62 files changed, 1303 insertions, 663 deletions

diff --git a/src/audio_core/CMakeLists.txt b/src/audio_core/CMakeLists.txt
index d1d177b51..a0ae07752 100644
--- a/src/audio_core/CMakeLists.txt
+++ b/src/audio_core/CMakeLists.txt
@@ -15,6 +15,8 @@ add_library(audio_core STATIC
     command_generator.cpp
     command_generator.h
     common.h
+    delay_line.cpp
+    delay_line.h
     effect_context.cpp
     effect_context.h
     info_updater.cpp
diff --git a/src/audio_core/command_generator.cpp b/src/audio_core/command_generator.cpp
index 5b1065520..437cc5ccd 100644
--- a/src/audio_core/command_generator.cpp
+++ b/src/audio_core/command_generator.cpp
@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cmath>
+#include <numbers>
 #include "audio_core/algorithm/interpolate.h"
 #include "audio_core/command_generator.h"
 #include "audio_core/effect_context.h"
@@ -13,6 +15,20 @@ namespace AudioCore {
 namespace {
 constexpr std::size_t MIX_BUFFER_SIZE = 0x3f00;
 constexpr std::size_t SCALED_MIX_BUFFER_SIZE = MIX_BUFFER_SIZE << 15ULL;
+using DelayLineTimes = std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT>;
+
+constexpr DelayLineTimes FDN_MIN_DELAY_LINE_TIMES{5.0f, 6.0f, 13.0f, 14.0f};
+constexpr DelayLineTimes FDN_MAX_DELAY_LINE_TIMES{45.704f, 82.782f, 149.94f, 271.58f};
+constexpr DelayLineTimes DECAY0_MAX_DELAY_LINE_TIMES{17.0f, 13.0f, 9.0f, 7.0f};
+constexpr DelayLineTimes DECAY1_MAX_DELAY_LINE_TIMES{19.0f, 11.0f, 10.0f, 6.0f};
+constexpr std::array<f32, AudioCommon::I3DL2REVERB_TAPS> EARLY_TAP_TIMES{
+    0.017136f, 0.059154f, 0.161733f, 0.390186f, 0.425262f, 0.455411f, 0.689737f,
+    0.745910f, 0.833844f, 0.859502f, 0.000000f, 0.075024f, 0.168788f, 0.299901f,
+    0.337443f, 0.371903f, 0.599011f, 0.716741f, 0.817859f, 0.851664f};
+constexpr std::array<f32, AudioCommon::I3DL2REVERB_TAPS> EARLY_GAIN{
+    0.67096f, 0.61027f, 1.0f,     0.35680f, 0.68361f, 0.65978f, 0.51939f,
+    0.24712f, 0.45945f, 0.45021f, 0.64196f, 0.54879f, 0.92925f, 0.38270f,
+    0.72867f, 0.69794f, 0.5464f,  0.24563f, 0.45214f, 0.44042f};
 
 template <std::size_t N>
 void ApplyMix(s32* output, const s32* input, s32 gain, s32 sample_count) {
@@ -65,6 +81,154 @@ s32 ApplyMixDepop(s32* output, s32 first_sample, s32 delta, s32 sample_count) {
     }
 }
 
+float Pow10(float x) {
+    if (x >= 0.0f) {
+        return 1.0f;
+    } else if (x <= -5.3f) {
+        return 0.0f;
+    }
+    return std::pow(10.0f, x);
+}
+
+float SinD(float degrees) {
+    return std::sin(degrees * std::numbers::pi_v<float> / 180.0f);
+}
+
+float CosD(float degrees) {
+    return std::cos(degrees * std::numbers::pi_v<float> / 180.0f);
+}
+
+float ToFloat(s32 sample) {
+    return static_cast<float>(sample) / 65536.f;
+}
+
+s32 ToS32(float sample) {
+    constexpr auto min = -8388608.0f;
+    constexpr auto max = 8388607.f;
+    float rescaled_sample = sample * 65536.0f;
+    if (rescaled_sample < min) {
+        rescaled_sample = min;
+    }
+    if (rescaled_sample > max) {
+        rescaled_sample = max;
+    }
+    return static_cast<s32>(rescaled_sample);
+}
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_1CH{0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+                                                           0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_2CH{0, 0, 0, 1, 1, 1, 1, 0, 0, 0,
+                                                           1, 1, 1, 0, 0, 0, 0, 1, 1, 1};
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_4CH{0, 0, 0, 1, 1, 1, 1, 2, 2, 2,
+                                                           1, 1, 1, 0, 0, 0, 0, 3, 3, 3};
+
+constexpr std::array<std::size_t, 20> REVERB_TAP_INDEX_6CH{4, 0, 0, 1, 1, 1, 1, 2, 2, 2,
+                                                           1, 1, 1, 0, 0, 0, 0, 3, 3, 3};
+
+template <std::size_t CHANNEL_COUNT>
+void ApplyReverbGeneric(I3dl2ReverbState& state,
+                        const std::array<const s32*, AudioCommon::MAX_CHANNEL_COUNT>& input,
+                        const std::array<s32*, AudioCommon::MAX_CHANNEL_COUNT>& output,
+                        s32 sample_count) {
+
+    auto GetTapLookup = []() {
+        if constexpr (CHANNEL_COUNT == 1) {
+            return REVERB_TAP_INDEX_1CH;
+        } else if constexpr (CHANNEL_COUNT == 2) {
+            return REVERB_TAP_INDEX_2CH;
+        } else if constexpr (CHANNEL_COUNT == 4) {
+            return REVERB_TAP_INDEX_4CH;
+        } else if constexpr (CHANNEL_COUNT == 6) {
+            return REVERB_TAP_INDEX_6CH;
+        }
+    };
+
+    const auto& tap_index_lut = GetTapLookup();
+    for (s32 sample = 0; sample < sample_count; sample++) {
+        std::array<f32, CHANNEL_COUNT> out_samples{};
+        std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> fsamp{};
+        std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> mixed{};
+        std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> osamp{};
+
+        // Mix everything into a single sample
+        s32 temp_mixed_sample = 0;
+        for (std::size_t i = 0; i < CHANNEL_COUNT; i++) {
+            temp_mixed_sample += input[i][sample];
+        }
+        const auto current_sample = ToFloat(temp_mixed_sample);
+        const auto early_tap = state.early_delay_line.TapOut(state.early_to_late_taps);
+
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_TAPS; i++) {
+            const auto tapped_samp =
+                state.early_delay_line.TapOut(state.early_tap_steps[i]) * EARLY_GAIN[i];
+            out_samples[tap_index_lut[i]] += tapped_samp;
+
+            if constexpr (CHANNEL_COUNT == 6) {
+                // handle lfe
+                out_samples[5] += tapped_samp;
+            }
+        }
+
+        state.lowpass_0 = current_sample * state.lowpass_2 + state.lowpass_0 * state.lowpass_1;
+        state.early_delay_line.Tick(state.lowpass_0);
+
+        for (std::size_t i = 0; i < CHANNEL_COUNT; i++) {
+            out_samples[i] *= state.early_gain;
+        }
+
+        // Two channel seems to apply a latet gain, we require to save this
+        f32 filter{};
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+            filter = state.fdn_delay_line[i].GetOutputSample();
+            const auto computed = filter * state.lpf_coefficients[0][i] + state.shelf_filter[i];
+            state.shelf_filter[i] =
+                filter * state.lpf_coefficients[1][i] + computed * state.lpf_coefficients[2][i];
+            fsamp[i] = computed;
+        }
+
+        // Mixing matrix
+        mixed[0] = fsamp[1] + fsamp[2];
+        mixed[1] = -fsamp[0] - fsamp[3];
+        mixed[2] = fsamp[0] - fsamp[3];
+        mixed[3] = fsamp[1] - fsamp[2];
+
+        if constexpr (CHANNEL_COUNT == 2) {
+            for (auto& mix : mixed) {
+                mix *= (filter * state.late_gain);
+            }
+        }
+
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+            const auto late = early_tap * state.late_gain;
+            osamp[i] = state.decay_delay_line0[i].Tick(late + mixed[i]);
+            osamp[i] = state.decay_delay_line1[i].Tick(osamp[i]);
+            state.fdn_delay_line[i].Tick(osamp[i]);
+        }
+
+        if constexpr (CHANNEL_COUNT == 1) {
+            output[0][sample] = ToS32(state.dry_gain * ToFloat(input[0][sample]) +
+                                      (out_samples[0] + osamp[0] + osamp[1]));
+        } else if constexpr (CHANNEL_COUNT == 2 || CHANNEL_COUNT == 4) {
+            for (std::size_t i = 0; i < CHANNEL_COUNT; i++) {
+                output[i][sample] =
+                    ToS32(state.dry_gain * ToFloat(input[i][sample]) + (out_samples[i] + osamp[i]));
+            }
+        } else if constexpr (CHANNEL_COUNT == 6) {
+            const auto temp_center = state.center_delay_line.Tick(0.5f * (osamp[2] - osamp[3]));
+            for (std::size_t i = 0; i < 4; i++) {
+                output[i][sample] =
+                    ToS32(state.dry_gain * ToFloat(input[i][sample]) + (out_samples[i] + osamp[i]));
+            }
+            output[4][sample] =
+                ToS32(state.dry_gain * ToFloat(input[4][sample]) + (out_samples[4] + temp_center));
+            output[5][sample] =
+                ToS32(state.dry_gain * ToFloat(input[5][sample]) + (out_samples[5] + osamp[3]));
+        }
+    }
+}
+
 } // namespace
 
 CommandGenerator::CommandGenerator(AudioCommon::AudioRendererParameter& worker_params_,
@@ -271,11 +435,10 @@ void CommandGenerator::GenerateBiquadFilterCommandForVoice(ServerVoiceInfo& voic
         }
 
         // Generate biquad filter
-        //        GenerateBiquadFilterCommand(mix_buffer_count, biquad_filter,
-        //        dsp_state.biquad_filter_state,
-        //                                    mix_buffer_count + channel, mix_buffer_count +
-        //                                    channel, worker_params.sample_count,
-        //                                    voice_info.GetInParams().node_id);
+        // GenerateBiquadFilterCommand(mix_buffer_count, biquad_filter,
+        // dsp_state.biquad_filter_state,
+        //                            mix_buffer_count + channel, mix_buffer_count + channel,
+        //                            worker_params.sample_count, voice_info.GetInParams().node_id);
     }
 }
 
@@ -376,21 +539,54 @@ void CommandGenerator::GenerateEffectCommand(ServerMixInfo& mix_info) {
 
 void CommandGenerator::GenerateI3dl2ReverbEffectCommand(s32 mix_buffer_offset, EffectBase* info,
                                                         bool enabled) {
-    if (!enabled) {
+    auto* reverb = dynamic_cast<EffectI3dl2Reverb*>(info);
+    const auto& params = reverb->GetParams();
+    auto& state = reverb->GetState();
+    const auto channel_count = params.channel_count;
+
+    if (channel_count != 1 && channel_count != 2 && channel_count != 4 && channel_count != 6) {
         return;
     }
-    const auto& params = dynamic_cast<EffectI3dl2Reverb*>(info)->GetParams();
-    const auto channel_count = params.channel_count;
+
+    std::array<const s32*, AudioCommon::MAX_CHANNEL_COUNT> input{};
+    std::array<s32*, AudioCommon::MAX_CHANNEL_COUNT> output{};
+
+    const auto status = params.status;
     for (s32 i = 0; i < channel_count; i++) {
-        // TODO(ogniK): Actually implement reverb
-        /*
-        if (params.input[i] != params.output[i]) {
-            const auto* input = GetMixBuffer(mix_buffer_offset + params.input[i]);
-            auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]);
-            ApplyMix<1>(output, input, 32768, worker_params.sample_count);
-        }*/
-        auto* output = GetMixBuffer(mix_buffer_offset + params.output[i]);
-        std::memset(output, 0, worker_params.sample_count * sizeof(s32));
+        input[i] = GetMixBuffer(mix_buffer_offset + params.input[i]);
+        output[i] = GetMixBuffer(mix_buffer_offset + params.output[i]);
+    }
+
+    if (enabled) {
+        if (status == ParameterStatus::Initialized) {
+            InitializeI3dl2Reverb(reverb->GetParams(), state, info->GetWorkBuffer());
+        } else if (status == ParameterStatus::Updating) {
+            UpdateI3dl2Reverb(reverb->GetParams(), state, false);
+        }
+    }
+
+    if (enabled) {
+        switch (channel_count) {
+        case 1:
+            ApplyReverbGeneric<1>(state, input, output, worker_params.sample_count);
+            break;
+        case 2:
+            ApplyReverbGeneric<2>(state, input, output, worker_params.sample_count);
+            break;
+        case 4:
+            ApplyReverbGeneric<4>(state, input, output, worker_params.sample_count);
+            break;
+        case 6:
+            ApplyReverbGeneric<6>(state, input, output, worker_params.sample_count);
+            break;
+        }
+    } else {
+        for (s32 i = 0; i < channel_count; i++) {
+            // Only copy if the buffer input and output do not match!
+            if ((mix_buffer_offset + params.input[i]) != (mix_buffer_offset + params.output[i])) {
+                std::memcpy(output[i], input[i], worker_params.sample_count * sizeof(s32));
+            }
+        }
     }
 }
 
@@ -528,6 +724,133 @@ s32 CommandGenerator::ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u3
     return sample_count;
 }
 
+void CommandGenerator::InitializeI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
+                                             std::vector<u8>& work_buffer) {
+    // Reset state
+    state.lowpass_0 = 0.0f;
+    state.lowpass_1 = 0.0f;
+    state.lowpass_2 = 0.0f;
+
+    state.early_delay_line.Reset();
+    state.early_tap_steps.fill(0);
+    state.early_gain = 0.0f;
+    state.late_gain = 0.0f;
+    state.early_to_late_taps = 0;
+    for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+        state.fdn_delay_line[i].Reset();
+        state.decay_delay_line0[i].Reset();
+        state.decay_delay_line1[i].Reset();
+    }
+    state.last_reverb_echo = 0.0f;
+    state.center_delay_line.Reset();
+    for (auto& coef : state.lpf_coefficients) {
+        coef.fill(0.0f);
+    }
+    state.shelf_filter.fill(0.0f);
+    state.dry_gain = 0.0f;
+
+    const auto sample_rate = info.sample_rate / 1000;
+    f32* work_buffer_ptr = reinterpret_cast<f32*>(work_buffer.data());
+
+    s32 delay_samples{};
+    for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+        delay_samples =
+            AudioCommon::CalculateDelaySamples(sample_rate, FDN_MAX_DELAY_LINE_TIMES[i]);
+        state.fdn_delay_line[i].Initialize(delay_samples, work_buffer_ptr);
+        work_buffer_ptr += delay_samples + 1;
+
+        delay_samples =
+            AudioCommon::CalculateDelaySamples(sample_rate, DECAY0_MAX_DELAY_LINE_TIMES[i]);
+        state.decay_delay_line0[i].Initialize(delay_samples, 0.0f, work_buffer_ptr);
+        work_buffer_ptr += delay_samples + 1;
+
+        delay_samples =
+            AudioCommon::CalculateDelaySamples(sample_rate, DECAY1_MAX_DELAY_LINE_TIMES[i]);
+        state.decay_delay_line1[i].Initialize(delay_samples, 0.0f, work_buffer_ptr);
+        work_buffer_ptr += delay_samples + 1;
+    }
+    delay_samples = AudioCommon::CalculateDelaySamples(sample_rate, 5.0f);
+    state.center_delay_line.Initialize(delay_samples, work_buffer_ptr);
+    work_buffer_ptr += delay_samples + 1;
+
+    delay_samples = AudioCommon::CalculateDelaySamples(sample_rate, 400.0f);
+    state.early_delay_line.Initialize(delay_samples, work_buffer_ptr);
+
+    UpdateI3dl2Reverb(info, state, true);
+}
+
+void CommandGenerator::UpdateI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
+                                         bool should_clear) {
+
+    state.dry_gain = info.dry_gain;
+    state.shelf_filter.fill(0.0f);
+    state.lowpass_0 = 0.0f;
+    state.early_gain = Pow10(std::min(info.room + info.reflection, 5000.0f) / 2000.0f);
+    state.late_gain = Pow10(std::min(info.room + info.reverb, 5000.0f) / 2000.0f);
+
+    const auto sample_rate = info.sample_rate / 1000;
+    const f32 hf_gain = Pow10(info.room_hf / 2000.0f);
+    if (hf_gain >= 1.0f) {
+        state.lowpass_2 = 1.0f;
+        state.lowpass_1 = 0.0f;
+    } else {
+        const auto a = 1.0f - hf_gain;
+        const auto b = 2.0f * (1.0f - hf_gain * CosD(256.0f * info.hf_reference /
+                                                     static_cast<f32>(info.sample_rate)));
+        const auto c = std::sqrt(b * b - 4.0f * a * a);
+
+        state.lowpass_1 = (b - c) / (2.0f * a);
+        state.lowpass_2 = 1.0f - state.lowpass_1;
+    }
+    state.early_to_late_taps = AudioCommon::CalculateDelaySamples(
+        sample_rate, 1000.0f * (info.reflection_delay + info.reverb_delay));
+
+    state.last_reverb_echo = 0.6f * info.diffusion * 0.01f;
+    for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+        const auto length =
+            FDN_MIN_DELAY_LINE_TIMES[i] +
+            (info.density / 100.0f) * (FDN_MAX_DELAY_LINE_TIMES[i] - FDN_MIN_DELAY_LINE_TIMES[i]);
+        state.fdn_delay_line[i].SetDelay(AudioCommon::CalculateDelaySamples(sample_rate, length));
+
+        const auto delay_sample_counts = state.fdn_delay_line[i].GetDelay() +
+                                         state.decay_delay_line0[i].GetDelay() +
+                                         state.decay_delay_line1[i].GetDelay();
+
+        float a = (-60.0f * static_cast<f32>(delay_sample_counts)) /
+                  (info.decay_time * static_cast<f32>(info.sample_rate));
+        float b = a / info.hf_decay_ratio;
+        float c = CosD(128.0f * 0.5f * info.hf_reference / static_cast<f32>(info.sample_rate)) /
+                  SinD(128.0f * 0.5f * info.hf_reference / static_cast<f32>(info.sample_rate));
+        float d = Pow10((b - a) / 40.0f);
+        float e = Pow10((b + a) / 40.0f) * 0.7071f;
+
+        state.lpf_coefficients[0][i] = e * ((d * c) + 1.0f) / (c + d);
+        state.lpf_coefficients[1][i] = e * (1.0f - (d * c)) / (c + d);
+        state.lpf_coefficients[2][i] = (c - d) / (c + d);
+
+        state.decay_delay_line0[i].SetCoefficient(state.last_reverb_echo);
+        state.decay_delay_line1[i].SetCoefficient(-0.9f * state.last_reverb_echo);
+    }
+
+    if (should_clear) {
+        for (std::size_t i = 0; i < AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT; i++) {
+            state.fdn_delay_line[i].Clear();
+            state.decay_delay_line0[i].Clear();
+            state.decay_delay_line1[i].Clear();
+        }
+        state.early_delay_line.Clear();
+        state.center_delay_line.Clear();
+    }
+
+    const auto max_early_delay = state.early_delay_line.GetMaxDelay();
+    const auto reflection_time = 1000.0f * (0.0098f * info.reverb_delay + 0.02f);
+    for (std::size_t tap = 0; tap < AudioCommon::I3DL2REVERB_TAPS; tap++) {
+        const auto length = AudioCommon::CalculateDelaySamples(
+            sample_rate, 1000.0f * info.reflection_delay + reflection_time * EARLY_TAP_TIMES[tap]);
+        state.early_tap_steps[tap] = std::min(length, max_early_delay);
+    }
+}
+
 void CommandGenerator::GenerateVolumeRampCommand(float last_volume, float current_volume,
                                                  s32 channel, s32 node_id) {
     const auto last = static_cast<s32>(last_volume * 32768.0f);
diff --git a/src/audio_core/command_generator.h b/src/audio_core/command_generator.h
index b937350b1..2ebb755b0 100644
--- a/src/audio_core/command_generator.h
+++ b/src/audio_core/command_generator.h
@@ -21,6 +21,8 @@ class ServerMixInfo;
 class EffectContext;
 class EffectBase;
 struct AuxInfoDSP;
+struct I3dl2ReverbParams;
+struct I3dl2ReverbState;
 using MixVolumeBuffer = std::array<float, AudioCommon::MAX_MIX_BUFFERS>;
 
 class CommandGenerator {
@@ -80,6 +82,9 @@ private:
     s32 ReadAuxBuffer(AuxInfoDSP& recv_info, VAddr recv_buffer, u32 max_samples, s32* out_data,
                       u32 sample_count, u32 read_offset, u32 read_count);
 
