early-access version 1341EA-1341

14 files changed, 152 insertions, 242 deletions

diff --git a/README.md b/README.md
index d851a7df5..ec04f01fe 100755
--- a/README.md
+++ b/README.md
@@ -1,7 +1,7 @@
 yuzu emulator early access
 =============
 
-This is the source code for early-access 1339.
+This is the source code for early-access 1341.
 
 ## Legal Notice
 
diff --git a/src/core/memory.cpp b/src/core/memory.cpp
index 961f1c028..11609682a 100755
--- a/src/core/memory.cpp
+++ b/src/core/memory.cpp
@@ -195,9 +195,8 @@ struct Memory::Impl {
             switch (type) {
             case Common::PageType::Unmapped: {
                 LOG_ERROR(HW_Memory,
-                          "Unmapped ReadBlock @ 0x{:016X} (start address = 0x{:016X}, size = {}) "
-                          "at PC 0x{:08X}",
-                          current_vaddr, src_addr, size, system.CurrentArmInterface().GetPC());
+                          "Unmapped ReadBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})",
+                          current_vaddr, src_addr, size);
                 std::memset(dest_buffer, 0, copy_amount);
                 break;
             }
@@ -241,9 +240,8 @@ struct Memory::Impl {
             switch (type) {
             case Common::PageType::Unmapped: {
                 LOG_ERROR(HW_Memory,
-                          "Unmapped ReadBlock @ 0x{:016X} (start address = 0x{:016X}, size = {}) "
-                          "at PC 0x{:08X}",
-                          current_vaddr, src_addr, size, system.CurrentArmInterface().GetPC());
+                          "Unmapped ReadBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})",
+                          current_vaddr, src_addr, size);
                 std::memset(dest_buffer, 0, copy_amount);
                 break;
             }
@@ -293,9 +291,8 @@ struct Memory::Impl {
             switch (type) {
             case Common::PageType::Unmapped: {
                 LOG_ERROR(HW_Memory,
-                          "Unmapped WriteBlock @ 0x{:016X} (start address = 0x{:016X}, size = {}) "
-                          "at PC 0x{:08X}",
-                          current_vaddr, dest_addr, size, system.CurrentArmInterface().GetPC());
+                          "Unmapped WriteBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})",
+                          current_vaddr, dest_addr, size);
                 break;
             }
             case Common::PageType::Memory: {
@@ -337,9 +334,8 @@ struct Memory::Impl {
             switch (type) {
             case Common::PageType::Unmapped: {
                 LOG_ERROR(HW_Memory,
-                          "Unmapped WriteBlock @ 0x{:016X} (start address = 0x{:016X}, size = {}) "
-                          "at PC 0x{:08X}",
-                          current_vaddr, dest_addr, size, system.CurrentArmInterface().GetPC());
+                          "Unmapped WriteBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})",
+                          current_vaddr, dest_addr, size);
                 break;
             }
             case Common::PageType::Memory: {
@@ -387,9 +383,8 @@ struct Memory::Impl {
             switch (type) {
             case Common::PageType::Unmapped: {
                 LOG_ERROR(HW_Memory,
-                          "Unmapped ZeroBlock @ 0x{:016X} (start address = 0x{:016X}, size = {}) "
-                          "at PC 0x{:08X}",
-                          current_vaddr, dest_addr, size, system.CurrentArmInterface().GetPC());
+                          "Unmapped ZeroBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})",
+                          current_vaddr, dest_addr, size);
                 break;
             }
             case Common::PageType::Memory: {
@@ -434,9 +429,8 @@ struct Memory::Impl {
             switch (type) {
             case Common::PageType::Unmapped: {
                 LOG_ERROR(HW_Memory,
-                          "Unmapped CopyBlock @ 0x{:016X} (start address = 0x{:016X}, size = {}) "
-                          "at PC 0x{:08X}",
-                          current_vaddr, src_addr, size, system.CurrentArmInterface().GetPC());
+                          "Unmapped CopyBlock @ 0x{:016X} (start address = 0x{:016X}, size = {})",
+                          current_vaddr, src_addr, size);
                 ZeroBlock(process, dest_addr, copy_amount);
                 break;
             }
@@ -607,8 +601,7 @@ struct Memory::Impl {
         }
         switch (Common::PageTable::PageInfo::ExtractType(raw_pointer)) {
         case Common::PageType::Unmapped:
-            LOG_ERROR(HW_Memory, "Unmapped Read{} @ 0x{:08X} at PC 0x{:08X}", sizeof(T) * 8, vaddr,
-                      system.CurrentArmInterface().GetPC());
+            LOG_ERROR(HW_Memory, "Unmapped Read{} @ 0x{:08X}", sizeof(T) * 8, vaddr);
             return 0;
         case Common::PageType::Memory:
             ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", vaddr);
@@ -645,9 +638,8 @@ struct Memory::Impl {
         }
         switch (Common::PageTable::PageInfo::ExtractType(raw_pointer)) {
         case Common::PageType::Unmapped:
-            LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X} at PC 0x{:08X}",
-                      sizeof(data) * 8, static_cast<u32>(data), vaddr,
-                      system.CurrentArmInterface().GetPC());
+            LOG_ERROR(HW_Memory, "Unmapped Write{} 0x{:08X} @ 0x{:016X}", sizeof(data) * 8,
+                      static_cast<u32>(data), vaddr);
             return;
         case Common::PageType::Memory:
             ASSERT_MSG(false, "Mapped memory page without a pointer @ {:016X}", vaddr);
diff --git a/src/video_core/buffer_cache/buffer_base.h b/src/video_core/buffer_cache/buffer_base.h
index 8a5e6a3e7..0c00ae280 100755
--- a/src/video_core/buffer_cache/buffer_base.h
+++ b/src/video_core/buffer_cache/buffer_base.h
@@ -251,6 +251,16 @@ public:
         flags &= ~BufferFlagBits::Picked;
     }
 
