast<void (*)(id, SEL, id)>(objc_msgSend)(view, sel_getUid("setLayer:"), layer);
+
+    // NSScreen* screen = [NSScreen mainScreen]
+    id screen = reinterpret_cast<id (*)(Class, SEL)>(objc_msgSend)(objc_getClass("NSScreen"),
+                                                         sel_getUid("mainScreen"));
+
+    // CGFloat factor = [screen backingScaleFactor]
+    double factor =
+    reinterpret_cast<double (*)(id, SEL)>(objc_msgSend)(screen, sel_getUid("backingScaleFactor"));
+
+    // layer.contentsScale = factor
+    reinterpret_cast<void (*)(id, SEL, double)>(objc_msgSend)(layer, sel_getUid("setContentsScale:"),
+                                                    factor);
+
+    // Store the layer pointer, that way MoltenVK doesn't call [NSView layer] outside the main thread.
+    wsi.render_surface = layer;
+#endif
+}
+
 } // Anonymous namespace
 
 RendererVulkan::RendererVulkan(Core::Frontend::EmuWindow& window, Core::System& system)
@@ -264,6 +340,7 @@ void RendererVulkan::TryPresent(int /*timeout_ms*/) {
 }
 
 bool RendererVulkan::Init() {
+    PrepareWindow(render_window.GetWindowInfo());
     library = OpenVulkanLibrary();
     instance = CreateInstance(library, dld, render_window.GetWindowInfo().type,
                               Settings::values.renderer_debug);
@@ -371,6 +448,23 @@ bool RendererVulkan::CreateSurface() {
             return false;
         }
     }
+#endif
+#ifdef VK_USE_PLATFORM_METAL_EXT
+    if (window_info.type == Core::Frontend::WindowSystemType::MacOS) {
+        VkMetalSurfaceCreateInfoEXT surface_create_info = {
+            VK_STRUCTURE_TYPE_METAL_SURFACE_CREATE_INFO_EXT, nullptr, 0,
+            static_cast<const CAMetalLayer*>(window_info.render_surface)};
+
+        const auto vkCreateMetalSurfaceEXT = reinterpret_cast<PFN_vkCreateMetalSurfaceEXT>(
+            dld.vkGetInstanceProcAddr(*instance, "vkCreateMetalSurfaceEXT"));
+
+        if (!vkCreateMetalSurfaceEXT ||
+            vkCreateMetalSurfaceEXT(*instance, &surface_create_info, nullptr, &unsafe_surface) !=
+                VK_SUCCESS) {
+            LOG_ERROR(Render_Vulkan, "Failed to initialize Metal surface");
+            return false;
+        }
+    }
 #endif
     if (!unsafe_surface) {
         LOG_ERROR(Render_Vulkan, "Presentation not supported on this platform");
diff --git a/src/video_core/renderer_vulkan/vk_device.cpp b/src/video_core/renderer_vulkan/vk_device.cpp
index e1f0c93515..33bbe43205 100644
--- a/src/video_core/renderer_vulkan/vk_device.cpp
+++ b/src/video_core/renderer_vulkan/vk_device.cpp
@@ -16,6 +16,7 @@
 #include "video_core/renderer_vulkan/vk_device.h"
 #include "video_core/renderer_vulkan/wrapper.h"
 
+
 namespace Vulkan {
 
 namespace {
@@ -38,11 +39,11 @@ constexpr std::array REQUIRED_EXTENSIONS = {
     VK_KHR_SHADER_DRAW_PARAMETERS_EXTENSION_NAME,
     VK_KHR_16BIT_STORAGE_EXTENSION_NAME,
     VK_KHR_8BIT_STORAGE_EXTENSION_NAME,
-    VK_KHR_DRIVER_PROPERTIES_EXTENSION_NAME,
+    //VK_KHR_DRIVER_PROPERTIES_EXTENSION_NAME,
     VK_KHR_DESCRIPTOR_UPDATE_TEMPLATE_EXTENSION_NAME,
     VK_EXT_VERTEX_ATTRIBUTE_DIVISOR_EXTENSION_NAME,
-    VK_EXT_SHADER_SUBGROUP_BALLOT_EXTENSION_NAME,
-    VK_EXT_SHADER_SUBGROUP_VOTE_EXTENSION_NAME,
+    //VK_EXT_SHADER_SUBGROUP_BALLOT_EXTENSION_NAME,
+    //VK_EXT_SHADER_SUBGROUP_VOTE_EXTENSION_NAME,
     VK_EXT_HOST_QUERY_RESET_EXTENSION_NAME,
 };
 