+    void InitializeI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state,
+                               std::vector<u8>& work_buffer);
+    void UpdateI3dl2Reverb(I3dl2ReverbParams& info, I3dl2ReverbState& state, bool should_clear);
     // DSP Code
     s32 DecodePcm16(ServerVoiceInfo& voice_info, VoiceState& dsp_state, s32 sample_count,
                     s32 channel, std::size_t mix_offset);
diff --git a/src/audio_core/common.h b/src/audio_core/common.h
index ec59a3ba9..fe546c55d 100644
--- a/src/audio_core/common.h
+++ b/src/audio_core/common.h
@@ -33,6 +33,29 @@ constexpr std::size_t TEMP_MIX_BASE_SIZE = 0x3f00; // TODO(ogniK): Work out this
 // and our const ends up being 0x3f04, the 4 bytes are most
 // likely the sample history
 constexpr std::size_t TOTAL_TEMP_MIX_SIZE = TEMP_MIX_BASE_SIZE + AudioCommon::MAX_SAMPLE_HISTORY;
+constexpr f32 I3DL2REVERB_MAX_LEVEL = 5000.0f;
+constexpr f32 I3DL2REVERB_MIN_REFLECTION_DURATION = 0.02f;
+constexpr std::size_t I3DL2REVERB_TAPS = 20;
+constexpr std::size_t I3DL2REVERB_DELAY_LINE_COUNT = 4;
+using Fractional = s32;
+
+template <typename T>
+constexpr Fractional ToFractional(T x) {
+    return static_cast<Fractional>(x * static_cast<T>(0x4000));
+}
+
+constexpr Fractional MultiplyFractional(Fractional lhs, Fractional rhs) {
+    return static_cast<Fractional>(static_cast<s64>(lhs) * rhs >> 14);
+}
+
+constexpr s32 FractionalToFixed(Fractional x) {
+    const auto s = x & (1 << 13);
+    return static_cast<s32>(x >> 14) + s;
+}
+
+constexpr s32 CalculateDelaySamples(s32 sample_rate_khz, float time) {
+    return FractionalToFixed(MultiplyFractional(ToFractional(sample_rate_khz), ToFractional(time)));
+}
 
 static constexpr u32 VersionFromRevision(u32_le rev) {
     // "REV7" -> 7
diff --git a/src/audio_core/delay_line.cpp b/src/audio_core/delay_line.cpp
new file mode 100644
index 000000000..f4e4dd8d2
--- /dev/null
+++ b/src/audio_core/delay_line.cpp
@@ -0,0 +1,104 @@
+#include <cstring>
+#include "audio_core/delay_line.h"
+
+namespace AudioCore {
+DelayLineBase::DelayLineBase() = default;
+DelayLineBase::~DelayLineBase() = default;
+
+void DelayLineBase::Initialize(s32 max_delay_, float* src_buffer) {
+    buffer = src_buffer;
+    buffer_end = buffer + max_delay_;
+    max_delay = max_delay_;
+    output = buffer;
+    SetDelay(max_delay_);
+    Clear();
+}
+
+void DelayLineBase::SetDelay(s32 new_delay) {
+    if (max_delay < new_delay) {
+        return;
+    }
+    delay = new_delay;
+    input = (buffer + ((output - buffer) + new_delay) % (max_delay + 1));
+}
+
+s32 DelayLineBase::GetDelay() const {
+    return delay;
+}
+
+s32 DelayLineBase::GetMaxDelay() const {
+    return max_delay;
+}
+
+f32 DelayLineBase::TapOut(s32 last_sample) {
+    const float* ptr = input - (last_sample + 1);
+    if (ptr < buffer) {
+        ptr += (max_delay + 1);
+    }
+
+    return *ptr;
+}
+
+f32 DelayLineBase::Tick(f32 sample) {
+    *(input++) = sample;
+    const auto out_sample = *(output++);
+
+    if (buffer_end < input) {
+        input = buffer;
+    }
+
+    if (buffer_end < output) {
+        output = buffer;
+    }
+
+    return out_sample;
+}
+
+float* DelayLineBase::GetInput() {
+    return input;
+}
+
+const float* DelayLineBase::GetInput() const {
+    return input;
+}
+
+f32 DelayLineBase::GetOutputSample() const {
+    return *output;
+}
+
+void DelayLineBase::Clear() {
+    std::memset(buffer, 0, sizeof(float) * max_delay);
+}
+
+void DelayLineBase::Reset() {
+    buffer = nullptr;
+    buffer_end = nullptr;
+    max_delay = 0;
+    input = nullptr;
+    output = nullptr;
+    delay = 0;
+}
+
+DelayLineAllPass::DelayLineAllPass() = default;
+DelayLineAllPass::~DelayLineAllPass() = default;
+
+void DelayLineAllPass::Initialize(u32 delay_, float coeffcient_, f32* src_buffer) {
+    DelayLineBase::Initialize(delay_, src_buffer);
+    SetCoefficient(coeffcient_);
+}
+
+void DelayLineAllPass::SetCoefficient(float coeffcient_) {
+    coefficient = coeffcient_;
+}
+
+f32 DelayLineAllPass::Tick(f32 sample) {
+    const auto temp = sample - coefficient * *output;
+    return coefficient * temp + DelayLineBase::Tick(temp);
+}
+
+void DelayLineAllPass::Reset() {
+    coefficient = 0.0f;
+    DelayLineBase::Reset();
+}
+
+} // namespace AudioCore
diff --git a/src/audio_core/delay_line.h b/src/audio_core/delay_line.h
new file mode 100644
index 000000000..cafddd432
--- /dev/null
+++ b/src/audio_core/delay_line.h
@@ -0,0 +1,46 @@
+#pragma once
+
+#include "common/common_types.h"
+
+namespace AudioCore {
+
+class DelayLineBase {
+public:
+    DelayLineBase();
+    ~DelayLineBase();
+
+    void Initialize(s32 max_delay_, float* src_buffer);
+    void SetDelay(s32 new_delay);
+    s32 GetDelay() const;
+    s32 GetMaxDelay() const;
+    f32 TapOut(s32 last_sample);
+    f32 Tick(f32 sample);
+    float* GetInput();
+    const float* GetInput() const;
+    f32 GetOutputSample() const;
+    void Clear();
+    void Reset();
+
+protected:
+    float* buffer{nullptr};
+    float* buffer_end{nullptr};
+    s32 max_delay{};
+    float* input{nullptr};
+    float* output{nullptr};
+    s32 delay{};
+};
+
+class DelayLineAllPass final : public DelayLineBase {
+public:
+    DelayLineAllPass();
+    ~DelayLineAllPass();
+
+    void Initialize(u32 delay, float coeffcient_, f32* src_buffer);
+    void SetCoefficient(float coeffcient_);
+    f32 Tick(f32 sample);
+    void Reset();
+
+private:
+    float coefficient{};
+};
+} // namespace AudioCore
diff --git a/src/audio_core/effect_context.cpp b/src/audio_core/effect_context.cpp
index f770b9608..89e4573c7 100644
--- a/src/audio_core/effect_context.cpp
+++ b/src/audio_core/effect_context.cpp
@@ -90,6 +90,14 @@ s32 EffectBase::GetProcessingOrder() const {
     return processing_order;
 }
 
+std::vector<u8>& EffectBase::GetWorkBuffer() {
+    return work_buffer;
+}
+
+const std::vector<u8>& EffectBase::GetWorkBuffer() const {
+    return work_buffer;
+}
+
 EffectI3dl2Reverb::EffectI3dl2Reverb() : EffectGeneric(EffectType::I3dl2Reverb) {}
 EffectI3dl2Reverb::~EffectI3dl2Reverb() = default;
 
@@ -117,6 +125,12 @@ void EffectI3dl2Reverb::Update(EffectInfo::InParams& in_params) {
         usage = UsageState::Initialized;
         params.status = ParameterStatus::Initialized;
         skipped = in_params.buffer_address == 0 || in_params.buffer_size == 0;
+        if (!skipped) {
+            auto& cur_work_buffer = GetWorkBuffer();
+            // Has two buffers internally
+            cur_work_buffer.resize(in_params.buffer_size * 2);
+            std::fill(cur_work_buffer.begin(), cur_work_buffer.end(), 0);
+        }
     }
 }
 
@@ -129,6 +143,14 @@ void EffectI3dl2Reverb::UpdateForCommandGeneration() {
     GetParams().status = ParameterStatus::Updated;
 }
 
+I3dl2ReverbState& EffectI3dl2Reverb::GetState() {
+    return state;
+}
+
+const I3dl2ReverbState& EffectI3dl2Reverb::GetState() const {
+    return state;
+}
+
 EffectBiquadFilter::EffectBiquadFilter() : EffectGeneric(EffectType::BiquadFilter) {}
 EffectBiquadFilter::~EffectBiquadFilter() = default;
 
diff --git a/src/audio_core/effect_context.h b/src/audio_core/effect_context.h
index c5e0b398c..5e0655dd7 100644
--- a/src/audio_core/effect_context.h
+++ b/src/audio_core/effect_context.h
@@ -8,6 +8,7 @@
 #include <memory>
 #include <vector>
 #include "audio_core/common.h"
+#include "audio_core/delay_line.h"
 #include "common/common_funcs.h"
 #include "common/common_types.h"
 #include "common/swap.h"
@@ -194,6 +195,8 @@ public:
     [[nodiscard]] bool IsEnabled() const;
     [[nodiscard]] s32 GetMixID() const;
     [[nodiscard]] s32 GetProcessingOrder() const;
+    [[nodiscard]] std::vector<u8>& GetWorkBuffer();
+    [[nodiscard]] const std::vector<u8>& GetWorkBuffer() const;
 
 protected:
     UsageState usage{UsageState::Invalid};
@@ -201,6 +204,7 @@ protected:
     s32 mix_id{};
     s32 processing_order{};
     bool enabled = false;
+    std::vector<u8> work_buffer{};
 };
 
 template <typename T>
@@ -212,7 +216,7 @@ public:
         return internal_params;
     }
 
-    const I3dl2ReverbParams& GetParams() const {
+    const T& GetParams() const {
         return internal_params;
     }
 
@@ -229,6 +233,27 @@ public:
     void UpdateForCommandGeneration() override;
 };
 
+struct I3dl2ReverbState {
+    f32 lowpass_0{};
+    f32 lowpass_1{};
+    f32 lowpass_2{};
+
+    DelayLineBase early_delay_line{};
+    std::array<u32, AudioCommon::I3DL2REVERB_TAPS> early_tap_steps{};
+    f32 early_gain{};
+    f32 late_gain{};
+
+    u32 early_to_late_taps{};
+    std::array<DelayLineBase, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> fdn_delay_line{};
+    std::array<DelayLineAllPass, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> decay_delay_line0{};
+    std::array<DelayLineAllPass, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> decay_delay_line1{};
+    f32 last_reverb_echo{};
+    DelayLineBase center_delay_line{};
+    std::array<std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT>, 3> lpf_coefficients{};
+    std::array<f32, AudioCommon::I3DL2REVERB_DELAY_LINE_COUNT> shelf_filter{};
+    f32 dry_gain{};
+};
+
 class EffectI3dl2Reverb : public EffectGeneric<I3dl2ReverbParams> {
 public:
     explicit EffectI3dl2Reverb();
@@ -237,8 +262,12 @@ public:
     void Update(EffectInfo::InParams& in_params) override;
     void UpdateForCommandGeneration() override;
 
+    I3dl2ReverbState& GetState();
+    const I3dl2ReverbState& GetState() const;
+
 private:
     bool skipped = false;
+    I3dl2ReverbState state{};
 };
 
 class EffectBiquadFilter : public EffectGeneric<BiquadFilterParams> {
diff --git a/src/common/CMakeLists.txt b/src/common/CMakeLists.txt
index 263c457cd..b657506b1 100644
--- a/src/common/CMakeLists.txt
+++ b/src/common/CMakeLists.txt
@@ -168,7 +168,6 @@ add_library(common STATIC
     time_zone.cpp
     time_zone.h
     tree.h
-    uint128.cpp
     uint128.h
     uuid.cpp
     uuid.h
diff --git a/src/common/cityhash.cpp b/src/common/cityhash.cpp
index 4e1d874b5..66218fc21 100644
--- a/src/common/cityhash.cpp
+++ b/src/common/cityhash.cpp
@@ -28,8 +28,10 @@
 // compromising on hash quality.
 
 #include <algorithm>
-#include <string.h> // for memcpy and memset
-#include "cityhash.h"
+#include <cstring>
+#include <utility>
+
+#include "common/cityhash.h"
 #include "common/swap.h"
 
 // #include "config.h"
@@ -42,21 +44,17 @@
 
 using namespace std;
 
-typedef uint8_t uint8;
-typedef uint32_t uint32;
-typedef uint64_t uint64;
-
 namespace Common {
 
-static uint64 UNALIGNED_LOAD64(const char* p) {
-    uint64 result;
-    memcpy(&result, p, sizeof(result));
+static u64 unaligned_load64(const char* p) {
+    u64 result;
+    std::memcpy(&result, p, sizeof(result));
     return result;
 }
 
-static uint32 UNALIGNED_LOAD32(const char* p) {
-    uint32 result;
-    memcpy(&result, p, sizeof(result));
+static u32 unaligned_load32(const char* p) {
+    u32 result;
+    std::memcpy(&result, p, sizeof(result));
     return result;
 }
 
@@ -76,64 +74,64 @@ static uint32 UNALIGNED_LOAD32(const char* p) {
 #endif
 #endif
 
-static uint64 Fetch64(const char* p) {
-    return uint64_in_expected_order(UNALIGNED_LOAD64(p));
+static u64 Fetch64(const char* p) {
+    return uint64_in_expected_order(unaligned_load64(p));
 }
 
-static uint32 Fetch32(const char* p) {
-    return uint32_in_expected_order(UNALIGNED_LOAD32(p));
+static u32 Fetch32(const char* p) {
+    return uint32_in_expected_order(unaligned_load32(p));
 }
 
 // Some primes between 2^63 and 2^64 for various uses.
-static const uint64 k0 = 0xc3a5c85c97cb3127ULL;
-static const uint64 k1 = 0xb492b66fbe98f273ULL;
-static const uint64 k2 = 0x9ae16a3b2f90404fULL;
+static constexpr u64 k0 = 0xc3a5c85c97cb3127ULL;
+static constexpr u64 k1 = 0xb492b66fbe98f273ULL;
+static constexpr u64 k2 = 0x9ae16a3b2f90404fULL;
 
 // Bitwise right rotate.  Normally this will compile to a single
 // instruction, especially if the shift is a manifest constant.
-static uint64 Rotate(uint64 val, int shift) {
+static u64 Rotate(u64 val, int shift) {
     // Avoid shifting by 64: doing so yields an undefined result.
     return shift == 0 ? val : ((val >> shift) | (val << (64 - shift)));
 }
 
-static uint64 ShiftMix(uint64 val) {
+static u64 ShiftMix(u64 val) {
     return val ^ (val >> 47);
 }
 
-static uint64 HashLen16(uint64 u, uint64 v) {
-    return Hash128to64(uint128(u, v));
+static u64 HashLen16(u64 u, u64 v) {
+    return Hash128to64(u128{u, v});
 }
 
-static uint64 HashLen16(uint64 u, uint64 v, uint64 mul) {
+static u64 HashLen16(u64 u, u64 v, u64 mul) {
     // Murmur-inspired hashing.
-    uint64 a = (u ^ v) * mul;
+    u64 a = (u ^ v) * mul;
     a ^= (a >> 47);
-    uint64 b = (v ^ a) * mul;
+    u64 b = (v ^ a) * mul;
     b ^= (b >> 47);
     b *= mul;
     return b;
 }
 
-static uint64 HashLen0to16(const char* s, std::size_t len) {
+static u64 HashLen0to16(const char* s, size_t len) {
     if (len >= 8) {
-        uint64 mul = k2 + len * 2;
-        uint64 a = Fetch64(s) + k2;
-        uint64 b = Fetch64(s + len - 8);
-        uint64 c = Rotate(b, 37) * mul + a;
-        uint64 d = (Rotate(a, 25) + b) * mul;
+        u64 mul = k2 + len * 2;
+        u64 a = Fetch64(s) + k2;
+        u64 b = Fetch64(s + len - 8);
+        u64 c = Rotate(b, 37) * mul + a;
+        u64 d = (Rotate(a, 25) + b) * mul;
         return HashLen16(c, d, mul);
     }
     if (len >= 4) {
-        uint64 mul = k2 + len * 2;
-        uint64 a = Fetch32(s);
+        u64 mul = k2 + len * 2;
+        u64 a = Fetch32(s);
         return HashLen16(len + (a << 3), Fetch32(s + len - 4), mul);
     }
     if (len > 0) {
-        uint8 a = s[0];
-        uint8 b = s[len >> 1];
-        uint8 c = s[len - 1];
-        uint32 y = static_cast<uint32>(a) + (static_cast<uint32>(b) << 8);
-        uint32 z = static_cast<uint32>(len) + (static_cast<uint32>(c) << 2);
+        u8 a = s[0];
+        u8 b = s[len >> 1];
+        u8 c = s[len - 1];
+        u32 y = static_cast<u32>(a) + (static_cast<u32>(b) << 8);
+        u32 z = static_cast<u32>(len) + (static_cast<u32>(c) << 2);
         return ShiftMix(y * k2 ^ z * k0) * k2;
     }
     return k2;
@@ -141,22 +139,21 @@ static uint64 HashLen0to16(const char* s, std::size_t len) {
 
 // This probably works well for 16-byte strings as well, but it may be overkill
 // in that case.
-static uint64 HashLen17to32(const char* s, std::size_t len) {
-    uint64 mul = k2 + len * 2;
-    uint64 a = Fetch64(s) * k1;
-    uint64 b = Fetch64(s + 8);
-    uint64 c = Fetch64(s + len - 8) * mul;
-    uint64 d = Fetch64(s + len - 16) * k2;
+static u64 HashLen17to32(const char* s, size_t len) {
+    u64 mul = k2 + len * 2;
+    u64 a = Fetch64(s) * k1;
+    u64 b = Fetch64(s + 8);
+    u64 c = Fetch64(s + len - 8) * mul;
+    u64 d = Fetch64(s + len - 16) * k2;
     return HashLen16(Rotate(a + b, 43) + Rotate(c, 30) + d, a + Rotate(b + k2, 18) + c, mul);
 }
 