+    /// Increases the likeliness of this being a stream buffer
+    void IncreaseStreamScore(int score) noexcept {
+        stream_score += score;
+    }
+
+    /// Returns the likeliness of this being a stream buffer
+    [[nodiscard]] int StreamScore() const noexcept {
+        return stream_score;
+    }
+
     /// Returns true when vaddr -> vaddr+size is fully contained in the buffer
     [[nodiscard]] bool IsInBounds(VAddr addr, u64 size) const noexcept {
         return addr >= cpu_addr && addr + size <= cpu_addr + SizeBytes();
@@ -574,6 +584,7 @@ private:
     VAddr cpu_addr = 0;
     Words words;
     BufferFlagBits flags{};
+    int stream_score = 0;
 };
 
 } // namespace VideoCommon
diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h
index bd8507610..a296036f4 100755
--- a/src/video_core/buffer_cache/buffer_cache.h
+++ b/src/video_core/buffer_cache/buffer_cache.h
@@ -71,6 +71,13 @@ class BufferCache {
 
     struct Empty {};
 
+    struct OverlapResult {
+        std::vector<BufferId> ids;
+        VAddr begin;
+        VAddr end;
+        bool has_stream_leap = false;
+    };
+
     struct Binding {
         VAddr cpu_addr{};
         u32 size{};
@@ -84,7 +91,7 @@ class BufferCache {
     };
 
 public:
-    static constexpr size_t SKIP_CACHE_SIZE = 4096;
+    static constexpr u32 SKIP_CACHE_SIZE = 4096;
 
     explicit BufferCache(VideoCore::RasterizerInterface& rasterizer_,
                          Tegra::Engines::Maxwell3D& maxwell3d_,
@@ -220,6 +227,10 @@ private:
 
     [[nodiscard]] BufferId FindBuffer(VAddr cpu_addr, u32 size);
 