@@ -179,7 +180,8 @@ bool VKDevice::Create() {
     features.fullDrawIndexUint32 = false;
     features.imageCubeArray = false;
     features.independentBlend = true;
-    features.geometryShader = true;
+    features.geometryShader = false;
+    //features.geometryShader = true;
     features.tessellationShader = true;
     features.sampleRateShading = false;
     features.dualSrcBlend = false;
@@ -473,7 +475,7 @@ bool VKDevice::IsSuitable(vk::PhysicalDevice physical, VkSurfaceKHR surface) {
         std::make_pair(features.largePoints, "largePoints"),
         std::make_pair(features.multiViewport, "multiViewport"),
         std::make_pair(features.depthBiasClamp, "depthBiasClamp"),
-        std::make_pair(features.geometryShader, "geometryShader"),
+        //std::make_pair(features.geometryShader, "geometryShader"),
         std::make_pair(features.tessellationShader, "tessellationShader"),
         std::make_pair(features.occlusionQueryPrecise, "occlusionQueryPrecise"),
         std::make_pair(features.fragmentStoresAndAtomics, "fragmentStoresAndAtomics"),
diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp
index f31f10faa1..e48a77431b 100644
--- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp
+++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp
@@ -779,7 +779,7 @@ void RasterizerVulkan::BeginTransformFeedback() {
     UNIMPLEMENTED_IF(binding.buffer_offset != 0);
 
     const GPUVAddr gpu_addr = binding.Address();
-    const std::size_t size = binding.buffer_size;
+    const VkDeviceSize size = static_cast<VkDeviceSize>(binding.buffer_size);
     const auto [buffer, offset] = buffer_cache.UploadMemory(gpu_addr, size, 4, true);
 
     scheduler.Record([buffer = *buffer, offset = offset, size](vk::CommandBuffer cmdbuf) {
@@ -1189,24 +1189,22 @@ std::size_t RasterizerVulkan::CalculateConstBufferSize(
 }
 
 RenderPassParams RasterizerVulkan::GetRenderPassParams(Texceptions texceptions) const {
-    using namespace VideoCore::Surface;
-
-    const auto& regs = system.GPU().Maxwell3D().regs;
     RenderPassParams renderpass_params;
 
-    for (std::size_t rt = 0; rt < static_cast<std::size_t>(regs.rt_control.count); ++rt) {
-        const auto& rendertarget = regs.rt[rt];
-        if (rendertarget.Address() == 0 || rendertarget.format == Tegra::RenderTargetFormat::NONE) {
+    const std::size_t num_attachments = color_attachments.size();
+    for (std::size_t rt = 0; rt < num_attachments; ++rt) {
+        auto& attachment = color_attachments[rt];
+        if (!attachment) {
             continue;
         }
         renderpass_params.color_attachments.push_back(RenderPassParams::ColorAttachment{
-            static_cast<u32>(rt), PixelFormatFromRenderTargetFormat(rendertarget.format),
+            static_cast<u32>(rt), attachment->GetFormat(),
             texceptions[rt]});
     }
 
-    renderpass_params.has_zeta = regs.zeta_enable;
-    if (renderpass_params.has_zeta) {
-        renderpass_params.zeta_pixel_format = PixelFormatFromDepthFormat(regs.zeta.format);
+    renderpass_params.has_zeta = static_cast<bool>(zeta_attachment);
+    if (zeta_attachment) {
+        renderpass_params.zeta_pixel_format = zeta_attachment->GetFormat();
         renderpass_params.zeta_texception = texceptions[ZETA_TEXCEPTION_INDEX];
     }
 
diff --git a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
index 0f369cabd6..d32d374504 100644
--- a/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
+++ b/src/video_core/renderer_vulkan/vk_shader_decompiler.cpp
@@ -291,10 +291,10 @@ public:
         AddCapability(spv::Capability::SampledBuffer);
         AddCapability(spv::Capability::StorageImageWriteWithoutFormat);
         AddCapability(spv::Capability::DrawParameters);
-        AddCapability(spv::Capability::SubgroupBallotKHR);
-        AddCapability(spv::Capability::SubgroupVoteKHR);
-        AddExtension("SPV_KHR_shader_ballot");
-        AddExtension("SPV_KHR_subgroup_vote");
+//        AddCapability(spv::Capability::SubgroupBallotKHR);
+//        AddCapability(spv::Capability::SubgroupVoteKHR);
+//        AddExtension("SPV_KHR_shader_ballot");
+//        AddExtension("SPV_KHR_subgroup_vote");
         AddExtension("SPV_KHR_storage_buffer_storage_class");
         AddExtension("SPV_KHR_variable_pointers");
         AddExtension("SPV_KHR_shader_draw_parameters");
@@ -514,8 +514,8 @@ private:
     }
 