 // Return a 16-byte hash for 48 bytes.  Quick and dirty.
 // Callers do best to use "random-looking" values for a and b.
-static pair<uint64, uint64> WeakHashLen32WithSeeds(uint64 w, uint64 x, uint64 y, uint64 z, uint64 a,
-                                                   uint64 b) {
+static pair<u64, u64> WeakHashLen32WithSeeds(u64 w, u64 x, u64 y, u64 z, u64 a, u64 b) {
     a += w;
     b = Rotate(b + a + z, 21);
-    uint64 c = a;
+    u64 c = a;
     a += x;
     a += y;
     b += Rotate(a, 44);
@@ -164,34 +161,34 @@ static pair<uint64, uint64> WeakHashLen32WithSeeds(uint64 w, uint64 x, uint64 y,
 }
 
 // Return a 16-byte hash for s[0] ... s[31], a, and b.  Quick and dirty.
-static pair<uint64, uint64> WeakHashLen32WithSeeds(const char* s, uint64 a, uint64 b) {
+static pair<u64, u64> WeakHashLen32WithSeeds(const char* s, u64 a, u64 b) {
     return WeakHashLen32WithSeeds(Fetch64(s), Fetch64(s + 8), Fetch64(s + 16), Fetch64(s + 24), a,
                                   b);
 }
 
 // Return an 8-byte hash for 33 to 64 bytes.
-static uint64 HashLen33to64(const char* s, std::size_t len) {
-    uint64 mul = k2 + len * 2;
-    uint64 a = Fetch64(s) * k2;
-    uint64 b = Fetch64(s + 8);
-    uint64 c = Fetch64(s + len - 24);
-    uint64 d = Fetch64(s + len - 32);
-    uint64 e = Fetch64(s + 16) * k2;
-    uint64 f = Fetch64(s + 24) * 9;
-    uint64 g = Fetch64(s + len - 8);
-    uint64 h = Fetch64(s + len - 16) * mul;
-    uint64 u = Rotate(a + g, 43) + (Rotate(b, 30) + c) * 9;
-    uint64 v = ((a + g) ^ d) + f + 1;
-    uint64 w = swap64((u + v) * mul) + h;
-    uint64 x = Rotate(e + f, 42) + c;
-    uint64 y = (swap64((v + w) * mul) + g) * mul;
-    uint64 z = e + f + c;
+static u64 HashLen33to64(const char* s, size_t len) {
+    u64 mul = k2 + len * 2;
+    u64 a = Fetch64(s) * k2;
+    u64 b = Fetch64(s + 8);
+    u64 c = Fetch64(s + len - 24);
+    u64 d = Fetch64(s + len - 32);
+    u64 e = Fetch64(s + 16) * k2;
+    u64 f = Fetch64(s + 24) * 9;
+    u64 g = Fetch64(s + len - 8);
+    u64 h = Fetch64(s + len - 16) * mul;
+    u64 u = Rotate(a + g, 43) + (Rotate(b, 30) + c) * 9;
+    u64 v = ((a + g) ^ d) + f + 1;
+    u64 w = swap64((u + v) * mul) + h;
+    u64 x = Rotate(e + f, 42) + c;
+    u64 y = (swap64((v + w) * mul) + g) * mul;
+    u64 z = e + f + c;
     a = swap64((x + z) * mul + y) + b;
     b = ShiftMix((z + a) * mul + d + h) * mul;
     return b + x;
 }
 
-uint64 CityHash64(const char* s, std::size_t len) {
+u64 CityHash64(const char* s, size_t len) {
     if (len <= 32) {
         if (len <= 16) {
             return HashLen0to16(s, len);
@@ -204,15 +201,15 @@ uint64 CityHash64(const char* s, std::size_t len) {
 
     // For strings over 64 bytes we hash the end first, and then as we
     // loop we keep 56 bytes of state: v, w, x, y, and z.
-    uint64 x = Fetch64(s + len - 40);
-    uint64 y = Fetch64(s + len - 16) + Fetch64(s + len - 56);
-    uint64 z = HashLen16(Fetch64(s + len - 48) + len, Fetch64(s + len - 24));
-    pair<uint64, uint64> v = WeakHashLen32WithSeeds(s + len - 64, len, z);
-    pair<uint64, uint64> w = WeakHashLen32WithSeeds(s + len - 32, y + k1, x);
+    u64 x = Fetch64(s + len - 40);
+    u64 y = Fetch64(s + len - 16) + Fetch64(s + len - 56);
+    u64 z = HashLen16(Fetch64(s + len - 48) + len, Fetch64(s + len - 24));
+    pair<u64, u64> v = WeakHashLen32WithSeeds(s + len - 64, len, z);
+    pair<u64, u64> w = WeakHashLen32WithSeeds(s + len - 32, y + k1, x);
     x = x * k1 + Fetch64(s);
 
     // Decrease len to the nearest multiple of 64, and operate on 64-byte chunks.
-    len = (len - 1) & ~static_cast<std::size_t>(63);
+    len = (len - 1) & ~static_cast<size_t>(63);
     do {
         x = Rotate(x + y + v.first + Fetch64(s + 8), 37) * k1;
         y = Rotate(y + v.second + Fetch64(s + 48), 42) * k1;
@@ -229,21 +226,21 @@ uint64 CityHash64(const char* s, std::size_t len) {
                      HashLen16(v.second, w.second) + x);
 }
 
-uint64 CityHash64WithSeed(const char* s, std::size_t len, uint64 seed) {
+u64 CityHash64WithSeed(const char* s, size_t len, u64 seed) {
     return CityHash64WithSeeds(s, len, k2, seed);
 }
 
-uint64 CityHash64WithSeeds(const char* s, std::size_t len, uint64 seed0, uint64 seed1) {
+u64 CityHash64WithSeeds(const char* s, size_t len, u64 seed0, u64 seed1) {
     return HashLen16(CityHash64(s, len) - seed0, seed1);
 }
 
 // A subroutine for CityHash128().  Returns a decent 128-bit hash for strings
 // of any length representable in signed long.  Based on City and Murmur.
-static uint128 CityMurmur(const char* s, std::size_t len, uint128 seed) {
-    uint64 a = Uint128Low64(seed);
-    uint64 b = Uint128High64(seed);
-    uint64 c = 0;
-    uint64 d = 0;
+static u128 CityMurmur(const char* s, size_t len, u128 seed) {
+    u64 a = seed[0];
+    u64 b = seed[1];
+    u64 c = 0;
+    u64 d = 0;
     signed long l = static_cast<long>(len) - 16;
     if (l <= 0) { // len <= 16
         a = ShiftMix(a * k1) * k1;
@@ -266,20 +263,20 @@ static uint128 CityMurmur(const char* s, std::size_t len, uint128 seed) {
     }
     a = HashLen16(a, c);
     b = HashLen16(d, b);
-    return uint128(a ^ b, HashLen16(b, a));
+    return u128{a ^ b, HashLen16(b, a)};
 }
 
-uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
+u128 CityHash128WithSeed(const char* s, size_t len, u128 seed) {
     if (len < 128) {
         return CityMurmur(s, len, seed);
     }
 
     // We expect len >= 128 to be the common case.  Keep 56 bytes of state:
     // v, w, x, y, and z.
-    pair<uint64, uint64> v, w;
-    uint64 x = Uint128Low64(seed);
-    uint64 y = Uint128High64(seed);
-    uint64 z = len * k1;
+    pair<u64, u64> v, w;
+    u64 x = seed[0];
+    u64 y = seed[1];
+    u64 z = len * k1;
     v.first = Rotate(y ^ k1, 49) * k1 + Fetch64(s);
     v.second = Rotate(v.first, 42) * k1 + Fetch64(s + 8);
     w.first = Rotate(y + z, 35) * k1 + x;
@@ -313,7 +310,7 @@ uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
     w.first *= 9;
     v.first *= k0;
     // If 0 < len < 128, hash up to 4 chunks of 32 bytes each from the end of s.
-    for (std::size_t tail_done = 0; tail_done < len;) {
+    for (size_t tail_done = 0; tail_done < len;) {
         tail_done += 32;
         y = Rotate(x + y, 42) * k0 + v.second;
         w.first += Fetch64(s + len - tail_done + 16);
@@ -328,13 +325,12 @@ uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed) {
     // different 56-byte-to-8-byte hashes to get a 16-byte final result.
     x = HashLen16(x, v.first);
     y = HashLen16(y + z, w.first);
-    return uint128(HashLen16(x + v.second, w.second) + y, HashLen16(x + w.second, y + v.second));
+    return u128{HashLen16(x + v.second, w.second) + y, HashLen16(x + w.second, y + v.second)};
 }
 
-uint128 CityHash128(const char* s, std::size_t len) {
-    return len >= 16
-               ? CityHash128WithSeed(s + 16, len - 16, uint128(Fetch64(s), Fetch64(s + 8) + k0))
-               : CityHash128WithSeed(s, len, uint128(k0, k1));
+u128 CityHash128(const char* s, size_t len) {
+    return len >= 16 ? CityHash128WithSeed(s + 16, len - 16, u128{Fetch64(s), Fetch64(s + 8) + k0})
+                     : CityHash128WithSeed(s, len, u128{k0, k1});
 }
 
 } // namespace Common
diff --git a/src/common/cityhash.h b/src/common/cityhash.h
index a00804e01..022d0f7cb 100644
--- a/src/common/cityhash.h
+++ b/src/common/cityhash.h
@@ -61,50 +61,38 @@
 
 #pragma once
 
-#include <cstddef>
-#include <cstdint>
-#include <utility>
+#include "common/common_types.h"
 
 namespace Common {
 
-using uint128 = std::pair<uint64_t, uint64_t>;
-
-[[nodiscard]] inline uint64_t Uint128Low64(const uint128& x) {
-    return x.first;
-}
-[[nodiscard]] inline uint64_t Uint128High64(const uint128& x) {
-    return x.second;
-}
-
 // Hash function for a byte array.
-[[nodiscard]] uint64_t CityHash64(const char* buf, std::size_t len);
+[[nodiscard]] u64 CityHash64(const char* buf, size_t len);
 
 // Hash function for a byte array.  For convenience, a 64-bit seed is also
 // hashed into the result.
-[[nodiscard]] uint64_t CityHash64WithSeed(const char* buf, std::size_t len, uint64_t seed);
+[[nodiscard]] u64 CityHash64WithSeed(const char* buf, size_t len, u64 seed);
 
 // Hash function for a byte array.  For convenience, two seeds are also
 // hashed into the result.
-[[nodiscard]] uint64_t CityHash64WithSeeds(const char* buf, std::size_t len, uint64_t seed0,
-                                           uint64_t seed1);
+[[nodiscard]] u64 CityHash64WithSeeds(const char* buf, size_t len, u64 seed0, u64 seed1);
 
 // Hash function for a byte array.
-[[nodiscard]] uint128 CityHash128(const char* s, std::size_t len);
+[[nodiscard]] u128 CityHash128(const char* s, size_t len);
 
 // Hash function for a byte array.  For convenience, a 128-bit seed is also
 // hashed into the result.
-[[nodiscard]] uint128 CityHash128WithSeed(const char* s, std::size_t len, uint128 seed);
+[[nodiscard]] u128 CityHash128WithSeed(const char* s, size_t len, u128 seed);
 
 // Hash 128 input bits down to 64 bits of output.
 // This is intended to be a reasonably good hash function.
-[[nodiscard]] inline uint64_t Hash128to64(const uint128& x) {
+[[nodiscard]] inline u64 Hash128to64(const u128& x) {
     // Murmur-inspired hashing.
-    const uint64_t kMul = 0x9ddfea08eb382d69ULL;
-    uint64_t a = (Uint128Low64(x) ^ Uint128High64(x)) * kMul;
+    const u64 mul = 0x9ddfea08eb382d69ULL;
+    u64 a = (x[0] ^ x[1]) * mul;
     a ^= (a >> 47);
-    uint64_t b = (Uint128High64(x) ^ a) * kMul;
+    u64 b = (x[1] ^ a) * mul;
     b ^= (b >> 47);
-    b *= kMul;
+    b *= mul;
     return b;
 }
 
diff --git a/src/common/uint128.cpp b/src/common/uint128.cpp
deleted file mode 100644
index 16bf7c828..000000000
--- a/src/common/uint128.cpp
+++ /dev/null
@@ -1,71 +0,0 @@
-// Copyright 2019 yuzu Emulator Project
-// Licensed under GPLv2 or any later version
-// Refer to the license.txt file included.
-
-#ifdef _MSC_VER
-#include <intrin.h>
-
-#pragma intrinsic(_umul128)
-#pragma intrinsic(_udiv128)
-#endif
-#include <cstring>
-#include "common/uint128.h"
-
-namespace Common {
-
-#ifdef _MSC_VER
-
-u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) {
-    u128 r{};
-    r[0] = _umul128(a, b, &r[1]);
-    u64 remainder;
-#if _MSC_VER < 1923
-    return udiv128(r[1], r[0], d, &remainder);
-#else
-    return _udiv128(r[1], r[0], d, &remainder);
-#endif
-}
-
-#else
-
-u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) {
-    const u64 diva = a / d;
-    const u64 moda = a % d;
-    const u64 divb = b / d;
-    const u64 modb = b % d;
-    return diva * b + moda * divb + moda * modb / d;
-}
-
-#endif
-
-u128 Multiply64Into128(u64 a, u64 b) {
-    u128 result;
-#ifdef _MSC_VER
-    result[0] = _umul128(a, b, &result[1]);
-#else
-    unsigned __int128 tmp = a;
-    tmp *= b;
-    std::memcpy(&result, &tmp, sizeof(u128));
-#endif
-    return result;
-}
-
-std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor) {
-    u64 remainder = dividend[0] % divisor;
-    u64 accum = dividend[0] / divisor;
-    if (dividend[1] == 0)
-        return {accum, remainder};
-    // We ignore dividend[1] / divisor as that overflows
-    const u64 first_segment = (dividend[1] % divisor) << 32;
-    accum += (first_segment / divisor) << 32;
-    const u64 second_segment = (first_segment % divisor) << 32;
-    accum += (second_segment / divisor);
-    remainder += second_segment % divisor;
-    if (remainder >= divisor) {
-        accum++;
-        remainder -= divisor;
-    }
-    return {accum, remainder};
-}
-
-} // namespace Common
diff --git a/src/common/uint128.h b/src/common/uint128.h
index 969259ab6..4780b2f9d 100644
--- a/src/common/uint128.h
+++ b/src/common/uint128.h
@@ -4,19 +4,118 @@
 
 #pragma once
 
+#include <cstring>
 #include <utility>
+
+#ifdef _MSC_VER
+#include <intrin.h>
+#pragma intrinsic(__umulh)
+#pragma intrinsic(_umul128)
+#pragma intrinsic(_udiv128)
+#else
+#include <x86intrin.h>
+#endif
+
 #include "common/common_types.h"
 
 namespace Common {
 
 // This function multiplies 2 u64 values and divides it by a u64 value.
-[[nodiscard]] u64 MultiplyAndDivide64(u64 a, u64 b, u64 d);
+[[nodiscard]] static inline u64 MultiplyAndDivide64(u64 a, u64 b, u64 d) {
+#ifdef _MSC_VER
+    u128 r{};
+    r[0] = _umul128(a, b, &r[1]);
+    u64 remainder;
+#if _MSC_VER < 1923
+    return udiv128(r[1], r[0], d, &remainder);
+#else
+    return _udiv128(r[1], r[0], d, &remainder);
+#endif
+#else
+    const u64 diva = a / d;
+    const u64 moda = a % d;
+    const u64 divb = b / d;
+    const u64 modb = b % d;
+    return diva * b + moda * divb + moda * modb / d;
+#endif
+}
 
 // This function multiplies 2 u64 values and produces a u128 value;
-[[nodiscard]] u128 Multiply64Into128(u64 a, u64 b);
+[[nodiscard]] static inline u128 Multiply64Into128(u64 a, u64 b) {
+    u128 result;
+#ifdef _MSC_VER
+    result[0] = _umul128(a, b, &result[1]);
+#else
+    unsigned __int128 tmp = a;
+    tmp *= b;
+    std::memcpy(&result, &tmp, sizeof(u128));
+#endif
+    return result;
+}
+
+[[nodiscard]] static inline u64 GetFixedPoint64Factor(u64 numerator, u64 divisor) {
+#ifdef __SIZEOF_INT128__
+    const auto base = static_cast<unsigned __int128>(numerator) << 64ULL;
+    return static_cast<u64>(base / divisor);
+#elif defined(_M_X64) || defined(_M_ARM64)
+    std::array<u64, 2> r = {0, numerator};
+    u64 remainder;
+#if _MSC_VER < 1923
+    return udiv128(r[1], r[0], divisor, &remainder);
+#else
+    return _udiv128(r[1], r[0], divisor, &remainder);
+#endif
+#else
+    // This one is bit more inaccurate.
+    return MultiplyAndDivide64(std::numeric_limits<u64>::max(), numerator, divisor);
+#endif
+}
+
+[[nodiscard]] static inline u64 MultiplyHigh(u64 a, u64 b) {
+#ifdef __SIZEOF_INT128__
+    return (static_cast<unsigned __int128>(a) * static_cast<unsigned __int128>(b)) >> 64;
+#elif defined(_M_X64) || defined(_M_ARM64)
+    return __umulh(a, b); // MSVC
+#else
+    // Generic fallback
+    const u64 a_lo = u32(a);
+    const u64 a_hi = a >> 32;
+    const u64 b_lo = u32(b);
+    const u64 b_hi = b >> 32;
+
+    const u64 a_x_b_hi = a_hi * b_hi;
+    const u64 a_x_b_mid = a_hi * b_lo;
+    const u64 b_x_a_mid = b_hi * a_lo;
+    const u64 a_x_b_lo = a_lo * b_lo;
+
+    const u64 carry_bit = (static_cast<u64>(static_cast<u32>(a_x_b_mid)) +
+                           static_cast<u64>(static_cast<u32>(b_x_a_mid)) + (a_x_b_lo >> 32)) >>
+                          32;
+
+    const u64 multhi = a_x_b_hi + (a_x_b_mid >> 32) + (b_x_a_mid >> 32) + carry_bit;
+
+    return multhi;
+#endif
+}
 
 // This function divides a u128 by a u32 value and produces two u64 values:
 // the result of division and the remainder
-[[nodiscard]] std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor);
+[[nodiscard]] static inline std::pair<u64, u64> Divide128On32(u128 dividend, u32 divisor) {
+    u64 remainder = dividend[0] % divisor;
+    u64 accum = dividend[0] / divisor;
+    if (dividend[1] == 0)
+        return {accum, remainder};
+    // We ignore dividend[1] / divisor as that overflows
+    const u64 first_segment = (dividend[1] % divisor) << 32;
+    accum += (first_segment / divisor) << 32;
+    const u64 second_segment = (first_segment % divisor) << 32;
+    accum += (second_segment / divisor);
+    remainder += second_segment % divisor;
+    if (remainder >= divisor) {
+        accum++;
+        remainder -= divisor;
+    }
+    return {accum, remainder};
+}
 
 } // namespace Common
diff --git a/src/common/wall_clock.cpp b/src/common/wall_clock.cpp
index a8c143f85..49830b8ab 100644
--- a/src/common/wall_clock.cpp
+++ b/src/common/wall_clock.cpp
@@ -2,6 +2,8 @@
 // Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
+#include <cstdint>
+
 #include "common/uint128.h"
 #include "common/wall_clock.h"
 
diff --git a/src/common/x64/native_clock.cpp b/src/common/x64/native_clock.cpp
index a65f6b832..87de40624 100644
--- a/src/common/x64/native_clock.cpp
+++ b/src/common/x64/native_clock.cpp
@@ -8,68 +8,10 @@
 #include <mutex>
 #include <thread>
 