+    [[nodiscard]] OverlapResult ResolveOverlaps(VAddr cpu_addr, u32 wanted_size);
+
+    void JoinOverlap(BufferId new_buffer_id, BufferId overlap_id, bool accumulate_stream_score);
+
     [[nodiscard]] BufferId CreateBuffer(VAddr cpu_addr, u32 wanted_size);
 
     void Register(BufferId buffer_id);
@@ -988,12 +999,15 @@ BufferId BufferCache<P>::FindBuffer(VAddr cpu_addr, u32 size) {
 }
 
 template <class P>
-BufferId BufferCache<P>::CreateBuffer(VAddr cpu_addr, u32 wanted_size) {
+typename BufferCache<P>::OverlapResult BufferCache<P>::ResolveOverlaps(VAddr cpu_addr,
+                                                                       u32 wanted_size) {
+    static constexpr int STREAM_LEAP_THRESHOLD = 16;
     std::vector<BufferId> overlap_ids;
-    VAddr cpu_addr_begin = cpu_addr;
-    VAddr cpu_addr_end = cpu_addr + wanted_size;
-    for (; cpu_addr >> PAGE_BITS < Common::DivCeil(cpu_addr_end, PAGE_SIZE);
-         cpu_addr += PAGE_SIZE) {
+    VAddr begin = cpu_addr;
+    VAddr end = cpu_addr + wanted_size;
+    int stream_score = 0;
+    bool has_stream_leap = false;
+    for (; cpu_addr >> PAGE_BITS < Common::DivCeil(end, PAGE_SIZE); cpu_addr += PAGE_SIZE) {
         const BufferId overlap_id = page_table[cpu_addr >> PAGE_BITS];
         if (!overlap_id) {
             continue;
@@ -1002,38 +1016,63 @@ BufferId BufferCache<P>::CreateBuffer(VAddr cpu_addr, u32 wanted_size) {
         if (overlap.IsPicked()) {
             continue;
         }
-        overlap.Pick();
         overlap_ids.push_back(overlap_id);
+        overlap.Pick();
         const VAddr overlap_cpu_addr = overlap.CpuAddr();
-        if (overlap_cpu_addr < cpu_addr_begin) {
-            cpu_addr = cpu_addr_begin = overlap_cpu_addr;
+        if (overlap_cpu_addr < begin) {
+            cpu_addr = begin = overlap_cpu_addr;
+        }
+        end = std::max(end, overlap_cpu_addr + overlap.SizeBytes());
+
+        stream_score += overlap.StreamScore();
+        if (stream_score > STREAM_LEAP_THRESHOLD && !has_stream_leap) {
+            // When this memory region has been joined a bunch of times, we assume it's being used
+            // as a stream buffer. Increase the size to skip constantly recreating buffers.
+            has_stream_leap = true;
+            end += PAGE_SIZE * 256;
         }
-        cpu_addr_end = std::max(cpu_addr_end, overlap_cpu_addr + overlap.SizeBytes());
     }
-    const u32 size = static_cast<u32>(cpu_addr_end - cpu_addr_begin);
-    const BufferId new_buffer_id = slot_buffers.insert(runtime, rasterizer, cpu_addr_begin, size);
-    Buffer& new_buffer = slot_buffers[new_buffer_id];
-
-    for (const BufferId overlap_id : overlap_ids) {
-        Buffer& overlap = slot_buffers[overlap_id];
-        overlap.Unpick();
+    return OverlapResult{
+        .ids = std::move(overlap_ids),
+        .begin = begin,
+        .end = end,
+        .has_stream_leap = has_stream_leap,
+    };
+}
 
