Fermi2D: Rework blit engine and add a software blitter.

src/video_core/CMakeLists.txt

@@ -28,6 +28,10 @@ add_library(video_core STATIC
     dirty_flags.h
     dma_pusher.cpp
     dma_pusher.h
+    engines/sw_blitter/blitter.cpp
+    engines/sw_blitter/blitter.h
+    engines/sw_blitter/converter.cpp
+    engines/sw_blitter/converter.h
     engines/const_buffer_info.h
     engines/engine_interface.h
     engines/engine_upload.cpp

src/video_core/control/channel_state.cpp

@@ -20,7 +20,7 @@ void ChannelState::Init(Core::System& system, GPU& gpu) {
     ASSERT(memory_manager);
     dma_pusher = std::make_unique<Tegra::DmaPusher>(system, gpu, *memory_manager, *this);
     maxwell_3d = std::make_unique<Engines::Maxwell3D>(system, *memory_manager);
-    fermi_2d = std::make_unique<Engines::Fermi2D>();
+    fermi_2d = std::make_unique<Engines::Fermi2D>(*memory_manager);
     kepler_compute = std::make_unique<Engines::KeplerCompute>(system, *memory_manager);
     maxwell_dma = std::make_unique<Engines::MaxwellDMA>(system, *memory_manager);
     kepler_memory = std::make_unique<Engines::KeplerMemory>(system, *memory_manager);

src/video_core/engines/fermi_2d.cpp

@@ -3,17 +3,25 @@
 
 #include "common/assert.h"
 #include "common/logging/log.h"
+#include "common/microprofile.h"
 #include "video_core/engines/fermi_2d.h"
-#include "video_core/memory_manager.h"
+#include "video_core/engines/sw_blitter/blitter.h"
 #include "video_core/rasterizer_interface.h"
 #include "video_core/surface.h"
+#include "video_core/textures/decoders.h"
+
+MICROPROFILE_DECLARE(GPU_BlitEngine);
+MICROPROFILE_DEFINE(GPU_BlitEngine, "GPU", "Blit Engine", MP_RGB(224, 224, 128));
 
 using VideoCore::Surface::BytesPerBlock;
 using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
 
 namespace Tegra::Engines {
 
-Fermi2D::Fermi2D() {
+using namespace Texture;
+
+Fermi2D::Fermi2D(MemoryManager& memory_manager_) {
+    sw_blitter = std::make_unique<Blitter::SoftwareBlitEngine>(memory_manager_);
     // Nvidia's OpenGL driver seems to assume these values
     regs.src.depth = 1;
     regs.dst.depth = 1;
@@ -42,6 +50,7 @@ void Fermi2D::CallMultiMethod(u32 method, const u32* base_start, u32 amount, u32
 }
 
 void Fermi2D::Blit() {
+    MICROPROFILE_SCOPE(GPU_BlitEngine);
     LOG_DEBUG(HW_GPU, "called. source address=0x{:x}, destination address=0x{:x}",
               regs.src.Address(), regs.dst.Address());
 
@@ -52,9 +61,12 @@ void Fermi2D::Blit() {
     UNIMPLEMENTED_IF_MSG(regs.clip_enable != 0, "Clipped blit enabled");
 
     const auto& args = regs.pixels_from_memory;
+    constexpr s64 null_derivate = 1ULL << 32;
     Config config{
         .operation = regs.operation,
         .filter = args.sample_mode.filter,
+        .must_accelerate = args.du_dx != null_derivate || args.dv_dy != null_derivate ||
+                           args.sample_mode.filter == Filter::Bilinear,
         .dst_x0 = args.dst_x0,
         .dst_y0 = args.dst_y0,
         .dst_x1 = args.dst_x0 + args.dst_width,
@@ -78,8 +90,9 @@ void Fermi2D::Blit() {
         config.src_x1 -= config.src_x0;
         config.src_x0 = 0;
     }
+
     if (!rasterizer->AccelerateSurfaceCopy(src, regs.dst, config)) {
-        UNIMPLEMENTED();
+        sw_blitter->Blit(src, regs.dst, config);
     }
 }
 

src/video_core/engines/fermi_2d.h

@@ -5,6 +5,7 @@
 
 #include <array>
 #include <cstddef>
+#include <memory>
 #include "common/bit_field.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
@@ -21,6 +22,10 @@ class RasterizerInterface;
 
 namespace Tegra::Engines {
 
+namespace Blitter {
+class SoftwareBlitEngine;
+}
+
 /**
  * This Engine is known as G80_2D. Documentation can be found in:
  * https://github.com/envytools/envytools/blob/master/rnndb/graph/g80_2d.xml
@@ -32,7 +37,7 @@ namespace Tegra::Engines {
 
 class Fermi2D final : public EngineInterface {
 public:
-    explicit Fermi2D();
+    explicit Fermi2D(MemoryManager& memory_manager_);
     ~Fermi2D() override;
 
     /// Binds a rasterizer to this engine.
@@ -286,6 +291,7 @@ public:
     struct Config {
         Operation operation;
         Filter filter;
+        bool must_accelerate;
         s32 dst_x0;
         s32 dst_y0;
         s32 dst_x1;
@@ -298,6 +304,7 @@ public:
 
 private:
     VideoCore::RasterizerInterface* rasterizer = nullptr;
+    std::unique_ptr<Blitter::SoftwareBlitEngine> sw_blitter;
 
     /// Performs the copy from the source surface to the destination surface as configured in the
     /// registers.