-#ifdef _MSC_VER
-#include <intrin.h>
-
-#pragma intrinsic(__umulh)
-#pragma intrinsic(_udiv128)
-#else
-#include <x86intrin.h>
-#endif
-
 #include "common/atomic_ops.h"
 #include "common/uint128.h"
 #include "common/x64/native_clock.h"
 
-namespace {
-
-[[nodiscard]] u64 GetFixedPoint64Factor(u64 numerator, u64 divisor) {
-#ifdef __SIZEOF_INT128__
-    const auto base = static_cast<unsigned __int128>(numerator) << 64ULL;
-    return static_cast<u64>(base / divisor);
-#elif defined(_M_X64) || defined(_M_ARM64)
-    std::array<u64, 2> r = {0, numerator};
-    u64 remainder;
-#if _MSC_VER < 1923
-    return udiv128(r[1], r[0], divisor, &remainder);
-#else
-    return _udiv128(r[1], r[0], divisor, &remainder);
-#endif
-#else
-    // This one is bit more inaccurate.
-    return MultiplyAndDivide64(std::numeric_limits<u64>::max(), numerator, divisor);
-#endif
-}
-
-[[nodiscard]] u64 MultiplyHigh(u64 a, u64 b) {
-#ifdef __SIZEOF_INT128__
-    return (static_cast<unsigned __int128>(a) * static_cast<unsigned __int128>(b)) >> 64;
-#elif defined(_M_X64) || defined(_M_ARM64)
-    return __umulh(a, b); // MSVC
-#else
-    // Generic fallback
-    const u64 a_lo = u32(a);
-    const u64 a_hi = a >> 32;
-    const u64 b_lo = u32(b);
-    const u64 b_hi = b >> 32;
-
-    const u64 a_x_b_hi = a_hi * b_hi;
-    const u64 a_x_b_mid = a_hi * b_lo;
-    const u64 b_x_a_mid = b_hi * a_lo;
-    const u64 a_x_b_lo = a_lo * b_lo;
-
-    const u64 carry_bit = (static_cast<u64>(static_cast<u32>(a_x_b_mid)) +
-                           static_cast<u64>(static_cast<u32>(b_x_a_mid)) + (a_x_b_lo >> 32)) >>
-                          32;
-
-    const u64 multhi = a_x_b_hi + (a_x_b_mid >> 32) + (b_x_a_mid >> 32) + carry_bit;
-
-    return multhi;
-#endif
-}
-
-} // namespace
-
 namespace Common {
 
 u64 EstimateRDTSCFrequency() {
diff --git a/src/core/CMakeLists.txt b/src/core/CMakeLists.txt
index e74e6a668..c6bdf72ec 100644
--- a/src/core/CMakeLists.txt
+++ b/src/core/CMakeLists.txt
@@ -19,7 +19,6 @@ add_library(core STATIC
     core.h
     core_timing.cpp
     core_timing.h
-    core_timing_util.cpp
     core_timing_util.h
     cpu_manager.cpp
     cpu_manager.h
@@ -266,6 +265,7 @@ add_library(core STATIC
     hle/service/am/applets/software_keyboard.h
     hle/service/am/applets/web_browser.cpp
     hle/service/am/applets/web_browser.h
+    hle/service/am/applets/web_types.h
     hle/service/am/idle.cpp
     hle/service/am/idle.h
     hle/service/am/omm.cpp
diff --git a/src/core/core.cpp b/src/core/core.cpp
index 30f5e1128..de6305e2a 100644
--- a/src/core/core.cpp
+++ b/src/core/core.cpp
@@ -308,6 +308,9 @@ struct System::Impl {
         // Close all CPU/threading state
         cpu_manager.Shutdown();
 
+        // Release the Time Manager's resources
+        time_manager.Shutdown();
+
         // Shutdown kernel and core timing
         core_timing.Shutdown();
         kernel.Shutdown();
diff --git a/src/core/core_timing_util.cpp b/src/core/core_timing_util.cpp
deleted file mode 100644
index 8ce8e602e..000000000
--- a/src/core/core_timing_util.cpp
+++ /dev/null
@@ -1,84 +0,0 @@
-// Copyright 2008 Dolphin Emulator Project / 2017 Citra Emulator Project
-// Licensed under GPLv2+
-// Refer to the license.txt file included.
-
-#include "core/core_timing_util.h"
-
-#include <cinttypes>
-#include <limits>
-#include "common/logging/log.h"
-#include "common/uint128.h"
-#include "core/hardware_properties.h"
-
-namespace Core::Timing {
-
-constexpr u64 MAX_VALUE_TO_MULTIPLY = std::numeric_limits<s64>::max() / Hardware::BASE_CLOCK_RATE;
-
-s64 msToCycles(std::chrono::milliseconds ms) {
-    if (static_cast<u64>(ms.count() / 1000) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_ERROR(Core_Timing, "Integer overflow, use max value");
-        return std::numeric_limits<s64>::max();
-    }
-    if (static_cast<u64>(ms.count()) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_DEBUG(Core_Timing, "Time very big, do rounding");
-        return Hardware::BASE_CLOCK_RATE * (ms.count() / 1000);
-    }
-    return (Hardware::BASE_CLOCK_RATE * ms.count()) / 1000;
-}
-
-s64 usToCycles(std::chrono::microseconds us) {
-    if (static_cast<u64>(us.count() / 1000000) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_ERROR(Core_Timing, "Integer overflow, use max value");
-        return std::numeric_limits<s64>::max();
-    }
-    if (static_cast<u64>(us.count()) > MAX_VALUE_TO_MULTIPLY) {
-        LOG_DEBUG(Core_Timing, "Time very big, do rounding");
-        return Hardware::BASE_CLOCK_RATE * (us.count() / 1000000);
-    }
-    return (Hardware::BASE_CLOCK_RATE * us.count()) / 1000000;
-}
-
-s64 nsToCycles(std::chrono::nanoseconds ns) {
-    const u128 temporal = Common::Multiply64Into128(ns.count(), Hardware::BASE_CLOCK_RATE);
-    return Common::Divide128On32(temporal, static_cast<u32>(1000000000)).first;
-}
-
-u64 msToClockCycles(std::chrono::milliseconds ns) {
-    const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ);
-    return Common::Divide128On32(temp, 1000).first;
-}
-
-u64 usToClockCycles(std::chrono::microseconds ns) {
-    const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ);
-    return Common::Divide128On32(temp, 1000000).first;
-}
-
-u64 nsToClockCycles(std::chrono::nanoseconds ns) {
-    const u128 temp = Common::Multiply64Into128(ns.count(), Hardware::CNTFREQ);
-    return Common::Divide128On32(temp, 1000000000).first;
-}
-
-u64 CpuCyclesToClockCycles(u64 ticks) {
-    const u128 temporal = Common::Multiply64Into128(ticks, Hardware::CNTFREQ);
-    return Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-}
-
-std::chrono::milliseconds CyclesToMs(s64 cycles) {
-    const u128 temporal = Common::Multiply64Into128(cycles, 1000);
-    u64 ms = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-    return std::chrono::milliseconds(ms);
-}
-
-std::chrono::nanoseconds CyclesToNs(s64 cycles) {
-    const u128 temporal = Common::Multiply64Into128(cycles, 1000000000);
-    u64 ns = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-    return std::chrono::nanoseconds(ns);
-}
-
-std::chrono::microseconds CyclesToUs(s64 cycles) {
-    const u128 temporal = Common::Multiply64Into128(cycles, 1000000);
-    u64 us = Common::Divide128On32(temporal, static_cast<u32>(Hardware::BASE_CLOCK_RATE)).first;
-    return std::chrono::microseconds(us);
-}
-
-} // namespace Core::Timing
diff --git a/src/core/core_timing_util.h b/src/core/core_timing_util.h
index e4a046bf9..14c36a485 100644
--- a/src/core/core_timing_util.h
+++ b/src/core/core_timing_util.h
@@ -1,24 +1,59 @@
-// Copyright 2008 Dolphin Emulator Project / 2017 Citra Emulator Project
-// Licensed under GPLv2+
+// Copyright 2020 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
 // Refer to the license.txt file included.
 
 #pragma once
 
 #include <chrono>
+
 #include "common/common_types.h"
+#include "core/hardware_properties.h"
 
 namespace Core::Timing {
 
-s64 msToCycles(std::chrono::milliseconds ms);
-s64 usToCycles(std::chrono::microseconds us);
-s64 nsToCycles(std::chrono::nanoseconds ns);
-u64 msToClockCycles(std::chrono::milliseconds ns);
-u64 usToClockCycles(std::chrono::microseconds ns);
-u64 nsToClockCycles(std::chrono::nanoseconds ns);
-std::chrono::milliseconds CyclesToMs(s64 cycles);
-std::chrono::nanoseconds CyclesToNs(s64 cycles);
-std::chrono::microseconds CyclesToUs(s64 cycles);
-
-u64 CpuCyclesToClockCycles(u64 ticks);
+namespace detail {
+constexpr u64 CNTFREQ_ADJUSTED = Hardware::CNTFREQ / 1000;
+constexpr u64 BASE_CLOCK_RATE_ADJUSTED = Hardware::BASE_CLOCK_RATE / 1000;
+} // namespace detail
+
+[[nodiscard]] constexpr s64 msToCycles(std::chrono::milliseconds ms) {
+    return ms.count() * detail::BASE_CLOCK_RATE_ADJUSTED;
+}
+
+[[nodiscard]] constexpr s64 usToCycles(std::chrono::microseconds us) {
+    return us.count() * detail::BASE_CLOCK_RATE_ADJUSTED / 1000;
+}
+
+[[nodiscard]] constexpr s64 nsToCycles(std::chrono::nanoseconds ns) {
+    return ns.count() * detail::BASE_CLOCK_RATE_ADJUSTED / 1000000;
+}
+
+[[nodiscard]] constexpr u64 msToClockCycles(std::chrono::milliseconds ms) {
+    return static_cast<u64>(ms.count()) * detail::CNTFREQ_ADJUSTED;
+}
+
+[[nodiscard]] constexpr u64 usToClockCycles(std::chrono::microseconds us) {
+    return us.count() * detail::CNTFREQ_ADJUSTED / 1000;
+}
+
+[[nodiscard]] constexpr u64 nsToClockCycles(std::chrono::nanoseconds ns) {
+    return ns.count() * detail::CNTFREQ_ADJUSTED / 1000000;
+}
+
+[[nodiscard]] constexpr u64 CpuCyclesToClockCycles(u64 ticks) {
+    return ticks * detail::CNTFREQ_ADJUSTED / detail::BASE_CLOCK_RATE_ADJUSTED;
+}
+
+[[nodiscard]] constexpr std::chrono::milliseconds CyclesToMs(s64 cycles) {
+    return std::chrono::milliseconds(cycles / detail::BASE_CLOCK_RATE_ADJUSTED);
+}
+
+[[nodiscard]] constexpr std::chrono::nanoseconds CyclesToNs(s64 cycles) {
+    return std::chrono::nanoseconds(cycles * 1000000 / detail::BASE_CLOCK_RATE_ADJUSTED);
+}
+
+[[nodiscard]] constexpr std::chrono::microseconds CyclesToUs(s64 cycles) {
+    return std::chrono::microseconds(cycles * 1000 / detail::BASE_CLOCK_RATE_ADJUSTED);
+}
 
 } // namespace Core::Timing
diff --git a/src/core/frontend/applets/controller.h b/src/core/frontend/applets/controller.h
index dff71d8d9..b0626a0f9 100644
--- a/src/core/frontend/applets/controller.h
+++ b/src/core/frontend/applets/controller.h
@@ -31,6 +31,7 @@ struct ControllerParameters {
     bool allow_dual_joycons{};
     bool allow_left_joycon{};
     bool allow_right_joycon{};
+    bool allow_gamecube_controller{};
 };
 
 class ControllerApplet {
diff --git a/src/core/hle/kernel/kernel.cpp b/src/core/hle/kernel/kernel.cpp
index b6e6f115e..39c5182c5 100644
--- a/src/core/hle/kernel/kernel.cpp
+++ b/src/core/hle/kernel/kernel.cpp
@@ -101,8 +101,6 @@ struct KernelCore::Impl {
 
         current_process = nullptr;
 
-        system_resource_limit = nullptr;
-
         global_handle_table.Clear();
 
         preemption_event = nullptr;
@@ -111,6 +109,13 @@ struct KernelCore::Impl {
 
         exclusive_monitor.reset();
 
+        hid_shared_mem = nullptr;
+        font_shared_mem = nullptr;
+        irs_shared_mem = nullptr;
+        time_shared_mem = nullptr;
+
+        system_resource_limit = nullptr;
+
         // Next host thead ID to use, 0-3 IDs represent core threads, >3 represent others
         next_host_thread_id = Core::Hardware::NUM_CPU_CORES;
     }
diff --git a/src/core/hle/service/hid/controllers/npad.cpp b/src/core/hle/service/hid/controllers/npad.cpp
index dbf198345..70b9f3824 100644
--- a/src/core/hle/service/hid/controllers/npad.cpp
+++ b/src/core/hle/service/hid/controllers/npad.cpp
@@ -21,6 +21,7 @@
 
 namespace Service::HID {
 constexpr s32 HID_JOYSTICK_MAX = 0x7fff;
+constexpr s32 HID_TRIGGER_MAX = 0x7fff;
 [[maybe_unused]] constexpr s32 HID_JOYSTICK_MIN = -0x7fff;
 constexpr std::size_t NPAD_OFFSET = 0x9A00;
 constexpr u32 BATTERY_FULL = 2;
@@ -48,6 +49,8 @@ Controller_NPad::NPadControllerType Controller_NPad::MapSettingsTypeToNPad(
         return NPadControllerType::JoyRight;
     case Settings::ControllerType::Handheld:
         return NPadControllerType::Handheld;
+    case Settings::ControllerType::GameCube:
+        return NPadControllerType::GameCube;
     default:
         UNREACHABLE();
         return NPadControllerType::ProController;
@@ -67,6 +70,8 @@ Settings::ControllerType Controller_NPad::MapNPadToSettingsType(
         return Settings::ControllerType::RightJoycon;
     case NPadControllerType::Handheld:
         return Settings::ControllerType::Handheld;
+    case NPadControllerType::GameCube:
+        return Settings::ControllerType::GameCube;
     default:
         UNREACHABLE();
         return Settings::ControllerType::ProController;
@@ -209,6 +214,13 @@ void Controller_NPad::InitNewlyAddedController(std::size_t controller_idx) {
         controller.assignment_mode = NpadAssignments::Single;
         controller.footer_type = AppletFooterUiType::JoyRightHorizontal;
         break;
+    case NPadControllerType::GameCube:
+        controller.style_set.gamecube.Assign(1);
+        // The GC Controller behaves like a wired Pro Controller
+        controller.device_type.fullkey.Assign(1);
+        controller.system_properties.is_vertical.Assign(1);
+        controller.system_properties.use_plus.Assign(1);
+        break;
     case NPadControllerType::Pokeball:
         controller.style_set.palma.Assign(1);
         controller.device_type.palma.Assign(1);
@@ -259,6 +271,7 @@ void Controller_NPad::OnInit() {
         style.joycon_right.Assign(1);
         style.joycon_dual.Assign(1);
         style.fullkey.Assign(1);
+        style.gamecube.Assign(1);
         style.palma.Assign(1);
     }
 
@@ -339,6 +352,7 @@ void Controller_NPad::RequestPadStateUpdate(u32 npad_id) {
     auto& pad_state = npad_pad_states[controller_idx].pad_states;
     auto& lstick_entry = npad_pad_states[controller_idx].l_stick;
     auto& rstick_entry = npad_pad_states[controller_idx].r_stick;
+    auto& trigger_entry = npad_trigger_states[controller_idx];
     const auto& button_state = buttons[controller_idx];
     const auto& analog_state = sticks[controller_idx];
     const auto [stick_l_x_f, stick_l_y_f] =
@@ -404,6 +418,17 @@ void Controller_NPad::RequestPadStateUpdate(u32 npad_id) {
         pad_state.left_sl.Assign(button_state[SL - BUTTON_HID_BEGIN]->GetStatus());
         pad_state.left_sr.Assign(button_state[SR - BUTTON_HID_BEGIN]->GetStatus());
     }
+
+    if (controller_type == NPadControllerType::GameCube) {
+        trigger_entry.l_analog = static_cast<s32>(
+            button_state[ZL - BUTTON_HID_BEGIN]->GetStatus() ? HID_TRIGGER_MAX : 0);
+        trigger_entry.r_analog = static_cast<s32>(
+            button_state[ZR - BUTTON_HID_BEGIN]->GetStatus() ? HID_TRIGGER_MAX : 0);
+        pad_state.zl.Assign(false);
+        pad_state.zr.Assign(button_state[R - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.l.Assign(button_state[ZL - BUTTON_HID_BEGIN]->GetStatus());
+        pad_state.r.Assign(button_state[ZR - BUTTON_HID_BEGIN]->GetStatus());
+    }
 }
 
 void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8* data,
@@ -418,6 +443,11 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8*
             &npad.joy_left_states,  &npad.joy_right_states, &npad.palma_states,
             &npad.system_ext_states};
 
+        // There is the posibility to have more controllers with analog triggers
+        const std::array<TriggerGeneric*, 1> controller_triggers{
+            &npad.gc_trigger_states,
+        };
+
         for (auto* main_controller : controller_npads) {
             main_controller->common.entry_count = 16;
             main_controller->common.total_entry_count = 17;
@@ -435,6 +465,21 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8*
             cur_entry.timestamp2 = cur_entry.timestamp;
         }
 
+        for (auto* analog_trigger : controller_triggers) {
+            analog_trigger->entry_count = 16;
+            analog_trigger->total_entry_count = 17;
+
+            const auto& last_entry = analog_trigger->trigger[analog_trigger->last_entry_index];
+
+            analog_trigger->timestamp = core_timing.GetCPUTicks();
+            analog_trigger->last_entry_index = (analog_trigger->last_entry_index + 1) % 17;
+
+            auto& cur_entry = analog_trigger->trigger[analog_trigger->last_entry_index];
+
+            cur_entry.timestamp = last_entry.timestamp + 1;
+            cur_entry.timestamp2 = cur_entry.timestamp;
+        }
+
         const auto& controller_type = connected_controllers[i].type;
 
         if (controller_type == NPadControllerType::None || !connected_controllers[i].is_connected) {
@@ -444,6 +489,7 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8*
 