-        std::vector<BufferCopy> copies;
-        const size_t dst_base_offset = overlap.CpuAddr() - new_buffer.CpuAddr();
-        overlap.ForEachDownloadRange([&](u64 begin, u64 range_size) {
-            copies.push_back(BufferCopy{
-                .src_offset = begin,
-                .dst_offset = dst_base_offset + begin,
-                .size = range_size,
-            });
-            new_buffer.UnmarkRegionAsCpuModified(begin, range_size);
-            new_buffer.MarkRegionAsGpuModified(begin, range_size);
+template <class P>
+void BufferCache<P>::JoinOverlap(BufferId new_buffer_id, BufferId overlap_id,
+                                 bool accumulate_stream_score) {
+    Buffer& new_buffer = slot_buffers[new_buffer_id];
+    Buffer& overlap = slot_buffers[overlap_id];
+    if (accumulate_stream_score) {
+        new_buffer.IncreaseStreamScore(overlap.StreamScore() + 1);
+    }
+    std::vector<BufferCopy> copies;
+    const size_t dst_base_offset = overlap.CpuAddr() - new_buffer.CpuAddr();
+    overlap.ForEachDownloadRange([&](u64 begin, u64 range_size) {
+        copies.push_back(BufferCopy{
+            .src_offset = begin,
+            .dst_offset = dst_base_offset + begin,
+            .size = range_size,
         });
-        if (!copies.empty()) {
-            runtime.CopyBuffer(slot_buffers[new_buffer_id], overlap, copies);
-        }
-        ReplaceBufferDownloads(overlap_id, new_buffer_id);
-        DeleteBuffer(overlap_id);
+        new_buffer.UnmarkRegionAsCpuModified(begin, range_size);
+        new_buffer.MarkRegionAsGpuModified(begin, range_size);
+    });
+    if (!copies.empty()) {
+        runtime.CopyBuffer(slot_buffers[new_buffer_id], overlap, copies);
+    }
+    ReplaceBufferDownloads(overlap_id, new_buffer_id);
+    DeleteBuffer(overlap_id);
+}
+
+template <class P>
+BufferId BufferCache<P>::CreateBuffer(VAddr cpu_addr, u32 wanted_size) {
+    const OverlapResult overlap = ResolveOverlaps(cpu_addr, wanted_size);
+    const u32 size = static_cast<u32>(overlap.end - overlap.begin);
+    const BufferId new_buffer_id = slot_buffers.insert(runtime, rasterizer, overlap.begin, size);
+    for (const BufferId overlap_id : overlap.ids) {
+        JoinOverlap(new_buffer_id, overlap_id, !overlap.has_stream_leap);
     }
     Register(new_buffer_id);
     return new_buffer_id;
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.cpp b/src/video_core/renderer_opengl/gl_buffer_cache.cpp
index 889ad6c56..6da3906a4 100755
--- a/src/video_core/renderer_opengl/gl_buffer_cache.cpp
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.cpp
@@ -7,10 +7,6 @@
 #include "video_core/buffer_cache/buffer_cache.h"
 #include "video_core/renderer_opengl/gl_buffer_cache.h"
 #include "video_core/renderer_opengl/gl_device.h"
-#include "video_core/vulkan_common/vulkan_device.h"
-#include "video_core/vulkan_common/vulkan_instance.h"
-#include "video_core/vulkan_common/vulkan_library.h"
-#include "video_core/vulkan_common/vulkan_memory_allocator.h"
 
 namespace OpenGL {
 namespace {
@@ -36,13 +32,8 @@ Buffer::Buffer(BufferCacheRuntime& runtime, VideoCore::RasterizerInterface& rast
     buffer.Create();
     const std::string name = fmt::format("Buffer 0x{:x}", CpuAddr());
     glObjectLabel(GL_BUFFER, buffer.handle, static_cast<GLsizei>(name.size()), name.data());
-    if (runtime.device.UseAssemblyShaders()) {
-        CreateMemoryObjects(runtime);
-        glNamedBufferStorageMemEXT(buffer.handle, SizeBytes(), memory_commit.ExportOpenGLHandle(),
-                                   memory_commit.Offset());
-    } else {
-        glNamedBufferData(buffer.handle, SizeBytes(), nullptr, GL_DYNAMIC_DRAW);
-    }
+    glNamedBufferData(buffer.handle, SizeBytes(), nullptr, GL_DYNAMIC_DRAW);
+
     if (runtime.has_unified_vertex_buffers) {
         glGetNamedBufferParameterui64vNV(buffer.handle, GL_BUFFER_GPU_ADDRESS_NV, &address);
     }
@@ -71,61 +62,30 @@ void Buffer::MakeResident(GLenum access) noexcept {
     glMakeNamedBufferResidentNV(buffer.handle, access);
 }
 