src/video_core/engines/sw_blitter/blitter.cpp

@@ -0,0 +1,213 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#include <vector>
+
+#include "video_core/engines/sw_blitter/blitter.h"
+#include "video_core/engines/sw_blitter/converter.h"
+#include "video_core/memory_manager.h"
+#include "video_core/surface.h"
+#include "video_core/textures/decoders.h"
+
+namespace Tegra {
+class MemoryManager;
+}
+
+using VideoCore::Surface::BytesPerBlock;
+using VideoCore::Surface::PixelFormatFromRenderTargetFormat;
+
+namespace Tegra::Engines::Blitter {
+
+using namespace Texture;
+
+namespace {
+
+void NeighrestNeighbor(std::span<u8> input, std::span<u8> output, u32 src_width, u32 src_height,
+                       u32 dst_width, u32 dst_height, size_t bpp) {
+    const size_t dx_du = std::llround((static_cast<f64>(src_width) / dst_width) * (1ULL << 32));
+    const size_t dy_dv = std::llround((static_cast<f64>(src_height) / dst_height) * (1ULL << 32));
+    size_t src_y = 0;
+    for (u32 y = 0; y < dst_height; y++) {
+        size_t src_x = 0;
+        for (u32 x = 0; x < dst_width; x++) {
+            const size_t read_from = ((src_y * src_width + src_x) >> 32) * bpp;
+            const size_t write_to = (y * dst_width + x) * bpp;
+
+            std::memcpy(&output[write_to], &input[read_from], bpp);
+            src_x += dx_du;
+        }
+        src_y += dy_dv;
+    }
+}
+
+void NeighrestNeighborFast(std::span<f32> input, std::span<f32> output, u32 src_width,
+                           u32 src_height, u32 dst_width, u32 dst_height) {
+    const size_t dx_du = std::llround((static_cast<f64>(src_width) / dst_width) * (1ULL << 32));
+    const size_t dy_dv = std::llround((static_cast<f64>(src_height) / dst_height) * (1ULL << 32));
+    size_t src_y = 0;
+    for (u32 y = 0; y < dst_height; y++) {
+        size_t src_x = 0;
+        for (u32 x = 0; x < dst_width; x++) {
+            const size_t read_from = ((src_y * src_width + src_x) >> 32) * 4;
+            const size_t write_to = (y * dst_width + x) * 4;
+
+            std::memcpy(&output[write_to], &input[read_from], sizeof(f32) * 4);
+            src_x += dx_du;
+        }
+        src_y += dy_dv;
+    }
+}
+
+/*
+void Bilinear(std::span<f32> input, std::span<f32> output, size_t src_width,
+                       size_t src_height, size_t dst_width, size_t dst_height) {
+    const auto inv_lerp = [](u32 coord, u32 end) { return
+static_cast<f32>(std::min(std::max(static_cast<s32>(coord), 0), end - 1)) / (end); };
+
+
+    for (u32 y = 0; y < dst_height; y++) {
+        const f32 ty_0 = inv_lerp(y, dst_extent_y);
+        const f32 ty_1 = inv_lerp(y + 1, dst_extent_y);
+        for (u32 x = 0; x < dst_width; x++) {
+            const f32 tx_0 = inv_lerp(x, dst_extent_x);
+            const f32 tx_1 = inv_lerp(x + 1, dst_extent_x);
+            const std::array<f32, 4> get_pixel = [&](f32 tx, f32 ty, u32 width, u32 height) {
+                std::array<f32, 4> result{};
+
+                return (std::llround(width * tx) + std::llround(height * ty) * width) * 4;
+            };
+            std::array<f32, 4> result{};
+
+            const size_t read_from = get_pixel(src_width, src_height);
+            const size_t write_to = get_pixel(tx_0, ty_0, dst_width, dst_height);
+
+            std::memcpy(&output[write_to], &input[read_from], bpp);
+        }
+    }
+}
+*/
+
+} // namespace
+
+struct SoftwareBlitEngine::BlitEngineImpl {
+    std::vector<u8> tmp_buffer;
+    std::vector<u8> src_buffer;
+    std::vector<u8> dst_buffer;
+    std::vector<f32> intermediate_src;
+    std::vector<f32> intermediate_dst;
+    ConverterFactory converter_factory;
+};
+
+SoftwareBlitEngine::SoftwareBlitEngine(MemoryManager& memory_manager_)
+    : memory_manager{memory_manager_} {
+    impl = std::make_unique<BlitEngineImpl>();
+}
+
+SoftwareBlitEngine::~SoftwareBlitEngine() = default;
+
+bool SoftwareBlitEngine::Blit(Fermi2D::Surface& src, Fermi2D::Surface& dst,
+                              Fermi2D::Config& config) {
+    UNIMPLEMENTED_IF(config.filter == Fermi2D::Filter::Bilinear);
+
+    const auto get_surface_size = [](Fermi2D::Surface& surface, u32 bytes_per_pixel) {
+        if (surface.linear == Fermi2D::MemoryLayout::BlockLinear) {
+            return CalculateSize(true, bytes_per_pixel, surface.width, surface.height,
+                                 surface.depth, surface.block_height, surface.block_depth);
+        }
+        return static_cast<size_t>(surface.pitch * surface.height);
+    };
+    const auto process_pitch_linear = [](bool unpack, std::span<u8> input, std::span<u8> output,
+                                         u32 extent_x, u32 extent_y, u32 pitch, u32 x0, u32 y0,
+                                         size_t bpp) {
+        const size_t base_offset = x0 * bpp;
+        const size_t copy_size = extent_x * bpp;
+        for (u32 y = y0; y < extent_y; y++) {
+            const size_t first_offset = y * pitch + base_offset;
+            const size_t second_offset = y * extent_x * bpp;
+            u8* write_to = unpack ? &output[first_offset] : &output[second_offset];
+            const u8* read_from = unpack ? &input[second_offset] : &input[first_offset];
+            std::memcpy(write_to, read_from, copy_size);
+        }
+    };
+
+    const u32 src_extent_x = config.src_x1 - config.src_x0;
+    const u32 src_extent_y = config.src_y1 - config.src_y0;
+
+    const u32 dst_extent_x = config.dst_x1 - config.dst_x0;
+    const u32 dst_extent_y = config.dst_y1 - config.dst_y0;
+    const auto src_bytes_per_pixel = BytesPerBlock(PixelFormatFromRenderTargetFormat(src.format));
+    const auto dst_bytes_per_pixel = BytesPerBlock(PixelFormatFromRenderTargetFormat(dst.format));
+    const size_t src_size = get_surface_size(src, src_bytes_per_pixel);
+    impl->tmp_buffer.resize(src_size);
+    memory_manager.ReadBlock(src.Address(), impl->tmp_buffer.data(), src_size);
+
+    const size_t src_copy_size = src_extent_x * src_extent_y * src_bytes_per_pixel;
+
+    const size_t dst_copy_size = dst_extent_x * dst_extent_y * dst_bytes_per_pixel;
+
+    impl->src_buffer.resize(src_copy_size);
+
+    const bool no_passthrough =
+        src.format != dst.format || src_extent_x != dst_extent_x || src_extent_y != dst_extent_y;
+
+    const auto convertion_phase_same_format = [&]() {
+        NeighrestNeighbor(impl->src_buffer, impl->dst_buffer, src_extent_x, src_extent_y,
+                          dst_extent_x, dst_extent_y, dst_bytes_per_pixel);
+    };
+
+    const auto convertion_phase_ir = [&]() {
+        auto* input_converter = impl->converter_factory.GetFormatConverter(src.format);
+        impl->intermediate_src.resize((src_copy_size / src_bytes_per_pixel) * 4);
+        impl->intermediate_dst.resize((dst_copy_size / dst_bytes_per_pixel) * 4);
+        input_converter->ConvertTo(impl->src_buffer, impl->intermediate_src);
+
+        NeighrestNeighborFast(impl->intermediate_src, impl->intermediate_dst, src_extent_x,
+                              src_extent_y, dst_extent_x, dst_extent_y);
+
+        auto* output_converter = impl->converter_factory.GetFormatConverter(dst.format);
+        output_converter->ConvertFrom(impl->intermediate_dst, impl->dst_buffer);
+    };
+
+    // Do actuall Blit
+
+    impl->dst_buffer.resize(dst_copy_size);
+    if (src.linear == Fermi2D::MemoryLayout::BlockLinear) {
+        UnswizzleSubrect(impl->src_buffer, impl->tmp_buffer, src_bytes_per_pixel, src.width,
+                         src.height, src.depth, config.src_x0, config.src_y0, src_extent_x,
+                         src_extent_y, src.block_height, src.block_depth,
+                         src_extent_x * src_bytes_per_pixel);
+    } else {
+        process_pitch_linear(false, impl->tmp_buffer, impl->src_buffer, src_extent_x, src_extent_y,
+                             src.pitch, config.src_x0, config.src_y0, src_bytes_per_pixel);
+    }
+
+    // Conversion Phase
+    if (no_passthrough) {
+        if (src.format != dst.format) {
+            convertion_phase_ir();
+        } else {
+            convertion_phase_same_format();
+        }
+    } else {
+        impl->dst_buffer.swap(impl->src_buffer);
+    }
+
+    const size_t dst_size = get_surface_size(dst, dst_bytes_per_pixel);
+    impl->tmp_buffer.resize(dst_size);
+    memory_manager.ReadBlock(dst.Address(), impl->tmp_buffer.data(), dst_size);
+
+    if (dst.linear == Fermi2D::MemoryLayout::BlockLinear) {
+        SwizzleSubrect(impl->tmp_buffer, impl->dst_buffer, dst_bytes_per_pixel, dst.width,
+                       dst.height, dst.depth, config.dst_x0, config.dst_y0, dst_extent_x,
+                       dst_extent_y, dst.block_height, dst.block_depth,
+                       dst_extent_x * dst_bytes_per_pixel);
+    } else {
+        process_pitch_linear(true, impl->dst_buffer, impl->tmp_buffer, dst_extent_x, dst_extent_y,
+                             dst.pitch, config.dst_x0, config.dst_y0,
+                             static_cast<size_t>(dst_bytes_per_pixel));
+    }
+    memory_manager.WriteBlock(dst.Address(), impl->tmp_buffer.data(), dst_size);
+    return true;
+}
+
+} // namespace Tegra::Engines::Blitter