         RequestPadStateUpdate(npad_index);
         auto& pad_state = npad_pad_states[npad_index];
+        auto& trigger_state = npad_trigger_states[npad_index];
 
         auto& main_controller =
             npad.fullkey_states.npad[npad.fullkey_states.common.last_entry_index];
@@ -456,6 +502,8 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8*
         auto& pokeball_entry = npad.palma_states.npad[npad.palma_states.common.last_entry_index];
         auto& libnx_entry =
             npad.system_ext_states.npad[npad.system_ext_states.common.last_entry_index];
+        auto& trigger_entry =
+            npad.gc_trigger_states.trigger[npad.gc_trigger_states.last_entry_index];
 
         libnx_entry.connection_status.raw = 0;
         libnx_entry.connection_status.is_connected.Assign(1);
@@ -524,6 +572,18 @@ void Controller_NPad::OnUpdate(const Core::Timing::CoreTiming& core_timing, u8*
 
             libnx_entry.connection_status.is_right_connected.Assign(1);
             break;
+        case NPadControllerType::GameCube:
+            main_controller.connection_status.raw = 0;
+            main_controller.connection_status.is_connected.Assign(1);
+            main_controller.connection_status.is_wired.Assign(1);
+            main_controller.pad.pad_states.raw = pad_state.pad_states.raw;
+            main_controller.pad.l_stick = pad_state.l_stick;
+            main_controller.pad.r_stick = pad_state.r_stick;
+            trigger_entry.l_analog = trigger_state.l_analog;
+            trigger_entry.r_analog = trigger_state.r_analog;
+
+            libnx_entry.connection_status.is_wired.Assign(1);
+            break;
         case NPadControllerType::Pokeball:
             pokeball_entry.connection_status.raw = 0;
             pokeball_entry.connection_status.is_connected.Assign(1);
@@ -674,6 +734,7 @@ void Controller_NPad::OnMotionUpdate(const Core::Timing::CoreTiming& core_timing
                 right_sixaxis_entry.orientation = motion_devices[1].orientation;
             }
             break;
+        case NPadControllerType::GameCube:
         case NPadControllerType::Pokeball:
             break;
         }
@@ -1135,6 +1196,8 @@ bool Controller_NPad::IsControllerSupported(NPadControllerType controller) const
             return style.joycon_left;
         case NPadControllerType::JoyRight:
             return style.joycon_right;
+        case NPadControllerType::GameCube:
+            return style.gamecube;
         case NPadControllerType::Pokeball:
             return style.palma;
         default:
diff --git a/src/core/hle/service/hid/controllers/npad.h b/src/core/hle/service/hid/controllers/npad.h
index 48bab988c..bc2e6779d 100644
--- a/src/core/hle/service/hid/controllers/npad.h
+++ b/src/core/hle/service/hid/controllers/npad.h
@@ -51,6 +51,7 @@ public:
         JoyDual,
         JoyLeft,
         JoyRight,
+        GameCube,
         Pokeball,
     };
 
@@ -60,6 +61,7 @@ public:
         JoyconDual = 5,
         JoyconLeft = 6,
         JoyconRight = 7,
+        GameCube = 8,
         Pokeball = 9,
         MaxNpadType = 10,
     };
@@ -389,6 +391,25 @@ private:
     };
     static_assert(sizeof(SixAxisGeneric) == 0x708, "SixAxisGeneric is an invalid size");
 
+    struct TriggerState {
+        s64_le timestamp{};
+        s64_le timestamp2{};
+        s32_le l_analog{};
+        s32_le r_analog{};
+    };
+    static_assert(sizeof(TriggerState) == 0x18, "TriggerState is an invalid size");
+
+    struct TriggerGeneric {
+        INSERT_PADDING_BYTES(0x4);
+        s64_le timestamp;
+        INSERT_PADDING_BYTES(0x4);
+        s64_le total_entry_count;
+        s64_le last_entry_index;
+        s64_le entry_count;
+        std::array<TriggerState, 17> trigger{};
+    };
+    static_assert(sizeof(TriggerGeneric) == 0x1C8, "TriggerGeneric is an invalid size");
+
     struct NPadSystemProperties {
         union {
             s64_le raw{};
@@ -509,7 +530,9 @@ private:
         AppletFooterUiType footer_type;
         // nfc_states needs to be checked switchbrew does not match with HW
         NfcXcdHandle nfc_states;
-        INSERT_PADDING_BYTES(0xdef);
+        INSERT_PADDING_BYTES(0x8); // Mutex
+        TriggerGeneric gc_trigger_states;
+        INSERT_PADDING_BYTES(0xc1f);
     };
     static_assert(sizeof(NPadEntry) == 0x5000, "NPadEntry is an invalid size");
 
@@ -560,6 +583,7 @@ private:
     f32 sixaxis_fusion_parameter2{};
     bool sixaxis_at_rest{true};
     std::array<ControllerPad, 10> npad_pad_states{};
+    std::array<TriggerState, 10> npad_trigger_states{};
     bool is_in_lr_assignment_mode{false};
     Core::System& system;
 };
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
index 36970f828..ecba1dba1 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_nvdec.cpp
@@ -34,8 +34,7 @@ NvResult nvhost_nvdec::Ioctl1(Ioctl command, const std::vector<u8>& input,
         case 0xa: {
             if (command.length == 0x1c) {
                 LOG_INFO(Service_NVDRV, "NVDEC video stream ended");
-                Tegra::ChCommandHeaderList cmdlist(1);
-                cmdlist[0] = Tegra::ChCommandHeader{0xDEADB33F};
+                Tegra::ChCommandHeaderList cmdlist{{0xDEADB33F}};
                 system.GPU().PushCommandBuffer(cmdlist);
             }
             return UnmapBuffer(input, output);
diff --git a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
index 72499654c..70849a9bd 100644
--- a/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
+++ b/src/core/hle/service/nvdrv/devices/nvhost_vic.cpp
@@ -28,8 +28,13 @@ NvResult nvhost_vic::Ioctl1(Ioctl command, const std::vector<u8>& input, std::ve
             return GetWaitbase(input, output);
         case 0x9:
             return MapBuffer(input, output);
-        case 0xa:
+        case 0xa: {
+            if (command.length == 0x1c) {
+                Tegra::ChCommandHeaderList cmdlist{{0xDEADB33F}};
+                system.GPU().PushCommandBuffer(cmdlist);
+            }
             return UnmapBuffer(input, output);
+        }
         default:
             break;
         }
diff --git a/src/core/hle/service/time/time_manager.cpp b/src/core/hle/service/time/time_manager.cpp
index 858623e2b..1f7309f6b 100644
--- a/src/core/hle/service/time/time_manager.cpp
+++ b/src/core/hle/service/time/time_manager.cpp
@@ -279,6 +279,10 @@ const SharedMemory& TimeManager::GetSharedMemory() const {
     return impl->shared_memory;
 }
 
+void TimeManager::Shutdown() {
+    impl.reset();
+}
+
 void TimeManager::UpdateLocalSystemClockTime(s64 posix_time) {
     impl->UpdateLocalSystemClockTime(system, posix_time);
 }
diff --git a/src/core/hle/service/time/time_manager.h b/src/core/hle/service/time/time_manager.h
index 993c7c288..4db8cc0e1 100644
--- a/src/core/hle/service/time/time_manager.h
+++ b/src/core/hle/service/time/time_manager.h
@@ -61,6 +61,8 @@ public:
 
     const SharedMemory& GetSharedMemory() const;
 
+    void Shutdown();
+
     void SetupTimeZoneManager(std::string location_name,
                               Clock::SteadyClockTimePoint time_zone_updated_time_point,
                               std::size_t total_location_name_count, u128 time_zone_rule_version,
diff --git a/src/input_common/mouse/mouse_input.cpp b/src/input_common/mouse/mouse_input.cpp
index 67a584d53..b864d26f2 100644
--- a/src/input_common/mouse/mouse_input.cpp
+++ b/src/input_common/mouse/mouse_input.cpp
@@ -33,11 +33,16 @@ void Mouse::UpdateThread() {
             info.motion.UpdateOrientation(update_time * 1000);
             info.tilt_speed = 0;
             info.data.motion = info.motion.GetMotion();
+            if (Settings::values.mouse_panning) {
+                info.last_mouse_change *= 0.96f;
+                info.data.axis = {static_cast<int>(16 * info.last_mouse_change.x),
+                                  static_cast<int>(16 * -info.last_mouse_change.y)};
+            }
         }
         if (configuring) {
             UpdateYuzuSettings();
         }
-        if (mouse_panning_timout++ > 8) {
+        if (mouse_panning_timout++ > 20) {
             StopPanning();
         }
         std::this_thread::sleep_for(std::chrono::milliseconds(update_time));
@@ -82,16 +87,27 @@ void Mouse::StopPanning() {
 void Mouse::MouseMove(int x, int y, int center_x, int center_y) {
     for (MouseInfo& info : mouse_info) {
         if (Settings::values.mouse_panning) {
-            const auto mouse_change = Common::MakeVec(x, y) - Common::MakeVec(center_x, center_y);
+            auto mouse_change =
+                (Common::MakeVec(x, y) - Common::MakeVec(center_x, center_y)).Cast<float>();
             mouse_panning_timout = 0;
 
             if (mouse_change.y == 0 && mouse_change.x == 0) {
                 continue;
             }
+            const auto mouse_change_length = mouse_change.Length();
+            if (mouse_change_length < 3.0f) {
+                mouse_change /= mouse_change_length / 3.0f;
+            }
+
+            info.last_mouse_change = (info.last_mouse_change * 0.91f) + (mouse_change * 0.09f);
+
+            const auto last_mouse_change_length = info.last_mouse_change.Length();
+            if (last_mouse_change_length > 8.0f) {
+                info.last_mouse_change /= last_mouse_change_length / 8.0f;
+            } else if (last_mouse_change_length < 1.0f) {
+                info.last_mouse_change = mouse_change / mouse_change.Length();
+            }
 
-            info.last_mouse_change = (info.last_mouse_change * 0.8f) + (mouse_change * 0.2f);
-            info.data.axis = {static_cast<int>(16 * info.last_mouse_change.x),
-                              static_cast<int>(16 * -info.last_mouse_change.y)};
             info.tilt_direction = info.last_mouse_change;
             info.tilt_speed = info.tilt_direction.Normalize() * info.sensitivity;
             continue;
diff --git a/src/input_common/settings.h b/src/input_common/settings.h
index 75486554b..a59f5d461 100644
--- a/src/input_common/settings.h
+++ b/src/input_common/settings.h
@@ -340,6 +340,7 @@ enum class ControllerType {
     LeftJoycon,
     RightJoycon,
     Handheld,
+    GameCube,
 };
 
 struct PlayerInput {
diff --git a/src/tests/CMakeLists.txt b/src/tests/CMakeLists.txt
index 6a5c18945..4ea0076e9 100644
--- a/src/tests/CMakeLists.txt
+++ b/src/tests/CMakeLists.txt
@@ -1,5 +1,6 @@
 add_executable(tests
     common/bit_field.cpp
+    common/cityhash.cpp
     common/fibers.cpp
     common/param_package.cpp
     common/ring_buffer.cpp
diff --git a/src/tests/common/cityhash.cpp b/src/tests/common/cityhash.cpp
new file mode 100644
index 000000000..7a40b6c4a
--- /dev/null
+++ b/src/tests/common/cityhash.cpp
@@ -0,0 +1,22 @@
+// Copyright 2021 yuzu Emulator Project
+// Licensed under GPLv2 or any later version
+// Refer to the license.txt file included.
+
+#include <catch2/catch.hpp>
+
+#include "common/cityhash.h"
+
+constexpr char msg[] = "The blue frogs are singing under the crimson sky.\n"
+                       "It is time to run, Robert.";
+
+using namespace Common;
+
+TEST_CASE("CityHash", "[common]") {
+    // These test results were built against a known good version.
+    REQUIRE(CityHash64(msg, sizeof(msg)) == 0x92d5c2e9cbfbbc01);
+    REQUIRE(CityHash64WithSeed(msg, sizeof(msg), 0xdead) == 0xbfbe93f21a2820dd);
+    REQUIRE(CityHash64WithSeeds(msg, sizeof(msg), 0xbeef, 0xcafe) == 0xb343317955fc8a06);
+    REQUIRE(CityHash128(msg, sizeof(msg)) == u128{0x98e60d0423747eaa, 0xd8694c5b6fcaede9});
+    REQUIRE(CityHash128WithSeed(msg, sizeof(msg), {0xdead, 0xbeef}) ==
+            u128{0xf0307dba81199ebe, 0xd77764e0c4a9eb74});
+}
diff --git a/src/video_core/cdma_pusher.cpp b/src/video_core/cdma_pusher.cpp
index 33b3c060b..a3fda1094 100644
--- a/src/video_core/cdma_pusher.cpp
+++ b/src/video_core/cdma_pusher.cpp
@@ -37,59 +37,43 @@ CDmaPusher::CDmaPusher(GPU& gpu_)
 
 CDmaPusher::~CDmaPusher() = default;
 
-void CDmaPusher::Push(ChCommandHeaderList&& entries) {
-    cdma_queue.push(std::move(entries));
-}
-
-void CDmaPusher::DispatchCalls() {
-    while (!cdma_queue.empty()) {
-        Step();
-    }
-}
-
-void CDmaPusher::Step() {
-    const auto entries{cdma_queue.front()};
-    cdma_queue.pop();
-
-    std::vector<u32> values(entries.size());
-    std::memcpy(values.data(), entries.data(), entries.size() * sizeof(u32));
-
-    for (const u32 value : values) {
+void CDmaPusher::ProcessEntries(ChCommandHeaderList&& entries) {
+    for (const auto& value : entries) {
         if (mask != 0) {
             const auto lbs = static_cast<u32>(std::countr_zero(mask));
             mask &= ~(1U << lbs);
-            ExecuteCommand(static_cast<u32>(offset + lbs), value);
+            ExecuteCommand(offset + lbs, value.raw);
             continue;
         } else if (count != 0) {
             --count;
-            ExecuteCommand(static_cast<u32>(offset), value);
+            ExecuteCommand(offset, value.raw);
             if (incrementing) {
                 ++offset;
             }
             continue;
         }
-        const auto mode = static_cast<ChSubmissionMode>((value >> 28) & 0xf);
+        const auto mode = value.submission_mode.Value();
         switch (mode) {
         case ChSubmissionMode::SetClass: {
-            mask = value & 0x3f;
-            offset = (value >> 16) & 0xfff;
-            current_class = static_cast<ChClassId>((value >> 6) & 0x3ff);
+            mask = value.value & 0x3f;
+            offset = value.method_offset;
+            current_class = static_cast<ChClassId>((value.value >> 6) & 0x3ff);
             break;
         }
         case ChSubmissionMode::Incrementing:
         case ChSubmissionMode::NonIncrementing:
-            count = value & 0xffff;
-            offset = (value >> 16) & 0xfff;
+            count = value.value;
+            offset = value.method_offset;
             incrementing = mode == ChSubmissionMode::Incrementing;
             break;
         case ChSubmissionMode::Mask:
-            mask = value & 0xffff;
-            offset = (value >> 16) & 0xfff;
+            mask = value.value;
+            offset = value.method_offset;
             break;
         case ChSubmissionMode::Immediate: {
-            const u32 data = value & 0xfff;
-            offset = (value >> 16) & 0xfff;
-            ExecuteCommand(static_cast<u32>(offset), data);
+            const u32 data = value.value & 0xfff;
+            offset = value.method_offset;
+            ExecuteCommand(offset, data);
             break;
         }
         default:
@@ -102,8 +86,8 @@ void CDmaPusher::Step() {
 void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) {
     switch (current_class) {
     case ChClassId::NvDec:
-        ThiStateWrite(nvdec_thi_state, state_offset, {data});
-        switch (static_cast<ThiMethod>(state_offset)) {
+        ThiStateWrite(nvdec_thi_state, offset, data);
+        switch (static_cast<ThiMethod>(offset)) {
         case ThiMethod::IncSyncpt: {
             LOG_DEBUG(Service_NVDRV, "NVDEC Class IncSyncpt Method");
             const auto syncpoint_id = static_cast<u32>(data & 0xFF);
@@ -120,7 +104,7 @@ void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) {
             LOG_DEBUG(Service_NVDRV, "NVDEC method 0x{:X}",
                       static_cast<u32>(nvdec_thi_state.method_0));
             nvdec_processor->ProcessMethod(static_cast<Nvdec::Method>(nvdec_thi_state.method_0),
-                                           {data});
+                                           data);
             break;
         default:
             break;
@@ -144,7 +128,7 @@ void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) {
         case ThiMethod::SetMethod1:
             LOG_DEBUG(Service_NVDRV, "VIC method 0x{:X}, Args=({})",
                       static_cast<u32>(vic_thi_state.method_0), data);
-            vic_processor->ProcessMethod(static_cast<Vic::Method>(vic_thi_state.method_0), {data});
+            vic_processor->ProcessMethod(static_cast<Vic::Method>(vic_thi_state.method_0), data);
             break;
         default:
             break;
@@ -153,7 +137,7 @@ void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) {
     case ChClassId::Host1x:
         // This device is mainly for syncpoint synchronization
         LOG_DEBUG(Service_NVDRV, "Host1X Class Method");
-        host1x_processor->ProcessMethod(static_cast<Host1x::Method>(state_offset), {data});
+        host1x_processor->ProcessMethod(static_cast<Host1x::Method>(offset), data);
         break;
     default:
         UNIMPLEMENTED_MSG("Current class not implemented {:X}", static_cast<u32>(current_class));
@@ -161,10 +145,9 @@ void CDmaPusher::ExecuteCommand(u32 state_offset, u32 data) {
     }
 }
 
-void CDmaPusher::ThiStateWrite(ThiRegisters& state, u32 state_offset,
-                               const std::vector<u32>& arguments) {
-    u8* const state_offset_ptr = reinterpret_cast<u8*>(&state) + sizeof(u32) * state_offset;
-    std::memcpy(state_offset_ptr, arguments.data(), sizeof(u32) * arguments.size());
+void CDmaPusher::ThiStateWrite(ThiRegisters& state, u32 state_offset, u32 argument) {
+    u8* const offset_ptr = reinterpret_cast<u8*>(&state) + sizeof(u32) * state_offset;
+    std::memcpy(offset_ptr, &argument, sizeof(u32));
 }
 
 } // namespace Tegra
diff --git a/src/video_core/cdma_pusher.h b/src/video_core/cdma_pusher.h
index e5f212c1a..1bada44dd 100644
--- a/src/video_core/cdma_pusher.h
+++ b/src/video_core/cdma_pusher.h
@@ -5,9 +5,7 @@
 #pragma once
 
 #include <memory>
-#include <unordered_map>
 #include <vector>
-#include <queue>
 
 #include "common/bit_field.h"
 #include "common/common_types.h"
@@ -16,9 +14,9 @@
 namespace Tegra {
 
 class GPU;
+class Host1x;
 class Nvdec;
 class Vic;
-class Host1x;
 
 enum class ChSubmissionMode : u32 {
     SetClass = 0,
@@ -48,16 +46,10 @@ enum class ChClassId : u32 {
     NvDec = 0xf0
 };
 
-enum class ChMethod : u32 {
-    Empty = 0,
-    SetMethod = 0x10,
-    SetData = 0x11,
-};
-
 union ChCommandHeader {
     u32 raw;
     BitField<0, 16, u32> value;
-    BitField<16, 12, ChMethod> method_offset;
+    BitField<16, 12, u32> method_offset;
     BitField<28, 4, ChSubmissionMode> submission_mode;
 };
 static_assert(sizeof(ChCommandHeader) == sizeof(u32), "ChCommand header is an invalid size");
@@ -99,21 +91,15 @@ public:
     explicit CDmaPusher(GPU& gpu_);
     ~CDmaPusher();
 
-    /// Push NVDEC command buffer entries into queue
-    void Push(ChCommandHeaderList&& entries);
-
-    /// Process queued command buffer entries
-    void DispatchCalls();
-
-    /// Process one queue element
-    void Step();
+    /// Process the command entry
+    void ProcessEntries(ChCommandHeaderList&& entries);
 
+private:
     /// Invoke command class devices to execute the command based on the current state
     void ExecuteCommand(u32 state_offset, u32 data);
 
-private:
     /// Write arguments value to the ThiRegisters member at the specified offset
-    void ThiStateWrite(ThiRegisters& state, u32 state_offset, const std::vector<u32>& arguments);
+    void ThiStateWrite(ThiRegisters& state, u32 offset, u32 argument);
 
     GPU& gpu;
     std::shared_ptr<Tegra::Nvdec> nvdec_processor;
@@ -124,13 +110,10 @@ private:
     ThiRegisters vic_thi_state{};
     ThiRegisters nvdec_thi_state{};
 