-GLuint Buffer::SubBuffer(u32 offset) {
-    if (offset == 0) {
-        return buffer.handle;
-    }
-    for (const auto& [sub_buffer, sub_offset] : subs) {
-        if (sub_offset == offset) {
-            return sub_buffer.handle;
-        }
-    }
-    OGLBuffer sub_buffer;
-    sub_buffer.Create();
-    glNamedBufferStorageMemEXT(sub_buffer.handle, SizeBytes() - offset,
-                               memory_commit.ExportOpenGLHandle(), memory_commit.Offset() + offset);
-    return subs.emplace_back(std::move(sub_buffer), offset).first.handle;
-}
-
-void Buffer::CreateMemoryObjects(BufferCacheRuntime& runtime) {
-    auto& allocator = runtime.vulkan_memory_allocator;
-    auto& device = runtime.vulkan_device->GetLogical();
-    auto vulkan_buffer = device.CreateBuffer(VkBufferCreateInfo{
-        .sType = VK_STRUCTURE_TYPE_BUFFER_CREATE_INFO,
-        .pNext = nullptr,
-        .flags = 0,
-        .size = SizeBytes(),
-        .usage = VK_BUFFER_USAGE_TRANSFER_SRC_BIT | VK_BUFFER_USAGE_TRANSFER_DST_BIT |
-                 VK_BUFFER_USAGE_UNIFORM_TEXEL_BUFFER_BIT |
-                 VK_BUFFER_USAGE_STORAGE_TEXEL_BUFFER_BIT | VK_BUFFER_USAGE_UNIFORM_BUFFER_BIT |
-                 VK_BUFFER_USAGE_STORAGE_BUFFER_BIT | VK_BUFFER_USAGE_INDEX_BUFFER_BIT |
-                 VK_BUFFER_USAGE_VERTEX_BUFFER_BIT,
-        .sharingMode = VK_SHARING_MODE_EXCLUSIVE,
-        .queueFamilyIndexCount = 0,
-        .pQueueFamilyIndices = nullptr,
-    });
-    const VkMemoryRequirements requirements = device.GetBufferMemoryRequirements(*vulkan_buffer);
-    memory_commit = allocator->Commit(requirements, Vulkan::MemoryUsage::DeviceLocal);
-}
-
-BufferCacheRuntime::BufferCacheRuntime(const Device& device_, const Vulkan::Device* vulkan_device_,
-                                       Vulkan::MemoryAllocator* vulkan_memory_allocator_)
-    : device{device_}, vulkan_device{vulkan_device_},
-      vulkan_memory_allocator{vulkan_memory_allocator_},
-      stream_buffer{device.HasFastBufferSubData() ? std::nullopt
-                                                  : std::make_optional<StreamBuffer>()} {
+BufferCacheRuntime::BufferCacheRuntime(const Device& device_)
+    : device{device_}, has_fast_buffer_sub_data{device.HasFastBufferSubData()},
+      use_assembly_shaders{device.UseAssemblyShaders()},
+      has_unified_vertex_buffers{device.HasVertexBufferUnifiedMemory()},
+      stream_buffer{has_fast_buffer_sub_data ? std::nullopt : std::make_optional<StreamBuffer>()} {
     GLint gl_max_attributes;
     glGetIntegerv(GL_MAX_VERTEX_ATTRIBS, &gl_max_attributes);
     max_attributes = static_cast<u32>(gl_max_attributes);
-    use_assembly_shaders = device.UseAssemblyShaders();
-    has_unified_vertex_buffers = device.HasVertexBufferUnifiedMemory();
-
     for (auto& stage_uniforms : fast_uniforms) {
         for (OGLBuffer& buffer : stage_uniforms) {
             buffer.Create();
             glNamedBufferData(buffer.handle, BufferCache::SKIP_CACHE_SIZE, nullptr, GL_STREAM_DRAW);
         }
     }
+    for (auto& stage_uniforms : copy_uniforms) {
+        for (OGLBuffer& buffer : stage_uniforms) {
+            buffer.Create();
+            glNamedBufferData(buffer.handle, 0x10'000, nullptr, GL_STREAM_COPY);
+        }
+    }
+    for (OGLBuffer& buffer : copy_compute_uniforms) {
+        buffer.Create();
+        glNamedBufferData(buffer.handle, 0x10'000, nullptr, GL_STREAM_COPY);
+    }
 }
 