src/video_core/engines/sw_blitter/blitter.h

@@ -0,0 +1,27 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#pragma once
+
+#include "video_core/engines/fermi_2d.h"
+
+namespace Tegra {
+class MemoryManager;
+}
+
+namespace Tegra::Engines::Blitter {
+
+class SoftwareBlitEngine {
+public:
+    SoftwareBlitEngine(MemoryManager& memory_manager_);
+    ~SoftwareBlitEngine();
+
+    bool Blit(Fermi2D::Surface& src, Fermi2D::Surface& dst, Fermi2D::Config& copy_config);
+
+private:
+    MemoryManager& memory_manager;
+    struct BlitEngineImpl;
+    std::unique_ptr<BlitEngineImpl> impl;
+};
+
+} // namespace Tegra::Engines::Blitter

src/video_core/engines/sw_blitter/converter.cpp

@@ -0,0 +1,1097 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#include <array>
+#include <bit>
+#include <cmath>
+#include <span>
+#include <unordered_map>
+
+#include "common/assert.h"
+#include "video_core/engines/sw_blitter/converter.h"
+#include "video_core/surface.h"
+#include "video_core/textures/decoders.h"
+
+#ifdef _MSC_VER
+#define FORCE_INLINE __forceinline
+#else
+#define FORCE_INLINE inline __attribute__((always_inline))
+#endif
+
+namespace Tegra::Engines::Blitter {
+
+enum class Swizzle : size_t {
+    R = 0,
+    G = 1,
+    B = 2,
+    A = 3,
+    None,
+};
+
+enum class ComponentType : u32 {
+    SNORM = 1,
+    UNORM = 2,
+    SINT = 3,
+    UINT = 4,
+    SNORM_FORCE_FP16 = 5,
+    UNORM_FORCE_FP16 = 6,
+    FLOAT = 7,
+    SRGB = 8,
+};
+
+namespace {
+
+constexpr std::array<f32, 256> SRGB_TO_RGB_LUT = {
+    0.000000e+00f, 3.035270e-04f, 6.070540e-04f, 9.105810e-04f, 1.214108e-03f, 1.517635e-03f,
+    1.821162e-03f, 2.124689e-03f, 2.428216e-03f, 2.731743e-03f, 3.035270e-03f, 3.346536e-03f,
+    3.676507e-03f, 4.024717e-03f, 4.391442e-03f, 4.776953e-03f, 5.181517e-03f, 5.605392e-03f,
+    6.048833e-03f, 6.512091e-03f, 6.995410e-03f, 7.499032e-03f, 8.023193e-03f, 8.568126e-03f,
+    9.134059e-03f, 9.721218e-03f, 1.032982e-02f, 1.096009e-02f, 1.161224e-02f, 1.228649e-02f,
+    1.298303e-02f, 1.370208e-02f, 1.444384e-02f, 1.520851e-02f, 1.599629e-02f, 1.680738e-02f,
+    1.764195e-02f, 1.850022e-02f, 1.938236e-02f, 2.028856e-02f, 2.121901e-02f, 2.217389e-02f,
+    2.315337e-02f, 2.415763e-02f, 2.518686e-02f, 2.624122e-02f, 2.732089e-02f, 2.842604e-02f,
+    2.955684e-02f, 3.071344e-02f, 3.189603e-02f, 3.310477e-02f, 3.433981e-02f, 3.560131e-02f,
+    3.688945e-02f, 3.820437e-02f, 3.954624e-02f, 4.091520e-02f, 4.231141e-02f, 4.373503e-02f,
+    4.518620e-02f, 4.666509e-02f, 4.817183e-02f, 4.970657e-02f, 5.126946e-02f, 5.286065e-02f,
+    5.448028e-02f, 5.612849e-02f, 5.780543e-02f, 5.951124e-02f, 6.124605e-02f, 6.301001e-02f,
+    6.480327e-02f, 6.662594e-02f, 6.847817e-02f, 7.036009e-02f, 7.227185e-02f, 7.421357e-02f,
+    7.618538e-02f, 7.818742e-02f, 8.021982e-02f, 8.228271e-02f, 8.437621e-02f, 8.650046e-02f,
+    8.865558e-02f, 9.084171e-02f, 9.305897e-02f, 9.530747e-02f, 9.758735e-02f, 9.989873e-02f,
+    1.022417e-01f, 1.046165e-01f, 1.070231e-01f, 1.094617e-01f, 1.119324e-01f, 1.144354e-01f,
+    1.169707e-01f, 1.195384e-01f, 1.221388e-01f, 1.247718e-01f, 1.274377e-01f, 1.301365e-01f,
+    1.328683e-01f, 1.356333e-01f, 1.384316e-01f, 1.412633e-01f, 1.441285e-01f, 1.470273e-01f,
+    1.499598e-01f, 1.529261e-01f, 1.559265e-01f, 1.589608e-01f, 1.620294e-01f, 1.651322e-01f,
+    1.682694e-01f, 1.714411e-01f, 1.746474e-01f, 1.778884e-01f, 1.811642e-01f, 1.844750e-01f,
+    1.878208e-01f, 1.912017e-01f, 1.946178e-01f, 1.980693e-01f, 2.015563e-01f, 2.050787e-01f,
+    2.086369e-01f, 2.122308e-01f, 2.158605e-01f, 2.195262e-01f, 2.232280e-01f, 2.269659e-01f,
+    2.307401e-01f, 2.345506e-01f, 2.383976e-01f, 2.422811e-01f, 2.462013e-01f, 2.501583e-01f,
+    2.541521e-01f, 2.581829e-01f, 2.622507e-01f, 2.663556e-01f, 2.704978e-01f, 2.746773e-01f,
+    2.788943e-01f, 2.831487e-01f, 2.874408e-01f, 2.917706e-01f, 2.961383e-01f, 3.005438e-01f,
+    3.049873e-01f, 3.094689e-01f, 3.139887e-01f, 3.185468e-01f, 3.231432e-01f, 3.277781e-01f,
+    3.324515e-01f, 3.371636e-01f, 3.419144e-01f, 3.467041e-01f, 3.515326e-01f, 3.564001e-01f,
+    3.613068e-01f, 3.662526e-01f, 3.712377e-01f, 3.762621e-01f, 3.813260e-01f, 3.864294e-01f,
+    3.915725e-01f, 3.967552e-01f, 4.019778e-01f, 4.072402e-01f, 4.125426e-01f, 4.178851e-01f,
+    4.232677e-01f, 4.286905e-01f, 4.341536e-01f, 4.396572e-01f, 4.452012e-01f, 4.507858e-01f,
+    4.564110e-01f, 4.620770e-01f, 4.677838e-01f, 4.735315e-01f, 4.793202e-01f, 4.851499e-01f,
+    4.910209e-01f, 4.969330e-01f, 5.028865e-01f, 5.088813e-01f, 5.149177e-01f, 5.209956e-01f,
+    5.271151e-01f, 5.332764e-01f, 5.394795e-01f, 5.457245e-01f, 5.520114e-01f, 5.583404e-01f,
+    5.647115e-01f, 5.711249e-01f, 5.775805e-01f, 5.840784e-01f, 5.906188e-01f, 5.972018e-01f,
+    6.038274e-01f, 6.104956e-01f, 6.172066e-01f, 6.239604e-01f, 6.307572e-01f, 6.375968e-01f,
+    6.444797e-01f, 6.514056e-01f, 6.583748e-01f, 6.653873e-01f, 6.724432e-01f, 6.795425e-01f,
+    6.866853e-01f, 6.938717e-01f, 7.011019e-01f, 7.083758e-01f, 7.156935e-01f, 7.230551e-01f,
+    7.304608e-01f, 7.379104e-01f, 7.454042e-01f, 7.529422e-01f, 7.605245e-01f, 7.681512e-01f,
+    7.758222e-01f, 7.835378e-01f, 7.912979e-01f, 7.991027e-01f, 8.069522e-01f, 8.148466e-01f,
+    8.227857e-01f, 8.307699e-01f, 8.387990e-01f, 8.468732e-01f, 8.549926e-01f, 8.631572e-01f,
+    8.713671e-01f, 8.796224e-01f, 8.879231e-01f, 8.962694e-01f, 9.046612e-01f, 9.130986e-01f,
+    9.215819e-01f, 9.301109e-01f, 9.386857e-01f, 9.473065e-01f, 9.559733e-01f, 9.646863e-01f,
+    9.734453e-01f, 9.822506e-01f, 9.911021e-01f, 1.000000e+00f};
+
+constexpr std::array<f32, 256> RGB_TO_SRGB_LUT = {
+    0.000000e+00f, 4.984009e-02f, 8.494473e-02f, 1.107021e-01f, 1.318038e-01f, 1.500052e-01f,
+    1.661857e-01f, 1.808585e-01f, 1.943532e-01f, 2.068957e-01f, 2.186491e-01f, 2.297351e-01f,
+    2.402475e-01f, 2.502604e-01f, 2.598334e-01f, 2.690152e-01f, 2.778465e-01f, 2.863614e-01f,
+    2.945889e-01f, 3.025538e-01f, 3.102778e-01f, 3.177796e-01f, 3.250757e-01f, 3.321809e-01f,
+    3.391081e-01f, 3.458689e-01f, 3.524737e-01f, 3.589320e-01f, 3.652521e-01f, 3.714419e-01f,
+    3.775084e-01f, 3.834581e-01f, 3.892968e-01f, 3.950301e-01f, 4.006628e-01f, 4.061998e-01f,
+    4.116451e-01f, 4.170030e-01f, 4.222770e-01f, 4.274707e-01f, 4.325873e-01f, 4.376298e-01f,
+    4.426010e-01f, 4.475037e-01f, 4.523403e-01f, 4.571131e-01f, 4.618246e-01f, 4.664766e-01f,
+    4.710712e-01f, 4.756104e-01f, 4.800958e-01f, 4.845292e-01f, 4.889122e-01f, 4.932462e-01f,
+    4.975329e-01f, 5.017734e-01f, 5.059693e-01f, 5.101216e-01f, 5.142317e-01f, 5.183006e-01f,
+    5.223295e-01f, 5.263194e-01f, 5.302714e-01f, 5.341862e-01f, 5.380651e-01f, 5.419087e-01f,
+    5.457181e-01f, 5.494938e-01f, 5.532369e-01f, 5.569480e-01f, 5.606278e-01f, 5.642771e-01f,
+    5.678965e-01f, 5.714868e-01f, 5.750484e-01f, 5.785821e-01f, 5.820884e-01f, 5.855680e-01f,
+    5.890211e-01f, 5.924487e-01f, 5.958509e-01f, 5.992285e-01f, 6.025819e-01f, 6.059114e-01f,
+    6.092176e-01f, 6.125010e-01f, 6.157619e-01f, 6.190008e-01f, 6.222180e-01f, 6.254140e-01f,
+    6.285890e-01f, 6.317436e-01f, 6.348780e-01f, 6.379926e-01f, 6.410878e-01f, 6.441637e-01f,
+    6.472208e-01f, 6.502595e-01f, 6.532799e-01f, 6.562824e-01f, 6.592672e-01f, 6.622347e-01f,
+    6.651851e-01f, 6.681187e-01f, 6.710356e-01f, 6.739363e-01f, 6.768209e-01f, 6.796897e-01f,
+    6.825429e-01f, 6.853807e-01f, 6.882034e-01f, 6.910111e-01f, 6.938041e-01f, 6.965826e-01f,
+    6.993468e-01f, 7.020969e-01f, 7.048331e-01f, 7.075556e-01f, 7.102645e-01f, 7.129600e-01f,