-    s32 count{};
-    s32 offset{};
+    u32 count{};
+    u32 offset{};
     u32 mask{};
     bool incrementing{};
-
-    // Queue of command lists to be processed
-    std::queue<ChCommandHeaderList> cdma_queue;
 };
 
 } // namespace Tegra
diff --git a/src/video_core/command_classes/codecs/codec.cpp b/src/video_core/command_classes/codecs/codec.cpp
index 39bc923a5..d02dc6260 100644
--- a/src/video_core/command_classes/codecs/codec.cpp
+++ b/src/video_core/command_classes/codecs/codec.cpp
@@ -44,8 +44,10 @@ Codec::~Codec() {
 }
 
 void Codec::SetTargetCodec(NvdecCommon::VideoCodec codec) {
-    LOG_INFO(Service_NVDRV, "NVDEC video codec initialized to {}", codec);
-    current_codec = codec;
+    if (current_codec != codec) {
+        LOG_INFO(Service_NVDRV, "NVDEC video codec initialized to {}", static_cast<u32>(codec));
+        current_codec = codec;
+    }
 }
 
 void Codec::StateWrite(u32 offset, u64 arguments) {
@@ -55,7 +57,6 @@ void Codec::StateWrite(u32 offset, u64 arguments) {
 
 void Codec::Decode() {
     bool is_first_frame = false;
-
     if (!initialized) {
         if (current_codec == NvdecCommon::VideoCodec::H264) {
             av_codec = avcodec_find_decoder(AV_CODEC_ID_H264);
diff --git a/src/video_core/command_classes/nvdec.cpp b/src/video_core/command_classes/nvdec.cpp
index 79e1f4e13..e4f919afd 100644
--- a/src/video_core/command_classes/nvdec.cpp
+++ b/src/video_core/command_classes/nvdec.cpp
@@ -12,16 +12,16 @@ Nvdec::Nvdec(GPU& gpu_) : gpu(gpu_), codec(std::make_unique<Codec>(gpu)) {}
 
 Nvdec::~Nvdec() = default;
 
-void Nvdec::ProcessMethod(Method method, const std::vector<u32>& arguments) {
+void Nvdec::ProcessMethod(Method method, u32 argument) {
     if (method == Method::SetVideoCodec) {
-        codec->StateWrite(static_cast<u32>(method), arguments[0]);
+        codec->StateWrite(static_cast<u32>(method), argument);
     } else {
-        codec->StateWrite(static_cast<u32>(method), static_cast<u64>(arguments[0]) << 8);
+        codec->StateWrite(static_cast<u32>(method), static_cast<u64>(argument) << 8);
     }
 
     switch (method) {
     case Method::SetVideoCodec:
-        codec->SetTargetCodec(static_cast<NvdecCommon::VideoCodec>(arguments[0]));
+        codec->SetTargetCodec(static_cast<NvdecCommon::VideoCodec>(argument));
         break;
     case Method::Execute:
         Execute();
diff --git a/src/video_core/command_classes/nvdec.h b/src/video_core/command_classes/nvdec.h
index e4877c533..e66be80b8 100644
--- a/src/video_core/command_classes/nvdec.h
+++ b/src/video_core/command_classes/nvdec.h
@@ -23,7 +23,7 @@ public:
     ~Nvdec();
 
     /// Writes the method into the state, Invoke Execute() if encountered
-    void ProcessMethod(Method method, const std::vector<u32>& arguments);
+    void ProcessMethod(Method method, u32 argument);
 
     /// Return most recently decoded frame
     [[nodiscard]] AVFramePtr GetFrame();
diff --git a/src/video_core/command_classes/vic.cpp b/src/video_core/command_classes/vic.cpp
index 2b7569335..0a8b82f2b 100644
--- a/src/video_core/command_classes/vic.cpp
+++ b/src/video_core/command_classes/vic.cpp
@@ -18,18 +18,14 @@ extern "C" {
 namespace Tegra {
 
 Vic::Vic(GPU& gpu_, std::shared_ptr<Nvdec> nvdec_processor_)
-    : gpu(gpu_), nvdec_processor(std::move(nvdec_processor_)) {}
-Vic::~Vic() = default;
+    : gpu(gpu_),
+      nvdec_processor(std::move(nvdec_processor_)), converted_frame_buffer{nullptr, av_free} {}
 
-void Vic::VicStateWrite(u32 offset, u32 arguments) {
-    u8* const state_offset = reinterpret_cast<u8*>(&vic_state) + offset * sizeof(u32);
-    std::memcpy(state_offset, &arguments, sizeof(u32));
-}
+Vic::~Vic() = default;
 
-void Vic::ProcessMethod(Method method, const std::vector<u32>& arguments) {
-    LOG_DEBUG(HW_GPU, "Vic method 0x{:X}", method);
-    VicStateWrite(static_cast<u32>(method), arguments[0]);
-    const u64 arg = static_cast<u64>(arguments[0]) << 8;
+void Vic::ProcessMethod(Method method, u32 argument) {
+    LOG_DEBUG(HW_GPU, "Vic method 0x{:X}", static_cast<u32>(method));
+    const u64 arg = static_cast<u64>(argument) << 8;
     switch (method) {
     case Method::Execute:
         Execute();
@@ -53,8 +49,7 @@ void Vic::ProcessMethod(Method method, const std::vector<u32>& arguments) {
 
 void Vic::Execute() {
     if (output_surface_luma_address == 0) {
-        LOG_ERROR(Service_NVDRV, "VIC Luma address not set. Received 0x{:X}",
-                  vic_state.output_surface.luma_offset);
+        LOG_ERROR(Service_NVDRV, "VIC Luma address not set.");
         return;
     }
     const VicConfig config{gpu.MemoryManager().Read<u64>(config_struct_address + 0x20)};
@@ -89,8 +84,10 @@ void Vic::Execute() {
         // Get Converted frame
         const std::size_t linear_size = frame->width * frame->height * 4;
 
-        using AVMallocPtr = std::unique_ptr<u8, decltype(&av_free)>;
-        AVMallocPtr converted_frame_buffer{static_cast<u8*>(av_malloc(linear_size)), av_free};
+        // Only allocate frame_buffer once per stream, as the size is not expected to change
+        if (!converted_frame_buffer) {
+            converted_frame_buffer = AVMallocPtr{static_cast<u8*>(av_malloc(linear_size)), av_free};
+        }
 
         const int converted_stride{frame->width * 4};
         u8* const converted_frame_buf_addr{converted_frame_buffer.get()};
@@ -104,12 +101,12 @@ void Vic::Execute() {
             const u32 block_height = static_cast<u32>(config.block_linear_height_log2);
             const auto size = Tegra::Texture::CalculateSize(true, 4, frame->width, frame->height, 1,
                                                             block_height, 0);
-            std::vector<u8> swizzled_data(size);
+            luma_buffer.resize(size);
             Tegra::Texture::SwizzleSubrect(frame->width, frame->height, frame->width * 4,
-                                           frame->width, 4, swizzled_data.data(),
+                                           frame->width, 4, luma_buffer.data(),
                                            converted_frame_buffer.get(), block_height, 0, 0);
 
-            gpu.MemoryManager().WriteBlock(output_surface_luma_address, swizzled_data.data(), size);
+            gpu.MemoryManager().WriteBlock(output_surface_luma_address, luma_buffer.data(), size);
         } else {
             // send pitch linear frame
             gpu.MemoryManager().WriteBlock(output_surface_luma_address, converted_frame_buf_addr,
@@ -132,15 +129,15 @@ void Vic::Execute() {
         const auto stride = frame->linesize[0];
         const auto half_stride = frame->linesize[1];
 
-        std::vector<u8> luma_buffer(aligned_width * surface_height);
-        std::vector<u8> chroma_buffer(aligned_width * half_height);
+        luma_buffer.resize(aligned_width * surface_height);
+        chroma_buffer.resize(aligned_width * half_height);
 
         // Populate luma buffer
         for (std::size_t y = 0; y < surface_height - 1; ++y) {
-            std::size_t src = y * stride;
-            std::size_t dst = y * aligned_width;
+            const std::size_t src = y * stride;
+            const std::size_t dst = y * aligned_width;
 
-            std::size_t size = surface_width;
+            const std::size_t size = surface_width;
 
             for (std::size_t offset = 0; offset < size; ++offset) {
                 luma_buffer[dst + offset] = luma_ptr[src + offset];
@@ -151,8 +148,8 @@ void Vic::Execute() {
 
         // Populate chroma buffer from both channels with interleaving.
         for (std::size_t y = 0; y < half_height; ++y) {
-            std::size_t src = y * half_stride;
-            std::size_t dst = y * aligned_width;
+            const std::size_t src = y * half_stride;
+            const std::size_t dst = y * aligned_width;
 
             for (std::size_t x = 0; x < half_width; ++x) {
                 chroma_buffer[dst + x * 2] = chroma_b_ptr[src + x];
diff --git a/src/video_core/command_classes/vic.h b/src/video_core/command_classes/vic.h
index 8c4e284a1..f5a2ed100 100644
--- a/src/video_core/command_classes/vic.h
+++ b/src/video_core/command_classes/vic.h
@@ -15,43 +15,6 @@ namespace Tegra {
 class GPU;
 class Nvdec;
 
-struct PlaneOffsets {
-    u32 luma_offset{};
-    u32 chroma_u_offset{};
-    u32 chroma_v_offset{};
-};
-
-struct VicRegisters {
-    INSERT_PADDING_WORDS(64);
-    u32 nop{};
-    INSERT_PADDING_WORDS(15);
-    u32 pm_trigger{};
-    INSERT_PADDING_WORDS(47);
-    u32 set_application_id{};
-    u32 set_watchdog_timer{};
-    INSERT_PADDING_WORDS(17);
-    u32 context_save_area{};
-    u32 context_switch{};
-    INSERT_PADDING_WORDS(43);
-    u32 execute{};
-    INSERT_PADDING_WORDS(63);
-    std::array<std::array<PlaneOffsets, 8>, 8> surfacex_slots{};
-    u32 picture_index{};
-    u32 control_params{};
-    u32 config_struct_offset{};
-    u32 filter_struct_offset{};
-    u32 palette_offset{};
-    u32 hist_offset{};
-    u32 context_id{};
-    u32 fce_ucode_size{};
-    PlaneOffsets output_surface{};
-    u32 fce_ucode_offset{};
-    INSERT_PADDING_WORDS(4);
-    std::array<u32, 8> slot_context_id{};
-    INSERT_PADDING_WORDS(16);
-};
-static_assert(sizeof(VicRegisters) == 0x7A0, "VicRegisters is an invalid size");
-
 class Vic {
 public:
     enum class Method : u32 {
@@ -67,14 +30,11 @@ public:
     ~Vic();
 
     /// Write to the device state.
-    void ProcessMethod(Method method, const std::vector<u32>& arguments);
+    void ProcessMethod(Method method, u32 argument);
 
 private:
     void Execute();
 
-    void VicStateWrite(u32 offset, u32 arguments);
-    VicRegisters vic_state{};
-
     enum class VideoPixelFormat : u64_le {
         RGBA8 = 0x1f,
         BGRA8 = 0x20,
@@ -88,8 +48,6 @@ private:
         BitField<9, 2, u64_le> chroma_loc_vert;
         BitField<11, 4, u64_le> block_linear_kind;
         BitField<15, 4, u64_le> block_linear_height_log2;
-        BitField<19, 3, u64_le> reserved0;
-        BitField<22, 10, u64_le> reserved1;
         BitField<32, 14, u64_le> surface_width_minus1;
         BitField<46, 14, u64_le> surface_height_minus1;
     };
@@ -97,6 +55,13 @@ private:
     GPU& gpu;
     std::shared_ptr<Tegra::Nvdec> nvdec_processor;
 
+    /// Avoid reallocation of the following buffers every frame, as their
+    /// size does not change during a stream
+    using AVMallocPtr = std::unique_ptr<u8, decltype(&av_free)>;
+    AVMallocPtr converted_frame_buffer;
+    std::vector<u8> luma_buffer;
+    std::vector<u8> chroma_buffer;
+
     GPUVAddr config_struct_address{};
     GPUVAddr output_surface_luma_address{};
     GPUVAddr output_surface_chroma_u_address{};
diff --git a/src/video_core/gpu.cpp b/src/video_core/gpu.cpp
index 2a9bd4121..51c63af4a 100644
--- a/src/video_core/gpu.cpp
+++ b/src/video_core/gpu.cpp
@@ -30,8 +30,7 @@ MICROPROFILE_DEFINE(GPU_wait, "GPU", "Wait for the GPU", MP_RGB(128, 128, 192));
 
 GPU::GPU(Core::System& system_, bool is_async_, bool use_nvdec_)
     : system{system_}, memory_manager{std::make_unique<Tegra::MemoryManager>(system)},
-      dma_pusher{std::make_unique<Tegra::DmaPusher>(system, *this)},
-      cdma_pusher{std::make_unique<Tegra::CDmaPusher>(*this)}, use_nvdec{use_nvdec_},
+      dma_pusher{std::make_unique<Tegra::DmaPusher>(system, *this)}, use_nvdec{use_nvdec_},
       maxwell_3d{std::make_unique<Engines::Maxwell3D>(system, *memory_manager)},
       fermi_2d{std::make_unique<Engines::Fermi2D>()},
       kepler_compute{std::make_unique<Engines::KeplerCompute>(system, *memory_manager)},
@@ -494,8 +493,7 @@ void GPU::PushCommandBuffer(Tegra::ChCommandHeaderList& entries) {
     // TODO(ameerj): RE proper async nvdec operation
     // gpu_thread.SubmitCommandBuffer(std::move(entries));
 
-    cdma_pusher->Push(std::move(entries));
-    cdma_pusher->DispatchCalls();
+    cdma_pusher->ProcessEntries(std::move(entries));
 }
 
 void GPU::SwapBuffers(const Tegra::FramebufferConfig* framebuffer) {
diff --git a/src/video_core/gpu_thread.cpp b/src/video_core/gpu_thread.cpp
index 50319f1d5..eb0e43c0c 100644
--- a/src/video_core/gpu_thread.cpp
+++ b/src/video_core/gpu_thread.cpp
@@ -48,8 +48,7 @@ static void RunThread(Core::System& system, VideoCore::RendererBase& renderer,
             dma_pusher.DispatchCalls();
         } else if (auto* command_list = std::get_if<SubmitChCommandEntries>(&next.data)) {
             // NVDEC
-            cdma_pusher.Push(std::move(command_list->entries));
-            cdma_pusher.DispatchCalls();
+            cdma_pusher.ProcessEntries(std::move(command_list->entries));
         } else if (const auto* data = std::get_if<SwapBuffersCommand>(&next.data)) {
             renderer.SwapBuffers(data->framebuffer ? &*data->framebuffer : nullptr);
         } else if (std::holds_alternative<OnCommandListEndCommand>(next.data)) {
diff --git a/src/video_core/renderer_opengl/gl_shader_cache.cpp b/src/video_core/renderer_opengl/gl_shader_cache.cpp
index 529570ff0..5cf7cd151 100644
--- a/src/video_core/renderer_opengl/gl_shader_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_shader_cache.cpp
@@ -335,6 +335,10 @@ void ShaderCacheOpenGL::LoadDiskCache(u64 title_id, const std::atomic_bool& stop
                                       const VideoCore::DiskResourceLoadCallback& callback) {
     disk_cache.BindTitleID(title_id);
     const std::optional transferable = disk_cache.LoadTransferable();
+
+    LOG_INFO(Render_OpenGL, "Total Shader Count: {}",
+             transferable.has_value() ? transferable->size() : 0);
+
     if (!transferable) {
         return;
     }
diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp
index 5be6dabd9..362278f01 100644
--- a/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp
+++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.cpp
@@ -12,14 +12,15 @@
 #include "common/cityhash.h"
 #include "common/common_types.h"
 #include "video_core/renderer_vulkan/fixed_pipeline_state.h"
+#include "video_core/renderer_vulkan/vk_state_tracker.h"
 
 namespace Vulkan {
 
 namespace {
 
-constexpr std::size_t POINT = 0;
-constexpr std::size_t LINE = 1;
-constexpr std::size_t POLYGON = 2;
+constexpr size_t POINT = 0;
+constexpr size_t LINE = 1;
+constexpr size_t POLYGON = 2;
 constexpr std::array POLYGON_OFFSET_ENABLE_LUT = {
     POINT,   // Points
     LINE,    // Lines
@@ -40,10 +41,14 @@ constexpr std::array POLYGON_OFFSET_ENABLE_LUT = {
 
 } // Anonymous namespace
 
-void FixedPipelineState::Fill(const Maxwell& regs, bool has_extended_dynamic_state) {
-    const std::array enabled_lut = {regs.polygon_offset_point_enable,
-                                    regs.polygon_offset_line_enable,
-                                    regs.polygon_offset_fill_enable};
+void FixedPipelineState::Refresh(Tegra::Engines::Maxwell3D& maxwell3d,
+                                 bool has_extended_dynamic_state) {
+    const Maxwell& regs = maxwell3d.regs;
+    const std::array enabled_lut{
+        regs.polygon_offset_point_enable,
+        regs.polygon_offset_line_enable,
+        regs.polygon_offset_fill_enable,
+    };
     const u32 topology_index = static_cast<u32>(regs.draw.topology.Value());
 
     raw1 = 0;
@@ -64,45 +69,53 @@ void FixedPipelineState::Fill(const Maxwell& regs, bool has_extended_dynamic_sta
 
     raw2 = 0;
     const auto test_func =
-        regs.alpha_test_enabled == 1 ? regs.alpha_test_func : Maxwell::ComparisonOp::Always;
+        regs.alpha_test_enabled != 0 ? regs.alpha_test_func : Maxwell::ComparisonOp::Always;
     alpha_test_func.Assign(PackComparisonOp(test_func));
     early_z.Assign(regs.force_early_fragment_tests != 0 ? 1 : 0);
 
     alpha_test_ref = Common::BitCast<u32>(regs.alpha_test_ref);
     point_size = Common::BitCast<u32>(regs.point_size);
 