     void DeclareCommon() {
-        thread_id =
-            DeclareInputBuiltIn(spv::BuiltIn::SubgroupLocalInvocationId, t_in_uint, "thread_id");
+        //thread_id =
+        //    DeclareInputBuiltIn(spv::BuiltIn::SubgroupLocalInvocationId, t_in_uint, "thread_id");
     }
 
     void DeclareVertex() {
@@ -621,7 +621,8 @@ private:
 
     void DeclareRegisters() {
         for (const u32 gpr : ir.GetRegisters()) {
-            const Id id = OpVariable(t_prv_float, spv::StorageClass::Private, v_float_zero);
+            const Id id = OpVariable(t_prv_float, spv::StorageClass::Private);
+//            const Id id = OpVariable(t_prv_float, spv::StorageClass::Private, v_float_zero);
             Name(id, fmt::format("gpr_{}", gpr));
             registers.emplace(gpr, AddGlobalVariable(id));
         }
@@ -630,7 +631,8 @@ private:
     void DeclareCustomVariables() {
         const u32 num_custom_variables = ir.GetNumCustomVariables();
         for (u32 i = 0; i < num_custom_variables; ++i) {
-            const Id id = OpVariable(t_prv_float, spv::StorageClass::Private, v_float_zero);
+            const Id id = OpVariable(t_prv_float, spv::StorageClass::Private);
+            //const Id id = OpVariable(t_prv_float, spv::StorageClass::Private, v_float_zero);
             Name(id, fmt::format("custom_var_{}", i));
             custom_variables.emplace(i, AddGlobalVariable(id));
         }
@@ -638,7 +640,8 @@ private:
 
     void DeclarePredicates() {
         for (const auto pred : ir.GetPredicates()) {
-            const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false);
+            const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private);
+            //const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false);
             Name(id, fmt::format("pred_{}", static_cast<u32>(pred)));
             predicates.emplace(pred, AddGlobalVariable(id));
         }
@@ -646,7 +649,8 @@ private:
 
     void DeclareFlowVariables() {
         for (u32 i = 0; i < ir.GetASTNumVariables(); i++) {
-            const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false);
+            const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private);
+            //const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false);
             Name(id, fmt::format("flow_var_{}", static_cast<u32>(i)));
             flow_variables.emplace(i, AddGlobalVariable(id));
         }
@@ -693,7 +697,8 @@ private:
         constexpr std::array names = {"zero", "sign", "carry", "overflow"};
         for (std::size_t flag = 0; flag < INTERNAL_FLAGS_COUNT; ++flag) {
             const auto flag_code = static_cast<InternalFlag>(flag);
-            const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false);
+            const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private);
+            //const Id id = OpVariable(t_prv_bool, spv::StorageClass::Private, v_false);
             internal_flags[flag] = AddGlobalVariable(Name(id, names[flag]));
         }
     }
@@ -1294,7 +1299,7 @@ private:
         }
 
         if (const auto comment = std::get_if<CommentNode>(&*node)) {
-            Name(OpUndef(t_void), comment->GetText());
+            //Name(OpUndef(t_void), comment->GetText());
             return {};
         }
 
diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.h b/src/video_core/renderer_vulkan/vk_texture_cache.h
index bbd33a049b..533a81f33e 100644
--- a/src/video_core/renderer_vulkan/vk_texture_cache.h
+++ b/src/video_core/renderer_vulkan/vk_texture_cache.h
@@ -175,6 +175,10 @@ public:
         return {surface.GetAspectMask(), base_level, base_layer, num_layers};
     }
 
+    VideoCore::Surface::PixelFormat GetFormat() const {
+        return params.pixel_format;
+    }
+
     void Transition(VkImageLayout new_layout, VkPipelineStageFlags new_stage_mask,
                     VkAccessFlags new_access) const {
         surface.Transition(base_layer, num_layers, base_level, num_levels, new_stage_mask,
diff --git a/src/video_core/renderer_vulkan/wrapper.h b/src/video_core/renderer_vulkan/wrapper.h
index fb36578198..0531db8c71 100644
--- a/src/video_core/renderer_vulkan/wrapper.h
+++ b/src/video_core/renderer_vulkan/wrapper.h
@@ -16,6 +16,10 @@
 #define VK_NO_PROTOTYPES
 #include <vulkan/vulkan.h>
 