+    7.156424e-01f, 7.183118e-01f, 7.209683e-01f, 7.236121e-01f, 7.262435e-01f, 7.288625e-01f,
+    7.314693e-01f, 7.340640e-01f, 7.366470e-01f, 7.392181e-01f, 7.417776e-01f, 7.443256e-01f,
+    7.468624e-01f, 7.493880e-01f, 7.519025e-01f, 7.544061e-01f, 7.568989e-01f, 7.593810e-01f,
+    7.618526e-01f, 7.643137e-01f, 7.667645e-01f, 7.692052e-01f, 7.716358e-01f, 7.740564e-01f,
+    7.764671e-01f, 7.788681e-01f, 7.812595e-01f, 7.836413e-01f, 7.860138e-01f, 7.883768e-01f,
+    7.907307e-01f, 7.930754e-01f, 7.954110e-01f, 7.977377e-01f, 8.000556e-01f, 8.023647e-01f,
+    8.046651e-01f, 8.069569e-01f, 8.092403e-01f, 8.115152e-01f, 8.137818e-01f, 8.160402e-01f,
+    8.182903e-01f, 8.205324e-01f, 8.227665e-01f, 8.249926e-01f, 8.272109e-01f, 8.294214e-01f,
+    8.316242e-01f, 8.338194e-01f, 8.360070e-01f, 8.381871e-01f, 8.403597e-01f, 8.425251e-01f,
+    8.446831e-01f, 8.468339e-01f, 8.489776e-01f, 8.511142e-01f, 8.532437e-01f, 8.553662e-01f,
+    8.574819e-01f, 8.595907e-01f, 8.616927e-01f, 8.637881e-01f, 8.658767e-01f, 8.679587e-01f,
+    8.700342e-01f, 8.721032e-01f, 8.741657e-01f, 8.762218e-01f, 8.782716e-01f, 8.803151e-01f,
+    8.823524e-01f, 8.843835e-01f, 8.864085e-01f, 8.884274e-01f, 8.904402e-01f, 8.924471e-01f,
+    8.944480e-01f, 8.964431e-01f, 8.984324e-01f, 9.004158e-01f, 9.023935e-01f, 9.043654e-01f,
+    9.063318e-01f, 9.082925e-01f, 9.102476e-01f, 9.121972e-01f, 9.141413e-01f, 9.160800e-01f,
+    9.180133e-01f, 9.199412e-01f, 9.218637e-01f, 9.237810e-01f, 9.256931e-01f, 9.276000e-01f,
+    9.295017e-01f, 9.313982e-01f, 9.332896e-01f, 9.351761e-01f, 9.370575e-01f, 9.389339e-01f,
+    9.408054e-01f, 9.426719e-01f, 9.445336e-01f, 9.463905e-01f, 9.482424e-01f, 9.500897e-01f,
+    9.519322e-01f, 9.537700e-01f, 9.556032e-01f, 9.574316e-01f, 9.592555e-01f, 9.610748e-01f,
+    9.628896e-01f, 9.646998e-01f, 9.665055e-01f, 9.683068e-01f, 9.701037e-01f, 9.718961e-01f,
+    9.736842e-01f, 9.754679e-01f, 9.772474e-01f, 9.790225e-01f, 9.807934e-01f, 9.825601e-01f,
+    9.843225e-01f, 9.860808e-01f, 9.878350e-01f, 9.895850e-01f, 9.913309e-01f, 9.930727e-01f,
+    9.948106e-01f, 9.965444e-01f, 9.982741e-01f, 1.000000e+00f};
+
+} // namespace
+
+struct R32B32G32A32_FLOATTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::G, Swizzle::B, Swizzle::R};
+};
+
+struct R32G32B32A32_SINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT, ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R32G32B32A32_UINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT, ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32, 32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R16G16B16A16_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16, 16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R16G16B16A16_SNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SNORM, ComponentType::SNORM, ComponentType::SNORM, ComponentType::SNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16, 16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R16G16B16A16_SINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT, ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16, 16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R16G16B16A16_UINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT, ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16, 16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R16G16B16A16_FLOATTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16, 16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::A, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct R32G32_FLOATTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R32G32_SINTTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R32G32_UINTTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32, 32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R16G16B16X16_FLOATTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16, 16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::None, Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct A8R8G8B8_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::B, Swizzle::G, Swizzle::R, Swizzle::A};
+};
+
+struct A8R8G8B8_SRGBTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SRGB, ComponentType::SRGB, ComponentType::SRGB, ComponentType::SRGB};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::B, Swizzle::G, Swizzle::R, Swizzle::A};
+};
+
+struct A2B10G10R10_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {10, 10, 10, 2};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct A2B10G10R10_UINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT, ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {10, 10, 10, 2};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct A8B8G8R8_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct A8B8G8R8_SRGBTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SRGB, ComponentType::SRGB, ComponentType::SRGB, ComponentType::SRGB};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct A8B8G8R8_SNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SNORM, ComponentType::SNORM, ComponentType::SNORM, ComponentType::SNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct A8B8G8R8_SINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT, ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct A8B8G8R8_UINTTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT, ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8, 8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B, Swizzle::A};
+};
+
+struct R16G16_UNORMTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R16G16_SNORMTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SNORM, ComponentType::SNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R16G16_SINTTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R16G16_UINTTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R16G16_FLOATTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16, 16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct B10G11R11_FLOATTraits {
+    static constexpr size_t num_components = 3;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT, ComponentType::FLOAT, ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {11, 11, 10};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::R, Swizzle::G, Swizzle::B};
+};
+
+struct R32_SINTTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R32_UINTTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R32_FLOATTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {32};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R5G6B5_UNORMTraits {
+    static constexpr size_t num_components = 3;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {5, 6, 5};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::B, Swizzle::G, Swizzle::R};
+};
+
+struct A1R5G5B5_UNORMTraits {
+    static constexpr size_t num_components = 4;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {5, 5, 5, 1};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {
+        Swizzle::B, Swizzle::G, Swizzle::R, Swizzle::A};
+};
+
+struct R8G8_UNORMTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM, ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R8G8_SNORMTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SNORM, ComponentType::SNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R8G8_SINTTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT, ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R8G8_UINTTraits {
+    static constexpr size_t num_components = 2;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT, ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {8, 8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::G,