-    for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) {
-        binding_divisors[index] =
-            regs.instanced_arrays.IsInstancingEnabled(index) ? regs.vertex_array[index].divisor : 0;
+    if (maxwell3d.dirty.flags[Dirty::InstanceDivisors]) {
+        maxwell3d.dirty.flags[Dirty::InstanceDivisors] = false;
+        for (size_t index = 0; index < Maxwell::NumVertexArrays; ++index) {
+            const bool is_enabled = regs.instanced_arrays.IsInstancingEnabled(index);
+            binding_divisors[index] = is_enabled ? regs.vertex_array[index].divisor : 0;
+        }
     }
-
-    for (size_t index = 0; index < Maxwell::NumVertexAttributes; ++index) {
-        const auto& input = regs.vertex_attrib_format[index];
-        auto& attribute = attributes[index];
-        attribute.raw = 0;
-        attribute.enabled.Assign(input.IsConstant() ? 0 : 1);
-        attribute.buffer.Assign(input.buffer);
-        attribute.offset.Assign(input.offset);
-        attribute.type.Assign(static_cast<u32>(input.type.Value()));
-        attribute.size.Assign(static_cast<u32>(input.size.Value()));
-        attribute.binding_index_enabled.Assign(regs.vertex_array[index].IsEnabled() ? 1 : 0);
+    if (maxwell3d.dirty.flags[Dirty::VertexAttributes]) {
+        maxwell3d.dirty.flags[Dirty::VertexAttributes] = false;
+        for (size_t index = 0; index < Maxwell::NumVertexAttributes; ++index) {
+            const auto& input = regs.vertex_attrib_format[index];
+            auto& attribute = attributes[index];
+            attribute.raw = 0;
+            attribute.enabled.Assign(input.IsConstant() ? 0 : 1);
+            attribute.buffer.Assign(input.buffer);
+            attribute.offset.Assign(input.offset);
+            attribute.type.Assign(static_cast<u32>(input.type.Value()));
+            attribute.size.Assign(static_cast<u32>(input.size.Value()));
+        }
     }
-
-    for (std::size_t index = 0; index < std::size(attachments); ++index) {
-        attachments[index].Fill(regs, index);
+    if (maxwell3d.dirty.flags[Dirty::Blending]) {
+        maxwell3d.dirty.flags[Dirty::Blending] = false;
+        for (size_t index = 0; index < attachments.size(); ++index) {
+            attachments[index].Refresh(regs, index);
+        }
+    }
+    if (maxwell3d.dirty.flags[Dirty::ViewportSwizzles]) {
+        maxwell3d.dirty.flags[Dirty::ViewportSwizzles] = false;
+        const auto& transform = regs.viewport_transform;
+        std::ranges::transform(transform, viewport_swizzles.begin(), [](const auto& viewport) {
+            return static_cast<u16>(viewport.swizzle.raw);
+        });
     }
-
-    const auto& transform = regs.viewport_transform;
-    std::transform(transform.begin(), transform.end(), viewport_swizzles.begin(),
-                   [](const auto& viewport) { return static_cast<u16>(viewport.swizzle.raw); });
-
     if (!has_extended_dynamic_state) {
         no_extended_dynamic_state.Assign(1);
-        dynamic_state.Fill(regs);
+        dynamic_state.Refresh(regs);
     }
 }
 
-void FixedPipelineState::BlendingAttachment::Fill(const Maxwell& regs, std::size_t index) {
+void FixedPipelineState::BlendingAttachment::Refresh(const Maxwell& regs, size_t index) {
     const auto& mask = regs.color_mask[regs.color_mask_common ? 0 : index];
 
     raw = 0;
@@ -141,7 +154,7 @@ void FixedPipelineState::BlendingAttachment::Fill(const Maxwell& regs, std::size
     enable.Assign(1);
 }
 
-void FixedPipelineState::DynamicState::Fill(const Maxwell& regs) {
+void FixedPipelineState::DynamicState::Refresh(const Maxwell& regs) {
     u32 packed_front_face = PackFrontFace(regs.front_face);
     if (regs.screen_y_control.triangle_rast_flip != 0) {
         // Flip front face
@@ -178,9 +191,9 @@ void FixedPipelineState::DynamicState::Fill(const Maxwell& regs) {
     });
 }
 
-std::size_t FixedPipelineState::Hash() const noexcept {
+size_t FixedPipelineState::Hash() const noexcept {
     const u64 hash = Common::CityHash64(reinterpret_cast<const char*>(this), Size());
-    return static_cast<std::size_t>(hash);
+    return static_cast<size_t>(hash);
 }
 
 bool FixedPipelineState::operator==(const FixedPipelineState& rhs) const noexcept {
diff --git a/src/video_core/renderer_vulkan/fixed_pipeline_state.h b/src/video_core/renderer_vulkan/fixed_pipeline_state.h
index 465a55fdb..a0eb83a68 100644
--- a/src/video_core/renderer_vulkan/fixed_pipeline_state.h
+++ b/src/video_core/renderer_vulkan/fixed_pipeline_state.h
@@ -58,7 +58,7 @@ struct FixedPipelineState {
             BitField<30, 1, u32> enable;
         };
 
-        void Fill(const Maxwell& regs, std::size_t index);
+        void Refresh(const Maxwell& regs, size_t index);
 
         constexpr std::array<bool, 4> Mask() const noexcept {
             return {mask_r != 0, mask_g != 0, mask_b != 0, mask_a != 0};
@@ -96,8 +96,6 @@ struct FixedPipelineState {
         BitField<6, 14, u32> offset;
         BitField<20, 3, u32> type;
         BitField<23, 6, u32> size;
-        // Not really an element of a vertex attribute, but it can be packed here
-        BitField<29, 1, u32> binding_index_enabled;
 
         constexpr Maxwell::VertexAttribute::Type Type() const noexcept {
             return static_cast<Maxwell::VertexAttribute::Type>(type.Value());
@@ -108,7 +106,7 @@ struct FixedPipelineState {
         }
     };
 
-    template <std::size_t Position>
+    template <size_t Position>
     union StencilFace {
         BitField<Position + 0, 3, u32> action_stencil_fail;
         BitField<Position + 3, 3, u32> action_depth_fail;
@@ -152,7 +150,7 @@ struct FixedPipelineState {
         // Vertex stride is a 12 bits value, we have 4 bits to spare per element
         std::array<u16, Maxwell::NumVertexArrays> vertex_strides;
 
-        void Fill(const Maxwell& regs);
+        void Refresh(const Maxwell& regs);
 
         Maxwell::ComparisonOp DepthTestFunc() const noexcept {
             return UnpackComparisonOp(depth_test_func);
@@ -199,9 +197,9 @@ struct FixedPipelineState {
     std::array<u16, Maxwell::NumViewports> viewport_swizzles;
     DynamicState dynamic_state;
 
-    void Fill(const Maxwell& regs, bool has_extended_dynamic_state);
+    void Refresh(Tegra::Engines::Maxwell3D& maxwell3d, bool has_extended_dynamic_state);
 
-    std::size_t Hash() const noexcept;
+    size_t Hash() const noexcept;
 
     bool operator==(const FixedPipelineState& rhs) const noexcept;
 
@@ -209,8 +207,8 @@ struct FixedPipelineState {
         return !operator==(rhs);
     }
 
-    std::size_t Size() const noexcept {
-        const std::size_t total_size = sizeof *this;
+    size_t Size() const noexcept {
+        const size_t total_size = sizeof *this;
         return total_size - (no_extended_dynamic_state != 0 ? 0 : sizeof(DynamicState));
     }
 };
@@ -224,7 +222,7 @@ namespace std {
 
 template <>
 struct hash<Vulkan::FixedPipelineState> {
-    std::size_t operator()(const Vulkan::FixedPipelineState& k) const noexcept {
+    size_t operator()(const Vulkan::FixedPipelineState& k) const noexcept {
         return k.Hash();
     }
 };
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
index 848eedd66..668633e7b 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.cpp
@@ -201,10 +201,6 @@ void BufferCacheRuntime::BindTransformFeedbackBuffer(u32 index, VkBuffer buffer,
     });
 }
 
-void BufferCacheRuntime::BindBuffer(VkBuffer buffer, u32 offset, u32 size) {
-    update_descriptor_queue.AddBuffer(buffer, offset, size);
-}
-
 void BufferCacheRuntime::ReserveQuadArrayLUT(u32 num_indices, bool wait_for_idle) {
     if (num_indices <= current_num_indices) {
         return;
diff --git a/src/video_core/renderer_vulkan/vk_buffer_cache.h b/src/video_core/renderer_vulkan/vk_buffer_cache.h
index 041e6515c..982e92191 100644
--- a/src/video_core/renderer_vulkan/vk_buffer_cache.h
+++ b/src/video_core/renderer_vulkan/vk_buffer_cache.h
@@ -8,6 +8,7 @@
 #include "video_core/engines/maxwell_3d.h"
 #include "video_core/renderer_vulkan/vk_compute_pass.h"
 #include "video_core/renderer_vulkan/vk_staging_buffer_pool.h"
+#include "video_core/renderer_vulkan/vk_update_descriptor.h"
 #include "video_core/vulkan_common/vulkan_memory_allocator.h"
 #include "video_core/vulkan_common/vulkan_wrapper.h"
 
@@ -16,7 +17,6 @@ namespace Vulkan {
 class Device;
 class VKDescriptorPool;
 class VKScheduler;
-class VKUpdateDescriptorQueue;
 
 class BufferCacheRuntime;
 
@@ -86,7 +86,9 @@ public:
     }
 
 private:
-    void BindBuffer(VkBuffer buffer, u32 offset, u32 size);
+    void BindBuffer(VkBuffer buffer, u32 offset, u32 size) {
+        update_descriptor_queue.AddBuffer(buffer, offset, size);
+    }
 
     void ReserveQuadArrayLUT(u32 num_indices, bool wait_for_idle);
 
diff --git a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp
index d50dca604..fc6dd83eb 100644
--- a/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp
+++ b/src/video_core/renderer_vulkan/vk_graphics_pipeline.cpp
@@ -221,9 +221,6 @@ vk::Pipeline VKGraphicsPipeline::CreatePipeline(const SPIRVProgram& program,
     std::vector<VkVertexInputBindingDescription> vertex_bindings;
     std::vector<VkVertexInputBindingDivisorDescriptionEXT> vertex_binding_divisors;
     for (std::size_t index = 0; index < Maxwell::NumVertexArrays; ++index) {
-        if (state.attributes[index].binding_index_enabled == 0) {
-            continue;
-        }
         const bool instanced = state.binding_divisors[index] != 0;
         const auto rate = instanced ? VK_VERTEX_INPUT_RATE_INSTANCE : VK_VERTEX_INPUT_RATE_VERTEX;
         vertex_bindings.push_back({
diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.cpp b/src/video_core/renderer_vulkan/vk_rasterizer.cpp
index 684d4e3a6..dfd38f575 100644
--- a/src/video_core/renderer_vulkan/vk_rasterizer.cpp
+++ b/src/video_core/renderer_vulkan/vk_rasterizer.cpp
@@ -267,8 +267,7 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) {
 
     query_cache.UpdateCounters();
 
-    GraphicsPipelineCacheKey key;
-    key.fixed_state.Fill(maxwell3d.regs, device.IsExtExtendedDynamicStateSupported());
+    graphics_key.fixed_state.Refresh(maxwell3d, device.IsExtExtendedDynamicStateSupported());
 
     std::scoped_lock lock{buffer_cache.mutex, texture_cache.mutex};
 
@@ -276,14 +275,15 @@ void RasterizerVulkan::Draw(bool is_indexed, bool is_instanced) {
     texture_cache.UpdateRenderTargets(false);
 
     const auto shaders = pipeline_cache.GetShaders();
-    key.shaders = GetShaderAddresses(shaders);
+    graphics_key.shaders = GetShaderAddresses(shaders);
+
     SetupShaderDescriptors(shaders, is_indexed);
 
     const Framebuffer* const framebuffer = texture_cache.GetFramebuffer();
-    key.renderpass = framebuffer->RenderPass();
+    graphics_key.renderpass = framebuffer->RenderPass();
 
-    auto* const pipeline =
-        pipeline_cache.GetGraphicsPipeline(key, framebuffer->NumColorBuffers(), async_shaders);
+    VKGraphicsPipeline* const pipeline = pipeline_cache.GetGraphicsPipeline(
+        graphics_key, framebuffer->NumColorBuffers(), async_shaders);
     if (pipeline == nullptr || pipeline->GetHandle() == VK_NULL_HANDLE) {
         // Async graphics pipeline was not ready.
         return;
diff --git a/src/video_core/renderer_vulkan/vk_rasterizer.h b/src/video_core/renderer_vulkan/vk_rasterizer.h
index 7fc6741da..acea1ba2d 100644
--- a/src/video_core/renderer_vulkan/vk_rasterizer.h
+++ b/src/video_core/renderer_vulkan/vk_rasterizer.h
@@ -20,6 +20,7 @@
 #include "video_core/renderer_vulkan/vk_buffer_cache.h"
 #include "video_core/renderer_vulkan/vk_descriptor_pool.h"
 #include "video_core/renderer_vulkan/vk_fence_manager.h"
+#include "video_core/renderer_vulkan/vk_graphics_pipeline.h"
 #include "video_core/renderer_vulkan/vk_pipeline_cache.h"
 #include "video_core/renderer_vulkan/vk_query_cache.h"
 #include "video_core/renderer_vulkan/vk_scheduler.h"
@@ -173,6 +174,8 @@ private:
     VKUpdateDescriptorQueue update_descriptor_queue;
     BlitImageHelper blit_image;
 
+    GraphicsPipelineCacheKey graphics_key;
+
     TextureCacheRuntime texture_cache_runtime;
     TextureCache texture_cache;
     BufferCacheRuntime buffer_cache_runtime;
diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.cpp b/src/video_core/renderer_vulkan/vk_state_tracker.cpp
index e81fad007..956f86845 100644
--- a/src/video_core/renderer_vulkan/vk_state_tracker.cpp
+++ b/src/video_core/renderer_vulkan/vk_state_tracker.cpp
@@ -18,9 +18,7 @@
 #define NUM(field_name) (sizeof(Maxwell3D::Regs::field_name) / (sizeof(u32)))
 
 namespace Vulkan {
-
 namespace {
-
 using namespace Dirty;
 using namespace VideoCommon::Dirty;
 using Tegra::Engines::Maxwell3D;
@@ -128,6 +126,34 @@ void SetupDirtyStencilTestEnable(Tables& tables) {
     tables[0][OFF(stencil_enable)] = StencilTestEnable;
 }
 
+void SetupDirtyBlending(Tables& tables) {
+    tables[0][OFF(color_mask_common)] = Blending;
+    tables[0][OFF(independent_blend_enable)] = Blending;
+    FillBlock(tables[0], OFF(color_mask), NUM(color_mask), Blending);
+    FillBlock(tables[0], OFF(blend), NUM(blend), Blending);
+    FillBlock(tables[0], OFF(independent_blend), NUM(independent_blend), Blending);
+}
+
+void SetupDirtyInstanceDivisors(Tables& tables) {
+    static constexpr size_t divisor_offset = 3;
+    for (size_t index = 0; index < Regs::NumVertexArrays; ++index) {
+        tables[0][OFF(instanced_arrays) + index] = InstanceDivisors;
+        tables[0][OFF(vertex_array) + index * NUM(vertex_array[0]) + divisor_offset] =
+            InstanceDivisors;
+    }
+}
+
+void SetupDirtyVertexAttributes(Tables& tables) {
+    FillBlock(tables[0], OFF(vertex_attrib_format), NUM(vertex_attrib_format), VertexAttributes);
+}
+
+void SetupDirtyViewportSwizzles(Tables& tables) {
+    static constexpr size_t swizzle_offset = 6;
+    for (size_t index = 0; index < Regs::NumViewports; ++index) {
+        tables[0][OFF(viewport_transform) + index * NUM(viewport_transform[0]) + swizzle_offset] =
+            ViewportSwizzles;
+    }
+}
 } // Anonymous namespace
 
 StateTracker::StateTracker(Tegra::GPU& gpu)
@@ -148,6 +174,10 @@ StateTracker::StateTracker(Tegra::GPU& gpu)
     SetupDirtyFrontFace(tables);
     SetupDirtyStencilOp(tables);
     SetupDirtyStencilTestEnable(tables);
+    SetupDirtyBlending(tables);
+    SetupDirtyInstanceDivisors(tables);
+    SetupDirtyVertexAttributes(tables);
+    SetupDirtyViewportSwizzles(tables);
 }
 
 } // namespace Vulkan
diff --git a/src/video_core/renderer_vulkan/vk_state_tracker.h b/src/video_core/renderer_vulkan/vk_state_tracker.h
index c335d2bdf..84e918a71 100644
--- a/src/video_core/renderer_vulkan/vk_state_tracker.h
+++ b/src/video_core/renderer_vulkan/vk_state_tracker.h
@@ -35,6 +35,11 @@ enum : u8 {
     StencilOp,
     StencilTestEnable,
 
+    Blending,
+    InstanceDivisors,
+    VertexAttributes,
+    ViewportSwizzles,
+
     Last
 };
 static_assert(Last <= std::numeric_limits<u8>::max());
diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp
index f99273c6a..dc45fdcb1 100644
--- a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp
+++ b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp
@@ -20,20 +20,20 @@ VKUpdateDescriptorQueue::VKUpdateDescriptorQueue(const Device& device_, VKSchedu
 VKUpdateDescriptorQueue::~VKUpdateDescriptorQueue() = default;
 
 void VKUpdateDescriptorQueue::TickFrame() {
-    payload.clear();
+    payload_cursor = payload.data();
 }
 
 void VKUpdateDescriptorQueue::Acquire() {
     // Minimum number of entries required.
     // This is the maximum number of entries a single draw call migth use.
-    static constexpr std::size_t MIN_ENTRIES = 0x400;
+    static constexpr size_t MIN_ENTRIES = 0x400;
 
-    if (payload.size() + MIN_ENTRIES >= payload.max_size()) {
+    if (std::distance(payload.data(), payload_cursor) + MIN_ENTRIES >= payload.max_size()) {
         LOG_WARNING(Render_Vulkan, "Payload overflow, waiting for worker thread");
         scheduler.WaitWorker();
-        payload.clear();
+        payload_cursor = payload.data();
     }
-    upload_start = &*payload.end();
+    upload_start = payload_cursor;
 }
 
 void VKUpdateDescriptorQueue::Send(VkDescriptorUpdateTemplateKHR update_template,
diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.h b/src/video_core/renderer_vulkan/vk_update_descriptor.h
index e214f7195..d35e77c44 100644
--- a/src/video_core/renderer_vulkan/vk_update_descriptor.h
+++ b/src/video_core/renderer_vulkan/vk_update_descriptor.h
@@ -4,8 +4,7 @@
 
 #pragma once
 
-#include <variant>
-#include <boost/container/static_vector.hpp>
+#include <array>
 
 #include "common/common_types.h"
 #include "video_core/vulkan_common/vulkan_wrapper.h"
@@ -16,13 +15,15 @@ class Device;
 class VKScheduler;
 
 struct DescriptorUpdateEntry {
-    DescriptorUpdateEntry(VkDescriptorImageInfo image_) : image{image_} {}
+    struct Empty {};
 
+    DescriptorUpdateEntry() = default;
+    DescriptorUpdateEntry(VkDescriptorImageInfo image_) : image{image_} {}
     DescriptorUpdateEntry(VkDescriptorBufferInfo buffer_) : buffer{buffer_} {}
-
     DescriptorUpdateEntry(VkBufferView texel_buffer_) : texel_buffer{texel_buffer_} {}
 
     union {
+        Empty empty{};
         VkDescriptorImageInfo image;
         VkDescriptorBufferInfo buffer;
         VkBufferView texel_buffer;
@@ -41,39 +42,40 @@ public:
     void Send(VkDescriptorUpdateTemplateKHR update_template, VkDescriptorSet set);
 
     void AddSampledImage(VkImageView image_view, VkSampler sampler) {
-        payload.emplace_back(VkDescriptorImageInfo{
+        *(payload_cursor++) = VkDescriptorImageInfo{
             .sampler = sampler,
             .imageView = image_view,
             .imageLayout = VK_IMAGE_LAYOUT_GENERAL,
-        });
+        };
     }
 
     void AddImage(VkImageView image_view) {
-        payload.emplace_back(VkDescriptorImageInfo{
+        *(payload_cursor++) = VkDescriptorImageInfo{
             .sampler = VK_NULL_HANDLE,
             .imageView = image_view,
             .imageLayout = VK_IMAGE_LAYOUT_GENERAL,
-        });
+        };
     }
 
-    void AddBuffer(VkBuffer buffer, u64 offset, size_t size) {
-        payload.emplace_back(VkDescriptorBufferInfo{
+    void AddBuffer(VkBuffer buffer, VkDeviceSize offset, VkDeviceSize size) {
+        *(payload_cursor++) = VkDescriptorBufferInfo{
             .buffer = buffer,
             .offset = offset,
             .range = size,
-        });
+        };
     }
 
     void AddTexelBuffer(VkBufferView texel_buffer) {
-        payload.emplace_back(texel_buffer);
+        *(payload_cursor++) = texel_buffer;
     }
 
 private:
     const Device& device;
     VKScheduler& scheduler;
 