+#ifdef __APPLE__
+#include <MoltenVK/vk_mvk_moltenvk.h>
+#endif
+
 #include "common/common_types.h"
 
 namespace Vulkan::vk {
diff --git a/src/yuzu/CMakeLists.txt b/src/yuzu/CMakeLists.txt
index e57ea06d7a..f63a3e3611 100644
--- a/src/yuzu/CMakeLists.txt
+++ b/src/yuzu/CMakeLists.txt
@@ -212,3 +212,50 @@ if (ENABLE_VULKAN)
     target_include_directories(yuzu PRIVATE ../../externals/Vulkan-Headers/include)
     target_compile_definitions(yuzu PRIVATE HAS_VULKAN)
 endif()
+
+
+if(APPLE)
+  target_include_directories(yuzu PRIVATE ../../externals/MoltenVK/include)
+
+  include(BundleUtilities)
+  set(BUNDLE_PATH ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/yuzu.app)
+
+  # Ask for an application bundle.
+  set_target_properties(yuzu PROPERTIES
+    MACOSX_BUNDLE true
+    MACOSX_BUNDLE_INFO_PLIST ${CMAKE_CURRENT_SOURCE_DIR}/Info.plist
+  )
+
+  # Copy Qt plugins into the bundle
+  get_target_property(qtcocoa_location Qt5::QCocoaIntegrationPlugin LOCATION)
+  target_sources(yuzu PRIVATE "${qtcocoa_location}")
+  set_source_files_properties("${qtcocoa_location}" PROPERTIES MACOSX_PACKAGE_LOCATION MacOS/platforms)
+
+  get_target_property(qtmacstyle_location Qt5::QMacStylePlugin LOCATION)
+  target_sources(yuzu PRIVATE "${qtmacstyle_location}")
+  set_source_files_properties("${qtmacstyle_location}" PROPERTIES MACOSX_PACKAGE_LOCATION MacOS/styles)
+
+  # Copy MoltenVK into the bundle
+  target_sources(yuzu PRIVATE "${CMAKE_SOURCE_DIR}/externals/MoltenVK/lib/libMoltenVK.dylib")
+  set_source_files_properties("${CMAKE_SOURCE_DIR}/externals/MoltenVK/lib/libMoltenVK.dylib" PROPERTIES MACOSX_PACKAGE_LOCATION Frameworks)
+  target_sources(yuzu PRIVATE "${CMAKE_SOURCE_DIR}/externals/MoltenVK/lib/libvulkan.dylib")
+  set_source_files_properties("${CMAKE_SOURCE_DIR}/externals/MoltenVK/lib/libvulkan.dylib" PROPERTIES MACOSX_PACKAGE_LOCATION Frameworks)
+
+  # Copy validation layers as well
+  target_sources(yuzu PRIVATE "${CMAKE_SOURCE_DIR}/externals/MoltenVK/lib/libVkLayer_khronos_validation.dylib")
+  set_source_files_properties("${CMAKE_SOURCE_DIR}/externals/MoltenVK/lib/libVkLayer_khronos_validation.dylib" PROPERTIES MACOSX_PACKAGE_LOCATION Frameworks)
+  target_sources(yuzu PRIVATE "${CMAKE_SOURCE_DIR}/externals/MoltenVK/Resources/vulkan/icd.d/MoltenVK_icd.json")
+  set_source_files_properties("${CMAKE_SOURCE_DIR}/externals/MoltenVK/Resources/vulkan/icd.d/MoltenVK_icd.json" PROPERTIES MACOSX_PACKAGE_LOCATION "Resources/vulkan/icd.d")
+
+
+  # Update library references to make the bundle portable
+  include(PostprocessBundle)
+  #postprocess_bundle(yuzu)
+  # Fix rpath
+  add_custom_command(TARGET yuzu
+    POST_BUILD COMMAND
+    ${CMAKE_INSTALL_NAME_TOOL} -add_rpath "@executable_path/../Frameworks/"
+    $<TARGET_FILE:yuzu>)
+else()
+  install(TARGETS yuzu RUNTIME DESTINATION ${bindir})
+endif()
diff --git a/src/yuzu/bootmanager.cpp b/src/yuzu/bootmanager.cpp
index 6c62def603..67896f4d22 100644
--- a/src/yuzu/bootmanager.cpp
+++ b/src/yuzu/bootmanager.cpp
@@ -8,11 +8,13 @@
 #include <QHBoxLayout>
 #include <QKeyEvent>
 #include <QMessageBox>
+#ifndef __APPLE__
 #include <QOffscreenSurface>
 #include <QOpenGLContext>
 #include <QOpenGLFunctions>
 #include <QOpenGLFunctions_4_3_Core>
 #include <QOpenGLWindow>
+#endif
 #include <QPainter>
 #include <QScreen>
 #include <QStringList>
@@ -112,6 +114,7 @@ void EmuThread::run() {
 #endif
 }
 