 void BufferCacheRuntime::CopyBuffer(Buffer& dst_buffer, Buffer& src_buffer,
@@ -167,8 +127,14 @@ void BufferCacheRuntime::BindVertexBuffer(u32 index, Buffer& buffer, u32 offset,
 void BufferCacheRuntime::BindUniformBuffer(size_t stage, u32 binding_index, Buffer& buffer,
                                            u32 offset, u32 size) {
     if (use_assembly_shaders) {
-        const GLuint sub_buffer = buffer.SubBuffer(offset);
-        glBindBufferRangeNV(PABO_LUT[stage], binding_index, sub_buffer, 0,
+        GLuint handle;
+        if (offset != 0) {
+            handle = copy_uniforms[stage][binding_index].handle;
+            glCopyNamedBufferSubData(buffer.Handle(), handle, offset, 0, size);
+        } else {
+            handle = buffer.Handle();
+        }
+        glBindBufferRangeNV(PABO_LUT[stage], binding_index, handle, 0,
                             static_cast<GLsizeiptr>(size));
     } else {
         const GLuint base_binding = device.GetBaseBindings(stage).uniform_buffer;
@@ -181,8 +147,15 @@ void BufferCacheRuntime::BindUniformBuffer(size_t stage, u32 binding_index, Buff
 void BufferCacheRuntime::BindComputeUniformBuffer(u32 binding_index, Buffer& buffer, u32 offset,
                                                   u32 size) {
     if (use_assembly_shaders) {
-        glBindBufferRangeNV(GL_COMPUTE_PROGRAM_PARAMETER_BUFFER_NV, binding_index,
-                            buffer.SubBuffer(offset), 0, static_cast<GLsizeiptr>(size));
+        GLuint handle;
+        if (offset != 0) {
+            handle = copy_compute_uniforms[binding_index].handle;
+            glCopyNamedBufferSubData(buffer.Handle(), handle, offset, 0, size);
+        } else {
+            handle = buffer.Handle();
+        }
+        glBindBufferRangeNV(GL_COMPUTE_PROGRAM_PARAMETER_BUFFER_NV, binding_index, handle, 0,
+                            static_cast<GLsizeiptr>(size));
     } else {
         glBindBufferRange(GL_UNIFORM_BUFFER, binding_index, buffer.Handle(),
                           static_cast<GLintptr>(offset), static_cast<GLsizeiptr>(size));
diff --git a/src/video_core/renderer_opengl/gl_buffer_cache.h b/src/video_core/renderer_opengl/gl_buffer_cache.h
index f4d8871a9..d8b20a9af 100755
--- a/src/video_core/renderer_opengl/gl_buffer_cache.h
+++ b/src/video_core/renderer_opengl/gl_buffer_cache.h
@@ -15,13 +15,6 @@
 #include "video_core/renderer_opengl/gl_device.h"
 #include "video_core/renderer_opengl/gl_resource_manager.h"
 #include "video_core/renderer_opengl/gl_stream_buffer.h"
-#include "video_core/vulkan_common/vulkan_device.h"
-#include "video_core/vulkan_common/vulkan_memory_allocator.h"
-
-namespace Vulkan {
-class Device;
-class MemoryAllocator;
-} // namespace Vulkan
 
 namespace OpenGL {
 
@@ -39,8 +32,6 @@ public:
 
     void MakeResident(GLenum access) noexcept;
 
-    [[nodiscard]] GLuint SubBuffer(u32 offset);
-
     [[nodiscard]] GLuint64EXT HostGpuAddr() const noexcept {
         return address;
     }
@@ -50,13 +41,9 @@ public:
     }
 
 private:
-    void CreateMemoryObjects(BufferCacheRuntime& runtime);
-
     GLuint64EXT address = 0;
-    Vulkan::MemoryCommit memory_commit;
     OGLBuffer buffer;
     GLenum current_residency_access = GL_NONE;
-    std::vector<std::pair<OGLBuffer, u32>> subs;
 };
 
 class BufferCacheRuntime {
@@ -65,8 +52,7 @@ class BufferCacheRuntime {
 public:
     static constexpr u8 INVALID_BINDING = std::numeric_limits<u8>::max();
 
-    explicit BufferCacheRuntime(const Device& device_, const Vulkan::Device* vulkan_device_,
-                                Vulkan::MemoryAllocator* vulkan_memory_allocator_);
+    explicit BufferCacheRuntime(const Device& device_);
 
     void CopyBuffer(Buffer& dst_buffer, Buffer& src_buffer,
                     std::span<const VideoCommon::BufferCopy> copies);
@@ -127,7 +113,7 @@ public:
     }
 