+                                                                              Swizzle::R};
+};
+
+struct R16_UNORMTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R16_SNORMTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R16_SINTTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R16_UINTTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R16_FLOATTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::FLOAT};
+    static constexpr std::array<size_t, num_components> component_sizes = {16};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R8_UNORMTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R8_SNORMTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SNORM};
+    static constexpr std::array<size_t, num_components> component_sizes = {8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R8_SINTTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::SINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+struct R8_UINTTraits {
+    static constexpr size_t num_components = 1;
+    static constexpr std::array<ComponentType, num_components> component_types = {
+        ComponentType::UINT};
+    static constexpr std::array<size_t, num_components> component_sizes = {8};
+    static constexpr std::array<Swizzle, num_components> component_swizzle = {Swizzle::R};
+};
+
+template <class ConverterTraits>
+class ConverterImpl : public Converter {
+private:
+    static constexpr size_t num_components = ConverterTraits::num_components;
+    static constexpr std::array<ComponentType, num_components> component_types =
+        ConverterTraits::component_types;
+    static constexpr std::array<size_t, num_components> component_sizes =
+        ConverterTraits::component_sizes;
+    static constexpr std::array<Swizzle, num_components> component_swizzle =
+        ConverterTraits::component_swizzle;
+
+    static constexpr size_t CalculateByteSize() {
+        size_t size = 0;
+        for (const size_t component_size : component_sizes) {
+            size += component_size;
+        }
+        const size_t power = (sizeof(size_t) * 8) - std::countl_zero(size) - 1ULL;
+        const size_t base_size = 1ULL << power;
+        const size_t mask = base_size - 1ULL;
+        return ((size & mask) != 0 ? base_size << 1ULL : base_size) / 8;
+    }
+
+    static constexpr size_t total_bytes_per_pixel = CalculateByteSize();
+    static constexpr size_t total_words_per_pixel =
+        (total_bytes_per_pixel + sizeof(u32) - 1U) / sizeof(u32);
+    static constexpr size_t components_per_ir_rep = 4;
+
+    template <bool get_offsets>
+    static constexpr std::array<size_t, num_components> GetBoundWordsOffsets() {
+        std::array<size_t, num_components> result;
+        result.fill(0);
+        constexpr size_t total_bits_per_word = sizeof(u32) * 8;
+        size_t accumulated_size = 0;
+        size_t count = 0;
+        for (size_t i = 0; i < num_components; i++) {
+            if constexpr (get_offsets) {
+                result[i] = accumulated_size;
+            } else {
+                result[i] = count;
+            }
+            accumulated_size += component_sizes[i];
+            if (accumulated_size > total_bits_per_word) {
+                if constexpr (get_offsets) {
+                    result[i] = 0;
+                } else {
+                    result[i]++;
+                }
+                count++;
+                accumulated_size = component_sizes[i];
+            }
+        }
+        return result;
+    }
+
+    static constexpr std::array<size_t, num_components> bound_words = GetBoundWordsOffsets<false>();
+    static constexpr std::array<size_t, num_components> bound_offsets =
+        GetBoundWordsOffsets<true>();
+
+    static constexpr std::array<u32, num_components> GetComponentsMask() {
+        std::array<u32, num_components> result;
+        for (size_t i = 0; i < num_components; i++) {
+            result[i] = (((u32)~0) >> (8 * sizeof(u32) - component_sizes[i])) << bound_offsets[i];
+        }
+        return result;
+    }
+
+    static constexpr std::array<u32, num_components> component_mask = GetComponentsMask();
+
+    // We are forcing inline so the compiler can SIMD the conversations, since it may do 4 function
+    // calls, it may fail to detect the benefit of inlining.
+    template <size_t which_component>
+    FORCE_INLINE void ConvertToComponent(u32 which_word, f32& out_component) {
+        const u32 value = (which_word >> bound_offsets[which_component]) &
+                          static_cast<u32>((1ULL << component_sizes[which_component]) - 1ULL);
+        const auto sign_extend = [](u32 base_value, size_t bits) {
+            const size_t shift_amount = sizeof(u32) * 8 - bits;
+            s32 shifted_value = static_cast<s32>(base_value << shift_amount);
+            return shifted_value >> shift_amount;
+        };
+        const auto force_to_fp16 = [](f32 base_value) {
+            u32 tmp = std::bit_cast<u32>(base_value);
+            constexpr size_t fp32_mantissa_bits = 23;
+            constexpr size_t fp16_mantissa_bits = 10;
+            constexpr size_t mantissa_mask =
+                ~((1ULL << (fp32_mantissa_bits - fp16_mantissa_bits)) - 1ULL);
+            tmp = tmp & mantissa_mask;
+            // TODO: force the exponent within the range of half float. Not needed in UNORM / SNORM
+            return std::bit_cast<f32>(tmp);
+        };
+        const auto from_fp_n = [&sign_extend](u32 base_value, size_t bits, size_t mantissa) {
+            constexpr size_t fp32_mantissa_bits = 23;
+            size_t shift_towards = fp32_mantissa_bits - mantissa;
+            const u32 new_value =
+                static_cast<u32>(sign_extend(base_value, bits) << shift_towards) & (~(1U << 31));
+            return std::bit_cast<f32>(new_value);
+        };
+        const auto calculate_snorm = [&]() {
+            return static_cast<f32>(
+                static_cast<f64>(sign_extend(value, component_sizes[which_component])) /
+                ((1ULL << (component_sizes[which_component] - 1ULL)) - 1ULL));
+        };
+        const auto calculate_unorm = [&]() {
+            return static_cast<f32>(static_cast<f32>(value) /
+                                    ((1ULL << (component_sizes[which_component])) - 1ULL));
+        };
+        if constexpr (component_types[which_component] == ComponentType::SNORM) {
+            out_component = calculate_snorm();
+        } else if constexpr (component_types[which_component] == ComponentType::UNORM) {
+            out_component = calculate_unorm();
+        } else if constexpr (component_types[which_component] == ComponentType::SINT) {
+            out_component = static_cast<f32>(
+                static_cast<s32>(sign_extend(value, component_sizes[which_component])));
+        } else if constexpr (component_types[which_component] == ComponentType::UINT) {
+            out_component = static_cast<f32>(
+                static_cast<s32>(sign_extend(value, component_sizes[which_component])));
+        } else if constexpr (component_types[which_component] == ComponentType::SNORM_FORCE_FP16) {
+            out_component = calculate_snorm();
+            out_component = force_to_fp16(out_component);
+        } else if constexpr (component_types[which_component] == ComponentType::UNORM_FORCE_FP16) {
+            out_component = calculate_unorm();
+            out_component = force_to_fp16(out_component);
+        } else if constexpr (component_types[which_component] == ComponentType::FLOAT) {
+            if constexpr (component_sizes[which_component] == 32) {