+#ifndef __APPLE__
 class OpenGLContext : public Core::Frontend::GraphicsContext {
 public:
     explicit OpenGLContext(QOpenGLContext* shared_context)
@@ -142,67 +145,7 @@ private:
     QOpenGLContext* context;
     QOffscreenSurface* surface;
 };
-
-class ChildRenderWindow : public QWindow {
-public:
-    ChildRenderWindow(QWindow* parent, QWidget* event_handler)
-        : QWindow{parent}, event_handler{event_handler} {}
-
-    virtual ~ChildRenderWindow() = default;
-
-    virtual void Present() = 0;
-
-protected:
-    bool event(QEvent* event) override {
-        switch (event->type()) {
-        case QEvent::UpdateRequest:
-            Present();
-            return true;
-        case QEvent::MouseButtonPress:
-        case QEvent::MouseButtonRelease:
-        case QEvent::MouseButtonDblClick:
-        case QEvent::MouseMove:
-        case QEvent::KeyPress:
-        case QEvent::KeyRelease:
-        case QEvent::FocusIn:
-        case QEvent::FocusOut:
-        case QEvent::FocusAboutToChange:
-        case QEvent::Enter:
-        case QEvent::Leave:
-        case QEvent::Wheel:
-        case QEvent::TabletMove:
-        case QEvent::TabletPress:
-        case QEvent::TabletRelease:
-        case QEvent::TabletEnterProximity:
-        case QEvent::TabletLeaveProximity:
-        case QEvent::TouchBegin:
-        case QEvent::TouchUpdate:
-        case QEvent::TouchEnd:
-        case QEvent::InputMethodQuery:
-        case QEvent::TouchCancel:
-            return QCoreApplication::sendEvent(event_handler, event);
-        case QEvent::Drop:
-            GetMainWindow()->DropAction(static_cast<QDropEvent*>(event));
-            return true;
-        case QEvent::DragResponse:
-        case QEvent::DragEnter:
-        case QEvent::DragLeave:
-        case QEvent::DragMove:
-            GetMainWindow()->AcceptDropEvent(static_cast<QDropEvent*>(event));
-            return true;
-        default:
-            return QWindow::event(event);
-        }
-    }
-
-    void exposeEvent(QExposeEvent* event) override {
-        QWindow::requestUpdate();
-        QWindow::exposeEvent(event);
-    }
-
-private:
-    QWidget* event_handler{};
-};
+#endif
 
 class RenderWidget : public QWidget {
 public:
@@ -274,14 +217,17 @@ public:
         return autoFillBackground() ? QWidget::paintEngine() : nullptr;
     }
 
+    virtual void Present() {}
+
 private:
     GRenderWindow* parent;
 };
 
-class OpenGLWindow final : public ChildRenderWindow {
+#ifndef __APPLE__
+class OpenGLWidget final : public RenderWidget {
 public:
-    OpenGLWindow(QWindow* parent, QWidget* event_handler, QOpenGLContext* shared_context)
-        : ChildRenderWindow{parent, event_handler},
+    OpenGLWidget(GRenderWindow* parent, QOpenGLContext* shared_context)
+        : RenderWidget{parent},
           context(new QOpenGLContext(shared_context->parent())) {
 
         // disable vsync for any shared contexts
@@ -300,7 +246,7 @@ public:
         // WA_DontShowOnScreen, WA_DeleteOnClose
     }
 
-    ~OpenGLWindow() override {
+    ~OpenGLWidget() override {
         context->doneCurrent();
     }
 