+    DescriptorUpdateEntry* payload_cursor = nullptr;
     const DescriptorUpdateEntry* upload_start = nullptr;
-    boost::container::static_vector<DescriptorUpdateEntry, 0x10000> payload;
+    std::array<DescriptorUpdateEntry, 0x10000> payload;
 };
 
 } // namespace Vulkan
diff --git a/src/yuzu/applets/controller.cpp b/src/yuzu/applets/controller.cpp
index c680fd2c2..b92cd6886 100644
--- a/src/yuzu/applets/controller.cpp
+++ b/src/yuzu/applets/controller.cpp
@@ -67,6 +67,8 @@ bool IsControllerCompatible(Settings::ControllerType controller_type,
         return parameters.allow_right_joycon;
     case Settings::ControllerType::Handheld:
         return parameters.enable_single_mode && parameters.allow_handheld;
+    case Settings::ControllerType::GameCube:
+        return parameters.allow_gamecube_controller;
     default:
         return false;
     }
@@ -370,7 +372,7 @@ void QtControllerSelectorDialog::SetSupportedControllers() {
             QStringLiteral("image: url(:/controller/applet_joycon_right%0_disabled); ").arg(theme));
     }
 
-    if (parameters.allow_pro_controller) {
+    if (parameters.allow_pro_controller || parameters.allow_gamecube_controller) {
         ui->controllerSupported5->setStyleSheet(
             QStringLiteral("image: url(:/controller/applet_pro_controller%0); ").arg(theme));
     } else {
@@ -420,6 +422,10 @@ void QtControllerSelectorDialog::SetEmulatedControllers(std::size_t player_index
                            Settings::ControllerType::Handheld);
         emulated_controllers[player_index]->addItem(tr("Handheld"));
     }
+
+    pairs.emplace_back(emulated_controllers[player_index]->count(),
+                       Settings::ControllerType::GameCube);
+    emulated_controllers[player_index]->addItem(tr("GameCube Controller"));
 }
 
 Settings::ControllerType QtControllerSelectorDialog::GetControllerTypeFromIndex(
@@ -461,6 +467,7 @@ void QtControllerSelectorDialog::UpdateControllerIcon(std::size_t player_index)
         switch (GetControllerTypeFromIndex(emulated_controllers[player_index]->currentIndex(),
                                            player_index)) {
         case Settings::ControllerType::ProController:
+        case Settings::ControllerType::GameCube:
             return QStringLiteral("image: url(:/controller/applet_pro_controller%0); ");
         case Settings::ControllerType::DualJoyconDetached:
             return QStringLiteral("image: url(:/controller/applet_dual_joycon%0); ");
diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp
index 0635d13d0..3d6f64300 100644
--- a/src/yuzu/configuration/config.cpp
+++ b/src/yuzu/configuration/config.cpp
@@ -614,12 +614,6 @@ void Config::ReadDataStorageValues() {
                                 QString::fromStdString(FS::GetUserPath(FS::UserPath::DumpDir)))
                         .toString()
                         .toStdString());
-    FS::GetUserPath(FS::UserPath::CacheDir,
-                    qt_config
-                        ->value(QStringLiteral("cache_directory"),
-                                QString::fromStdString(FS::GetUserPath(FS::UserPath::CacheDir)))
-                        .toString()
-                        .toStdString());
     Settings::values.gamecard_inserted =
         ReadSetting(QStringLiteral("gamecard_inserted"), false).toBool();
     Settings::values.gamecard_current_game =
@@ -1218,9 +1212,6 @@ void Config::SaveDataStorageValues() {
     WriteSetting(QStringLiteral("dump_directory"),
                  QString::fromStdString(FS::GetUserPath(FS::UserPath::DumpDir)),
                  QString::fromStdString(FS::GetUserPath(FS::UserPath::DumpDir)));
-    WriteSetting(QStringLiteral("cache_directory"),
-                 QString::fromStdString(FS::GetUserPath(FS::UserPath::CacheDir)),
-                 QString::fromStdString(FS::GetUserPath(FS::UserPath::CacheDir)));
     WriteSetting(QStringLiteral("gamecard_inserted"), Settings::values.gamecard_inserted, false);
     WriteSetting(QStringLiteral("gamecard_current_game"), Settings::values.gamecard_current_game,
                  false);
diff --git a/src/yuzu/configuration/configure_filesystem.cpp b/src/yuzu/configuration/configure_filesystem.cpp
index 7ab4a80f7..bde2d4620 100644
--- a/src/yuzu/configuration/configure_filesystem.cpp
+++ b/src/yuzu/configuration/configure_filesystem.cpp
@@ -26,8 +26,6 @@ ConfigureFilesystem::ConfigureFilesystem(QWidget* parent)
             [this] { SetDirectory(DirectoryTarget::Dump, ui->dump_path_edit); });
     connect(ui->load_path_button, &QToolButton::pressed, this,
             [this] { SetDirectory(DirectoryTarget::Load, ui->load_path_edit); });
-    connect(ui->cache_directory_button, &QToolButton::pressed, this,
-            [this] { SetDirectory(DirectoryTarget::Cache, ui->cache_directory_edit); });
 
     connect(ui->reset_game_list_cache, &QPushButton::pressed, this,
             &ConfigureFilesystem::ResetMetadata);
@@ -50,8 +48,6 @@ void ConfigureFilesystem::setConfiguration() {
         QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::DumpDir)));
     ui->load_path_edit->setText(
         QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::LoadDir)));
-    ui->cache_directory_edit->setText(
-        QString::fromStdString(Common::FS::GetUserPath(Common::FS::UserPath::CacheDir)));
 
     ui->gamecard_inserted->setChecked(Settings::values.gamecard_inserted);
     ui->gamecard_current_game->setChecked(Settings::values.gamecard_current_game);
@@ -72,9 +68,6 @@ void ConfigureFilesystem::applyConfiguration() {
                             ui->dump_path_edit->text().toStdString());
     Common::FS::GetUserPath(Common::FS::UserPath::LoadDir,
                             ui->load_path_edit->text().toStdString());
-    Common::FS::GetUserPath(Common::FS::UserPath::CacheDir,
-                            ui->cache_directory_edit->text().toStdString());
-    Settings::values.gamecard_path = ui->gamecard_path_edit->text().toStdString();
 
     Settings::values.gamecard_inserted = ui->gamecard_inserted->isChecked();
     Settings::values.gamecard_current_game = ui->gamecard_current_game->isChecked();
@@ -103,9 +96,6 @@ void ConfigureFilesystem::SetDirectory(DirectoryTarget target, QLineEdit* edit)
     case DirectoryTarget::Load:
         caption = tr("Select Mod Load Directory...");
         break;
-    case DirectoryTarget::Cache:
-        caption = tr("Select Cache Directory...");
-        break;
     }
 
     QString str;
diff --git a/src/yuzu/configuration/configure_filesystem.h b/src/yuzu/configuration/configure_filesystem.h
index a79303760..2147cd405 100644
--- a/src/yuzu/configuration/configure_filesystem.h
+++ b/src/yuzu/configuration/configure_filesystem.h
@@ -32,7 +32,6 @@ private:
         Gamecard,
         Dump,
         Load,
-        Cache,
     };
 
     void SetDirectory(DirectoryTarget target, QLineEdit* edit);
diff --git a/src/yuzu/configuration/configure_filesystem.ui b/src/yuzu/configuration/configure_filesystem.ui
index 84bea0600..62b9abc7a 100644
--- a/src/yuzu/configuration/configure_filesystem.ui
+++ b/src/yuzu/configuration/configure_filesystem.ui
@@ -198,40 +198,7 @@
         <string>Caching</string>
        </property>
        <layout class="QGridLayout" name="gridLayout_5">
-        <item row="0" column="0">
-         <widget class="QLabel" name="label_10">
-          <property name="text">
-           <string>Cache Directory</string>
-          </property>
-         </widget>
-        </item>
-        <item row="0" column="1">
-         <spacer name="horizontalSpacer_3">
-          <property name="orientation">
-           <enum>Qt::Horizontal</enum>
-          </property>
-          <property name="sizeType">
-           <enum>QSizePolicy::Fixed</enum>
-          </property>
-          <property name="sizeHint" stdset="0">
-           <size>
-            <width>40</width>
-            <height>20</height>
-           </size>
-          </property>
-         </spacer>
-        </item>
-        <item row="0" column="2">
-         <widget class="QLineEdit" name="cache_directory_edit"/>
-        </item>
-        <item row="0" column="3">
-         <widget class="QToolButton" name="cache_directory_button">
-          <property name="text">
-           <string>...</string>
-          </property>
-         </widget>
-        </item>
-        <item row="1" column="0" colspan="4">
+        <item row="0" column="0" colspan="2">
          <layout class="QHBoxLayout" name="horizontalLayout_2">
           <item>
            <widget class="QCheckBox" name="cache_game_list">
diff --git a/src/yuzu/configuration/configure_input_player.cpp b/src/yuzu/configuration/configure_input_player.cpp
index c9d19c948..21d0d3449 100644
--- a/src/yuzu/configuration/configure_input_player.cpp
+++ b/src/yuzu/configuration/configure_input_player.cpp
@@ -467,10 +467,14 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
 
     UpdateControllerIcon();
     UpdateControllerAvailableButtons();
+    UpdateControllerEnabledButtons();
+    UpdateControllerButtonNames();
     UpdateMotionButtons();
     connect(ui->comboControllerType, qOverload<int>(&QComboBox::currentIndexChanged), [this](int) {
         UpdateControllerIcon();
         UpdateControllerAvailableButtons();
+        UpdateControllerEnabledButtons();
+        UpdateControllerButtonNames();
         UpdateMotionButtons();
     });
 
@@ -558,9 +562,6 @@ ConfigureInputPlayer::ConfigureInputPlayer(QWidget* parent, std::size_t player_i
             &ConfigureInputPlayer::SaveProfile);
 
     LoadConfiguration();
-
-    // TODO(wwylele): enable this when we actually emulate it
-    ui->buttonHome->setEnabled(false);
     ui->controllerFrame->SetPlayerInput(player_index, buttons_param, analogs_param);
     ui->controllerFrame->SetConnectedStatus(ui->groupConnectedController->isChecked());
 }
@@ -924,6 +925,12 @@ void ConfigureInputPlayer::SetConnectableControllers() {
                                                      Settings::ControllerType::Handheld);
             ui->comboControllerType->addItem(tr("Handheld"));
         }
+
+        if (enable_all || npad_style_set.gamecube == 1) {
+            index_controller_type_pairs.emplace_back(ui->comboControllerType->count(),
+                                                     Settings::ControllerType::GameCube);
+            ui->comboControllerType->addItem(tr("GameCube Controller"));
+        }
     };
 
     Core::System& system{Core::System::GetInstance()};
@@ -1014,7 +1021,7 @@ void ConfigureInputPlayer::UpdateControllerAvailableButtons() {
 
     // List of all the widgets that will be hidden by any of the following layouts that need
     // "unhidden" after the controller type changes
-    const std::array<QWidget*, 9> layout_show = {
+    const std::array<QWidget*, 11> layout_show = {
         ui->buttonShoulderButtonsSLSR,
         ui->horizontalSpacerShoulderButtonsWidget,
         ui->horizontalSpacerShoulderButtonsWidget2,
@@ -1024,6 +1031,8 @@ void ConfigureInputPlayer::UpdateControllerAvailableButtons() {
         ui->buttonShoulderButtonsRight,
         ui->buttonMiscButtonsPlusHome,
         ui->bottomRight,
+        ui->buttonMiscButtonsMinusGroup,
+        ui->buttonMiscButtonsScreenshotGroup,
     };
 
     for (auto* widget : layout_show) {
@@ -1056,6 +1065,14 @@ void ConfigureInputPlayer::UpdateControllerAvailableButtons() {
             ui->bottomLeft,
         };
         break;
+    case Settings::ControllerType::GameCube:
+        layout_hidden = {
+            ui->buttonShoulderButtonsSLSR,
+            ui->horizontalSpacerShoulderButtonsWidget2,
+            ui->buttonMiscButtonsMinusGroup,
+            ui->buttonMiscButtonsScreenshotGroup,
+        };
+        break;
     }
 
     for (auto* widget : layout_hidden) {
@@ -1063,6 +1080,52 @@ void ConfigureInputPlayer::UpdateControllerAvailableButtons() {
     }
 }
 
+void ConfigureInputPlayer::UpdateControllerEnabledButtons() {
+    auto layout = GetControllerTypeFromIndex(ui->comboControllerType->currentIndex());
+    if (debug) {
+        layout = Settings::ControllerType::ProController;
+    }
+
+    // List of all the widgets that will be disabled by any of the following layouts that need
+    // "enabled" after the controller type changes
+    const std::array<QWidget*, 4> layout_enable = {
+        ui->buttonHome,
+        ui->buttonLStickPressedGroup,
+        ui->groupRStickPressed,
+        ui->buttonShoulderButtonsButtonLGroup,
+    };
+
+    for (auto* widget : layout_enable) {
+        widget->setEnabled(true);
+    }
+
+    std::vector<QWidget*> layout_disable;
+    switch (layout) {
+    case Settings::ControllerType::ProController:
+    case Settings::ControllerType::DualJoyconDetached:
+    case Settings::ControllerType::Handheld:
+    case Settings::ControllerType::LeftJoycon:
+    case Settings::ControllerType::RightJoycon:
+        // TODO(wwylele): enable this when we actually emulate it
+        layout_disable = {
+            ui->buttonHome,
+        };
+        break;
+    case Settings::ControllerType::GameCube:
+        layout_disable = {
+            ui->buttonHome,
+            ui->buttonLStickPressedGroup,
+            ui->groupRStickPressed,
+            ui->buttonShoulderButtonsButtonLGroup,
+        };
+        break;
+    }
+
+    for (auto* widget : layout_disable) {
+        widget->setEnabled(false);
+    }
+}
+
 void ConfigureInputPlayer::UpdateMotionButtons() {
     if (debug) {
         // Motion isn't used with the debug controller, hide both groupboxes.
@@ -1085,6 +1148,11 @@ void ConfigureInputPlayer::UpdateMotionButtons() {
         ui->buttonMotionLeftGroup->hide();
         ui->buttonMotionRightGroup->show();
         break;
+    case Settings::ControllerType::GameCube:
+        // Hide both "Motion 1/2".
+        ui->buttonMotionLeftGroup->hide();
+        ui->buttonMotionRightGroup->hide();
+        break;
     case Settings::ControllerType::DualJoyconDetached:
     default:
         // Show both "Motion 1/2".
@@ -1094,6 +1162,36 @@ void ConfigureInputPlayer::UpdateMotionButtons() {
     }
 }
 
+void ConfigureInputPlayer::UpdateControllerButtonNames() {
+    auto layout = GetControllerTypeFromIndex(ui->comboControllerType->currentIndex());
+    if (debug) {
+        layout = Settings::ControllerType::ProController;
+    }
+
+    switch (layout) {
+    case Settings::ControllerType::ProController:
+    case Settings::ControllerType::DualJoyconDetached:
+    case Settings::ControllerType::Handheld:
+    case Settings::ControllerType::LeftJoycon:
+    case Settings::ControllerType::RightJoycon:
+        ui->buttonMiscButtonsPlusGroup->setTitle(tr("Plus"));
+        ui->buttonShoulderButtonsButtonZLGroup->setTitle(tr("ZL"));
+        ui->buttonShoulderButtonsZRGroup->setTitle(tr("ZR"));
+        ui->buttonShoulderButtonsRGroup->setTitle(tr("R"));
+        ui->LStick->setTitle(tr("Left Stick"));
+        ui->RStick->setTitle(tr("Right Stick"));
+        break;
+    case Settings::ControllerType::GameCube:
+        ui->buttonMiscButtonsPlusGroup->setTitle(tr("Start / Pause"));
+        ui->buttonShoulderButtonsButtonZLGroup->setTitle(tr("L"));
+        ui->buttonShoulderButtonsZRGroup->setTitle(tr("R"));
+        ui->buttonShoulderButtonsRGroup->setTitle(tr("Z"));
+        ui->LStick->setTitle(tr("Control Stick"));
+        ui->RStick->setTitle(tr("C-Stick"));
+        break;
+    }
+}
+
 void ConfigureInputPlayer::UpdateMappingWithDefaults() {
     if (ui->comboDevices->currentIndex() == 0) {
         return;
diff --git a/src/yuzu/configuration/configure_input_player.h b/src/yuzu/configuration/configure_input_player.h
index da2b89136..efe953fbc 100644
--- a/src/yuzu/configuration/configure_input_player.h
+++ b/src/yuzu/configuration/configure_input_player.h
@@ -143,9 +143,15 @@ private:
     /// Hides and disables controller settings based on the current controller type.
     void UpdateControllerAvailableButtons();
 
+    /// Disables controller settings based on the current controller type.
+    void UpdateControllerEnabledButtons();
+
     /// Shows or hides motion groupboxes based on the current controller type.
     void UpdateMotionButtons();
 
+    /// Alters the button names based on the current controller type.
+    void UpdateControllerButtonNames();
+
     /// Gets the default controller mapping for this device and auto configures the input to match.
     void UpdateMappingWithDefaults();
 
diff --git a/src/yuzu/configuration/configure_input_player_widget.cpp b/src/yuzu/configuration/configure_input_player_widget.cpp
index 0e8a964d2..61ba91cef 100644
--- a/src/yuzu/configuration/configure_input_player_widget.cpp
+++ b/src/yuzu/configuration/configure_input_player_widget.cpp
@@ -227,6 +227,9 @@ void PlayerControlPreview::paintEvent(QPaintEvent* event) {
     case Settings::ControllerType::RightJoycon:
         DrawRightController(p, center);
         break;
+    case Settings::ControllerType::GameCube:
+        DrawGCController(p, center);
+        break;
     case Settings::ControllerType::ProController:
     default:
         DrawProController(p, center);
diff --git a/src/yuzu/debugger/controller.cpp b/src/yuzu/debugger/controller.cpp
index 85724a8f3..2731d948d 100644
--- a/src/yuzu/debugger/controller.cpp
+++ b/src/yuzu/debugger/controller.cpp
@@ -42,7 +42,7 @@ void ControllerDialog::refreshConfiguration() {
 
 QAction* ControllerDialog::toggleViewAction() {
     if (toggle_view_action == nullptr) {
-        toggle_view_action = new QAction(windowTitle(), this);
+        toggle_view_action = new QAction(tr("&Controller P1"), this);
         toggle_view_action->setCheckable(true);
         toggle_view_action->setChecked(isVisible());
         connect(toggle_view_action, &QAction::toggled, this, &ControllerDialog::setVisible);
diff --git a/src/yuzu_cmd/config.cpp b/src/yuzu_cmd/config.cpp
index aa0a9f288..6d8bc5509 100644
--- a/src/yuzu_cmd/config.cpp
+++ b/src/yuzu_cmd/config.cpp
@@ -329,9 +329,6 @@ void Config::ReadValues() {
     FS::GetUserPath(
         FS::UserPath::DumpDir,
         sdl2_config->Get("Data Storage", "dump_directory", FS::GetUserPath(FS::UserPath::DumpDir)));
-    FS::GetUserPath(FS::UserPath::CacheDir,
-                    sdl2_config->Get("Data Storage", "cache_directory",
-                                     FS::GetUserPath(FS::UserPath::CacheDir)));
     Settings::values.gamecard_inserted =
         sdl2_config->GetBoolean("Data Storage", "gamecard_inserted", false);
     Settings::values.gamecard_current_game =