+                out_component = std::bit_cast<f32>(value);
+            } else if constexpr (component_sizes[which_component] == 16) {
+                static constexpr u32 sign_mask = 0x8000;
+                static constexpr u32 mantissa_mask = 0x8000;
+                out_component = std::bit_cast<f32>(((value & sign_mask) << 16) |
+                                                   (((value & 0x7c00) + 0x1C000) << 13) |
+                                                   ((value & mantissa_mask) << 13));
+            } else {
+                out_component = from_fp_n(value, component_sizes[which_component],
+                                          component_sizes[which_component] - 5);
+            }
+        } else if constexpr (component_types[which_component] == ComponentType::SRGB) {
+            if constexpr (component_swizzle[which_component] == Swizzle::A) {
+                out_component = calculate_unorm();
+            } else if constexpr (component_sizes[which_component] == 8) {
+                out_component = SRGB_TO_RGB_LUT[value];
+            } else {
+                out_component = calculate_unorm();
+                UNIMPLEMENTED_MSG("SRGB Conversion with component sizes of {} is unimplemented",
+                                  component_sizes[which_component]);
+            }
+        }
+    }
+
+    // We are forcing inline so the compiler can SIMD the conversations, since it may do 4 function
+    // calls, it may fail to detect the benefit of inlining.
+    template <size_t which_component>
+    FORCE_INLINE void ConvertFromComponent(u32& which_word, f32 in_component) {
+        const auto insert_to_word = [&]<typename T>(T new_word) {
+            which_word |= (static_cast<u32>(new_word) << bound_offsets[which_component]) &
+                          component_mask[which_component];
+        };
+        const auto to_fp_n = [](f32 base_value, size_t bits, size_t mantissa) {
+            constexpr size_t fp32_mantissa_bits = 23;
+            u32 tmp_value = std::bit_cast<u32>(std::max(base_value, 0.0f));
+            size_t shift_towards = fp32_mantissa_bits - mantissa;
+            return tmp_value >> shift_towards;
+        };
+        const auto calculate_unorm = [&]() {
+            return static_cast<u32>(static_cast<f32>(in_component) *
+                                    ((1ULL << (component_sizes[which_component])) - 1ULL));
+        };
+        if constexpr (component_types[which_component] == ComponentType::SNORM ||
+                      component_types[which_component] == ComponentType::SNORM_FORCE_FP16) {
+            s32 tmp_word =
+                static_cast<s32>(static_cast<f64>(in_component) *
+                                 ((1ULL << (component_sizes[which_component] - 1ULL)) - 1ULL));
+            insert_to_word(tmp_word);
+
+        } else if constexpr (component_types[which_component] == ComponentType::UNORM ||
+                             component_types[which_component] == ComponentType::UNORM_FORCE_FP16) {
+            u32 tmp_word = calculate_unorm();
+            insert_to_word(tmp_word);
+        } else if constexpr (component_types[which_component] == ComponentType::SINT) {
+            s32 tmp_word = static_cast<s32>(in_component);
+            insert_to_word(tmp_word);
+        } else if constexpr (component_types[which_component] == ComponentType::UINT) {
+            u32 tmp_word = static_cast<u32>(in_component);
+            insert_to_word(tmp_word);
+        } else if constexpr (component_types[which_component] == ComponentType::FLOAT) {
+            if constexpr (component_sizes[which_component] == 32) {
+                u32 tmp_word = std::bit_cast<u32>(in_component);
+                insert_to_word(tmp_word);
+            } else if constexpr (component_sizes[which_component] == 16) {
+                static constexpr u32 sign_mask = 0x8000;
+                static constexpr u32 mantissa_mask = 0x8000;
+                const u32 tmp_word = std::bit_cast<u32>(in_component);
+                const u32 half = ((tmp_word >> 16) & sign_mask) |
+                                 ((((tmp_word & 0x7f800000) - 0x38000000) >> 13) & 0x7c00) |
+                                 ((tmp_word >> 13) & 0x03ff);
+                insert_to_word(half);
+            } else {
+                insert_to_word(to_fp_n(in_component, component_sizes[which_component],
+                                       component_sizes[which_component] - 5));
+            }
+        } else if constexpr (component_types[which_component] == ComponentType::SRGB) {
+            if constexpr (component_swizzle[which_component] != Swizzle::A) {
+                if constexpr (component_sizes[which_component] == 8) {
+                    const u32 index = calculate_unorm();
+                    in_component = RGB_TO_SRGB_LUT[index];
+                } else {
+                    UNIMPLEMENTED_MSG("SRGB Conversion with component sizes of {} is unimplemented",
+                                      component_sizes[which_component]);
+                }
+            }
+            const u32 tmp_word = calculate_unorm();
+            insert_to_word(tmp_word);
+        }
+    }
+
+public:
+    void ConvertTo(std::span<u8> input, std::span<f32> output) override {
+        const size_t num_pixels = output.size() / components_per_ir_rep;
+        for (size_t pixel = 0; pixel < num_pixels; pixel++) {
+            std::array<u32, total_words_per_pixel> words{};
+
+            std::memcpy(words.data(), &input[pixel * total_bytes_per_pixel], total_bytes_per_pixel);
+            std::span<f32> new_components(&output[pixel * components_per_ir_rep],
+                                          components_per_ir_rep);
+            if constexpr (component_swizzle[0] != Swizzle::None) {
+                ConvertToComponent<0>(words[bound_words[0]],
+                                      new_components[static_cast<size_t>(component_swizzle[0])]);
+            } else {
+                new_components[0] = 0.0f;
+            }
+            if constexpr (num_components >= 2) {
+                if constexpr (component_swizzle[1] != Swizzle::None) {
+                    ConvertToComponent<1>(
+                        words[bound_words[1]],
+                        new_components[static_cast<size_t>(component_swizzle[1])]);
+                } else {
+                    new_components[1] = 0.0f;
+                }
+            } else {
+                new_components[1] = 0.0f;
+            }
+            if constexpr (num_components >= 3) {
+                if constexpr (component_swizzle[2] != Swizzle::None) {
+                    ConvertToComponent<2>(
+                        words[bound_words[2]],
+                        new_components[static_cast<size_t>(component_swizzle[2])]);
+                } else {
+                    new_components[2] = 0.0f;
+                }
+            } else {
+                new_components[2] = 0.0f;
+            }
+            if constexpr (num_components >= 4) {
+                if constexpr (component_swizzle[3] != Swizzle::None) {
+                    ConvertToComponent<3>(
+                        words[bound_words[3]],
+                        new_components[static_cast<size_t>(component_swizzle[3])]);
+                } else {
+                    new_components[3] = 0.0f;
+                }
+            } else {
+                new_components[3] = 0.0f;
+            }
+        }
+    }
+
+    void ConvertFrom(std::span<f32> input, std::span<u8> output) override {
+        const size_t num_pixels = output.size() / total_bytes_per_pixel;