@@ -320,25 +266,6 @@ public:
 private:
     QOpenGLContext* context{};
 };
-
-#ifdef HAS_VULKAN
-class VulkanWindow final : public ChildRenderWindow {
-public:
-    VulkanWindow(QWindow* parent, QWidget* event_handler, QVulkanInstance* instance)
-        : ChildRenderWindow{parent, event_handler} {
-        setSurfaceType(QSurface::SurfaceType::VulkanSurface);
-        setVulkanInstance(instance);
-    }
-
-    ~VulkanWindow() override = default;
-
-    void Present() override {
-        // TODO(bunnei): ImplementMe
-    }
-
-private:
-    QWidget* event_handler = nullptr;
-};
 #endif
 
 static Core::Frontend::WindowSystemType GetWindowSystemType() {
@@ -346,6 +273,8 @@ static Core::Frontend::WindowSystemType GetWindowSystemType() {
     QString platform_name = QGuiApplication::platformName();
     if (platform_name == QStringLiteral("windows"))
         return Core::Frontend::WindowSystemType::Windows;
+    else if (platform_name == QStringLiteral("cocoa"))
+        return Core::Frontend::WindowSystemType::MacOS;
     else if (platform_name == QStringLiteral("xcb"))
         return Core::Frontend::WindowSystemType::X11;
     else if (platform_name == QStringLiteral("wayland"))
@@ -396,15 +325,19 @@ GRenderWindow::~GRenderWindow() {
 }
 
 void GRenderWindow::MakeCurrent() {
+#ifndef __APPLE__
     if (core_context) {
         core_context->MakeCurrent();
     }
+#endif
 }
 
 void GRenderWindow::DoneCurrent() {
+#ifndef __APPLE__
     if (core_context) {
         core_context->DoneCurrent();
     }
+#endif
 }
 
 void GRenderWindow::PollEvents() {
@@ -567,14 +500,18 @@ void GRenderWindow::resizeEvent(QResizeEvent* event) {
 }
 
 std::unique_ptr<Core::Frontend::GraphicsContext> GRenderWindow::CreateSharedContext() const {
+#ifndef __APPLE__
     if (Settings::values.renderer_backend == Settings::RendererBackend::OpenGL) {
         return std::make_unique<OpenGLContext>(QOpenGLContext::globalShareContext());
     }
+#endif
     return {};
 }
 
 bool GRenderWindow::ReloadRenderTarget() {
+#ifndef __APPLE__
     core_context.reset();
+#endif
     delete child;
     delete layout();
     first_frame = false;
@@ -656,6 +593,7 @@ void GRenderWindow::CaptureScreenshot(u32 res_scale, const QString& screenshot_p
 }
 
 bool GRenderWindow::InitializeOpenGL() {
+#ifndef __APPLE__
     // TODO: One of these flags might be interesting: WA_OpaquePaintEvent, WA_NoBackground,
     // WA_DontShowOnScreen, WA_DeleteOnClose
     QSurfaceFormat fmt;
@@ -667,20 +605,18 @@ bool GRenderWindow::InitializeOpenGL() {
     fmt.setSwapInterval(0);
     QSurfaceFormat::setDefaultFormat(fmt);
 
-    GMainWindow* parent = GetMainWindow();
-    QWindow* parent_win_handle = parent ? parent->windowHandle() : nullptr;
-    child_window = new OpenGLWindow(parent_win_handle, this, QOpenGLContext::globalShareContext());
-    child_window->create();
-    child_widget = createWindowContainer(child_window, this);
-    child_widget->resize(Layout::ScreenUndocked::Width, Layout::ScreenUndocked::Height);
     layout()->addWidget(child_widget);
 
     core_context = CreateSharedContext();
     return true;
+#else
+    return false;
+#endif
 }
 
 bool GRenderWindow::InitializeVulkan() {
 #ifdef HAS_VULKAN
+    child->windowHandle()->setSurfaceType(QSurface::SurfaceType::VulkanSurface);
     layout()->addWidget(child);
     return true;
 #else
diff --git a/src/yuzu/bootmanager.h b/src/yuzu/bootmanager.h
index ac974c9c19..3e556f74fa 100644
--- a/src/yuzu/bootmanager.h
+++ b/src/yuzu/bootmanager.h
@@ -19,7 +19,6 @@
 
 class GRenderWindow;
 class QKeyEvent;
-class QOpenGLContext;
 class QScreen;
 class QTouchEvent;
 class QStringList;
@@ -188,7 +187,9 @@ private:
 
     EmuThread* emu_thread;
 
+#ifndef __APPLE__
     std::unique_ptr<GraphicsContext> core_context;
+#endif
 
     /// Temporary storage of the screenshot taken
     QImage screenshot_image;