+        for (size_t pixel = 0; pixel < num_pixels; pixel++) {
+            std::span<f32> old_components(&input[pixel * components_per_ir_rep],
+                                          components_per_ir_rep);
+            std::array<u32, total_words_per_pixel> words{};
+            if constexpr (component_swizzle[0] != Swizzle::None) {
+                ConvertFromComponent<0>(words[bound_words[0]],
+                                        old_components[static_cast<size_t>(component_swizzle[0])]);
+            }
+            if constexpr (num_components >= 2) {
+                if constexpr (component_swizzle[1] != Swizzle::None) {
+                    ConvertFromComponent<1>(
+                        words[bound_words[1]],
+                        old_components[static_cast<size_t>(component_swizzle[1])]);
+                }
+            }
+            if constexpr (num_components >= 3) {
+                if constexpr (component_swizzle[2] != Swizzle::None) {
+                    ConvertFromComponent<2>(
+                        words[bound_words[2]],
+                        old_components[static_cast<size_t>(component_swizzle[2])]);
+                }
+            }
+            if constexpr (num_components >= 4) {
+                if constexpr (component_swizzle[3] != Swizzle::None) {
+                    ConvertFromComponent<3>(
+                        words[bound_words[3]],
+                        old_components[static_cast<size_t>(component_swizzle[3])]);
+                }
+            }
+            std::memcpy(&output[pixel * total_bytes_per_pixel], words.data(),
+                        total_bytes_per_pixel);
+        }
+    }
+
+    ConverterImpl() = default;
+    ~ConverterImpl() = default;
+};
+
+struct ConverterFactory::ConverterFactoryImpl {
+    std::unordered_map<RenderTargetFormat, std::unique_ptr<Converter>> converters_cache;
+};
+
+ConverterFactory::ConverterFactory() {
+    impl = std::make_unique<ConverterFactoryImpl>();
+}
+
+ConverterFactory::~ConverterFactory() = default;
+
+Converter* ConverterFactory::GetFormatConverter(RenderTargetFormat format) {
+    auto it = impl->converters_cache.find(format);
+    if (it == impl->converters_cache.end()) [[unlikely]] {
+        return BuildConverter(format);
+    }
+    return it->second.get();
+}
+
+class NullConverter : public Converter {
+public:
+    void ConvertTo([[maybe_unused]] std::span<u8> input, std::span<f32> output) override {
+        std::fill(output.begin(), output.end(), 0.0f);
+    }
+    void ConvertFrom([[maybe_unused]] std::span<f32> input, std::span<u8> output) override {
+        const u8 fill_value = 0U;
+        std::fill(output.begin(), output.end(), fill_value);
+    }
+};
+
+Converter* ConverterFactory::BuildConverter(RenderTargetFormat format) {
+    switch (format) {
+    case RenderTargetFormat::R32B32G32A32_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32B32G32A32_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32B32A32_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32B32A32_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32B32A32_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32B32A32_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16B16A16_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16B16A16_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16B16A16_SNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16B16A16_SNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16B16A16_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16B16A16_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16B16A16_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16B16A16_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16B16A16_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16B16A16_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32G32_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32G32_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16B16X16_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16B16X16_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8R8G8B8_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8R8G8B8_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8R8G8B8_SRGB:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8R8G8B8_SRGBTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A2B10G10R10_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A2B10G10R10_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A2B10G10R10_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A2B10G10R10_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8B8G8R8_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8B8G8R8_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8B8G8R8_SRGB:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8B8G8R8_SRGBTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8B8G8R8_SNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8B8G8R8_SNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8B8G8R8_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8B8G8R8_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A8B8G8R8_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A8B8G8R8_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16_SNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16_SNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16G16_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16G16_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::B10G11R11_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<B10G11R11_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R32_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R32_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R5G6B5_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R5G6B5_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::A1R5G5B5_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<A1R5G5B5_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8G8_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8G8_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8G8_SNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8G8_SNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8G8_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8G8_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8G8_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8G8_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16_SNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16_SNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16_UINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R16_FLOAT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R16_FLOATTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8_UNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8_UNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8_SNORM:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8_SNORMTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8_SINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8_SINTTraits>>())
+            .first->second.get();
+        break;
+    case RenderTargetFormat::R8_UINT:
+        return impl->converters_cache
+            .emplace(format, std::make_unique<ConverterImpl<R8_UINTTraits>>())
+            .first->second.get();
+        break;
+    default: {
+        UNIMPLEMENTED_MSG("This format {} converter is not implemented", format);
+        return impl->converters_cache.emplace(format, std::make_unique<NullConverter>())
+            .first->second.get();
+    }
+    }
+}
+
+} // namespace Tegra::Engines::Blitter

src/video_core/engines/sw_blitter/converter.h

@@ -0,0 +1,35 @@
+// SPDX-FileCopyrightText: Copyright 2022 yuzu Emulator Project
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+#include <memory>
+#include <span>
+
+#include "common/common_types.h"
+
+#pragma once
+
+#include "video_core/gpu.h"
+
+namespace Tegra::Engines::Blitter {
+
+class Converter {
+public:
+    virtual void ConvertTo(std::span<u8> input, std::span<f32> output) = 0;
+    virtual void ConvertFrom(std::span<f32> input, std::span<u8> output) = 0;
+};
+
+class ConverterFactory {
+public:
+    ConverterFactory();
+    ~ConverterFactory();
+
+    Converter* GetFormatConverter(RenderTargetFormat format);
+
+private:
+    Converter* BuildConverter(RenderTargetFormat format);
+
+    struct ConverterFactoryImpl;
+    std::unique_ptr<ConverterFactoryImpl> impl;
+};
+
+} // namespace Tegra::Engines::Blitter

src/video_core/renderer_opengl/gl_rasterizer.cpp

@@ -466,8 +466,7 @@ bool RasterizerOpenGL::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Surf
                                              const Tegra::Engines::Fermi2D::Config& copy_config) {
     MICROPROFILE_SCOPE(OpenGL_Blits);
     std::scoped_lock lock{texture_cache.mutex};
-    texture_cache.BlitImage(dst, src, copy_config);
-    return true;
+    return texture_cache.BlitImage(dst, src, copy_config);
 }
 
 Tegra::Engines::AccelerateDMAInterface& RasterizerOpenGL::AccessAccelerateDMA() {

src/video_core/renderer_vulkan/vk_rasterizer.cpp

@@ -542,8 +542,7 @@ bool RasterizerVulkan::AccelerateSurfaceCopy(const Tegra::Engines::Fermi2D::Surf
                                              const Tegra::Engines::Fermi2D::Surface& dst,
                                              const Tegra::Engines::Fermi2D::Config& copy_config) {
     std::scoped_lock lock{texture_cache.mutex};
-    texture_cache.BlitImage(dst, src, copy_config);
-    return true;
+    return texture_cache.BlitImage(dst, src, copy_config);
 }
 
 Tegra::Engines::AccelerateDMAInterface& RasterizerVulkan::AccessAccelerateDMA() {

src/video_core/texture_cache/texture_cache.h

@@ -506,10 +506,14 @@ void TextureCache<P>::UnmapGPUMemory(size_t as_id, GPUVAddr gpu_addr, size_t siz
 }
 
 template <class P>
-void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
+bool TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
                                 const Tegra::Engines::Fermi2D::Surface& src,
                                 const Tegra::Engines::Fermi2D::Config& copy) {
-    const BlitImages images = GetBlitImages(dst, src, copy);
+    const auto result = GetBlitImages(dst, src, copy);
+    if (!result) {
+        return false;
+    }
+    const BlitImages images = *result;
     const ImageId dst_id = images.dst_id;
     const ImageId src_id = images.src_id;
 
@@ -596,6 +600,7 @@ void TextureCache<P>::BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
         runtime.BlitImage(dst_framebuffer, dst_view, src_view, dst_region, src_region, copy.filter,
                           copy.operation);
     }
+    return true;
 }
 
 template <class P>
@@ -1133,7 +1138,7 @@ ImageId TextureCache<P>::JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VA
 }
 
 template <class P>
-typename TextureCache<P>::BlitImages TextureCache<P>::GetBlitImages(
+std::optional<typename TextureCache<P>::BlitImages> TextureCache<P>::GetBlitImages(
     const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Surface& src,
     const Tegra::Engines::Fermi2D::Config& copy) {
 
@@ -1154,6 +1159,20 @@ typename TextureCache<P>::BlitImages TextureCache<P>::GetBlitImages(
         has_deleted_images = false;
         src_id = FindImage(src_info, src_addr, try_options);
         dst_id = FindImage(dst_info, dst_addr, try_options);
+        if (!copy.must_accelerate) {
+            do {
+                if (!src_id && !dst_id) {
+                    return std::nullopt;
+                }
+                if (src_id && True(slot_images[src_id].flags & ImageFlagBits::GpuModified)) {
+                    break;
+                }
+                if (dst_id && True(slot_images[dst_id].flags & ImageFlagBits::GpuModified)) {
+                    break;
+                }
+                return std::nullopt;
+            } while (false);
+        }
         const ImageBase* const src_image = src_id ? &slot_images[src_id] : nullptr;
         if (src_image && src_image->info.num_samples > 1) {
             RelaxedOptions find_options{FIND_OPTIONS | RelaxedOptions::ForceBrokenViews};
@@ -1194,12 +1213,12 @@ typename TextureCache<P>::BlitImages TextureCache<P>::GetBlitImages(
             dst_id = FindOrInsertImage(dst_info, dst_addr, RelaxedOptions{});
         } while (has_deleted_images);
     }
-    return BlitImages{
+    return {BlitImages{
         .dst_id = dst_id,
         .src_id = src_id,
         .dst_format = dst_info.format,
         .src_format = src_info.format,
-    };
+    }};
 }
 
 template <class P>

src/video_core/texture_cache/texture_cache_base.h

@@ -174,7 +174,7 @@ public:
     void UnmapGPUMemory(size_t as_id, GPUVAddr gpu_addr, size_t size);
 
     /// Blit an image with the given parameters
-    void BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
+    bool BlitImage(const Tegra::Engines::Fermi2D::Surface& dst,
                    const Tegra::Engines::Fermi2D::Surface& src,
                    const Tegra::Engines::Fermi2D::Config& copy);
 
@@ -285,9 +285,9 @@ private:
     [[nodiscard]] ImageId JoinImages(const ImageInfo& info, GPUVAddr gpu_addr, VAddr cpu_addr);
 
     /// Return a blit image pair from the given guest blit parameters
-    [[nodiscard]] BlitImages GetBlitImages(const Tegra::Engines::Fermi2D::Surface& dst,
-                                           const Tegra::Engines::Fermi2D::Surface& src,
-                                           const Tegra::Engines::Fermi2D::Config& copy);
+    [[nodiscard]] std::optional<BlitImages> GetBlitImages(
+        const Tegra::Engines::Fermi2D::Surface& dst, const Tegra::Engines::Fermi2D::Surface& src,
+        const Tegra::Engines::Fermi2D::Config& copy);
 
     /// Find or create a sampler from a guest descriptor sampler
     [[nodiscard]] SamplerId FindSampler(const TSCEntry& config);