     [[nodiscard]] bool HasFastBufferSubData() const noexcept {
-        return device.HasFastBufferSubData();
+        return has_fast_buffer_sub_data;
     }
 
 private:
@@ -138,18 +124,22 @@ private:
     };
 
     const Device& device;
-    const Vulkan::Device* vulkan_device;
-    Vulkan::MemoryAllocator* vulkan_memory_allocator;
-    std::optional<StreamBuffer> stream_buffer;
-
-    u32 max_attributes = 0;
 
+    bool has_fast_buffer_sub_data = false;
     bool use_assembly_shaders = false;
     bool has_unified_vertex_buffers = false;
 
+    u32 max_attributes = 0;
+
+    std::optional<StreamBuffer> stream_buffer;
+
     std::array<std::array<OGLBuffer, VideoCommon::NUM_GRAPHICS_UNIFORM_BUFFERS>,
                VideoCommon::NUM_STAGES>
         fast_uniforms;
+    std::array<std::array<OGLBuffer, VideoCommon::NUM_GRAPHICS_UNIFORM_BUFFERS>,
+               VideoCommon::NUM_STAGES>
+        copy_uniforms;
+    std::array<OGLBuffer, VideoCommon::NUM_COMPUTE_UNIFORM_BUFFERS> copy_compute_uniforms;
 
     u32 index_buffer_offset = 0;
 };
diff --git a/src/video_core/renderer_opengl/gl_device.cpp b/src/video_core/renderer_opengl/gl_device.cpp
index 7310185e6..703391ca5 100755
--- a/src/video_core/renderer_opengl/gl_device.cpp
+++ b/src/video_core/renderer_opengl/gl_device.cpp
@@ -197,7 +197,7 @@ bool IsASTCSupported() {
 }
 } // Anonymous namespace
 
-Device::Device(bool has_vulkan_instance) {
+Device::Device() {
     if (!GLAD_GL_VERSION_4_3) {
         LOG_ERROR(Render_OpenGL, "OpenGL 4.3 is not available");
         throw std::runtime_error{"Insufficient version"};
@@ -246,8 +246,7 @@ Device::Device(bool has_vulkan_instance) {
 
     use_assembly_shaders = Settings::values.use_assembly_shaders.GetValue() &&
                            GLAD_GL_NV_gpu_program5 && GLAD_GL_NV_compute_program5 &&
-                           GLAD_GL_NV_transform_feedback && GLAD_GL_NV_transform_feedback2 &&
-                           has_vulkan_instance;
+                           GLAD_GL_NV_transform_feedback && GLAD_GL_NV_transform_feedback2;
 
     use_asynchronous_shaders = Settings::values.use_asynchronous_shaders.GetValue();
 
diff --git a/src/video_core/renderer_opengl/gl_device.h b/src/video_core/renderer_opengl/gl_device.h
index 21a38544f..9e620f603 100755
--- a/src/video_core/renderer_opengl/gl_device.h
+++ b/src/video_core/renderer_opengl/gl_device.h
@@ -19,7 +19,7 @@ public:
         u32 image{};
     };
 
-    explicit Device(bool has_vulkan_instance);
+    explicit Device();
     explicit Device(std::nullptr_t);
 
     u32 GetMaxUniformBuffers(Tegra::Engines::ShaderType shader_type) const noexcept {
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.cpp b/src/video_core/renderer_opengl/gl_rasterizer.cpp
index 3e90e6481..8298d6787 100755
--- a/src/video_core/renderer_opengl/gl_rasterizer.cpp
+++ b/src/video_core/renderer_opengl/gl_rasterizer.cpp
@@ -170,8 +170,6 @@ ImageViewType ImageViewTypeFromEntry(const ImageEntry& entry) {
 
 RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_,
                                    Core::Memory::Memory& cpu_memory_, const Device& device_,
-                                   const Vulkan::Device* vulkan_device,
-                                   Vulkan::MemoryAllocator* vulkan_memory_allocator,
                                    ScreenInfo& screen_info_, ProgramManager& program_manager_,
                                    StateTracker& state_tracker_)
     : RasterizerAccelerated(cpu_memory_), gpu(gpu_), maxwell3d(gpu.Maxwell3D()),
@@ -179,7 +177,7 @@ RasterizerOpenGL::RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra
       screen_info(screen_info_), program_manager(program_manager_), state_tracker(state_tracker_),
       texture_cache_runtime(device, program_manager, state_tracker),
       texture_cache(texture_cache_runtime, *this, maxwell3d, kepler_compute, gpu_memory),
-      buffer_cache_runtime(device, vulkan_device, vulkan_memory_allocator),
+      buffer_cache_runtime(device),
       buffer_cache(*this, maxwell3d, kepler_compute, gpu_memory, cpu_memory_, buffer_cache_runtime),
       shader_cache(*this, emu_window_, gpu, maxwell3d, kepler_compute, gpu_memory, device),
       query_cache(*this, maxwell3d, gpu_memory),
diff --git a/src/video_core/renderer_opengl/gl_rasterizer.h b/src/video_core/renderer_opengl/gl_rasterizer.h
index b1094075a..9730544d9 100755
--- a/src/video_core/renderer_opengl/gl_rasterizer.h
+++ b/src/video_core/renderer_opengl/gl_rasterizer.h
@@ -46,11 +46,6 @@ namespace Tegra {
 class MemoryManager;
 }
 
-namespace Vulkan {
-class Device;
-class MemoryAllocator;
-} // namespace Vulkan
-
 namespace OpenGL {
 
 struct ScreenInfo;
@@ -67,8 +62,6 @@ class RasterizerOpenGL : public VideoCore::RasterizerAccelerated {
 public:
     explicit RasterizerOpenGL(Core::Frontend::EmuWindow& emu_window_, Tegra::GPU& gpu_,
                               Core::Memory::Memory& cpu_memory_, const Device& device_,
-                              const Vulkan::Device* vulkan_device,
-                              Vulkan::MemoryAllocator* vulkan_memory_allocator,
                               ScreenInfo& screen_info_, ProgramManager& program_manager_,
                               StateTracker& state_tracker_);
     ~RasterizerOpenGL() override;
diff --git a/src/video_core/renderer_opengl/renderer_opengl.cpp b/src/video_core/renderer_opengl/renderer_opengl.cpp
index 8fcb86581..9d2acd4d9 100755
--- a/src/video_core/renderer_opengl/renderer_opengl.cpp
+++ b/src/video_core/renderer_opengl/renderer_opengl.cpp
@@ -27,11 +27,6 @@
 #include "video_core/renderer_opengl/gl_shader_manager.h"
 #include "video_core/renderer_opengl/renderer_opengl.h"
 #include "video_core/textures/decoders.h"
-#include "video_core/vulkan_common/vulkan_debug_callback.h"
-#include "video_core/vulkan_common/vulkan_device.h"
-#include "video_core/vulkan_common/vulkan_instance.h"
-#include "video_core/vulkan_common/vulkan_library.h"
-#include "video_core/vulkan_common/vulkan_memory_allocator.h"
 
 namespace OpenGL {
 namespace {
@@ -127,93 +122,16 @@ void APIENTRY DebugHandler(GLenum source, GLenum type, GLuint id, GLenum severit
         break;
     }
 }
-
-Vulkan::vk::PhysicalDevice FindPhysicalDevice(Vulkan::vk::Instance& instance) {
-    using namespace Vulkan;
-    using UUID = std::array<GLubyte, GL_UUID_SIZE_EXT>;
-
-    GLint num_device_uuids;
-    glGetIntegerv(GL_NUM_DEVICE_UUIDS_EXT, &num_device_uuids);
-    std::vector<UUID> device_uuids(num_device_uuids);
-    for (GLint index = 0; index < num_device_uuids; ++index) {
-        glGetUnsignedBytei_vEXT(GL_DEVICE_UUID_EXT, 0, device_uuids[index].data());
-    }
-    UUID driver_uuid;
-    glGetUnsignedBytevEXT(GL_DRIVER_UUID_EXT, driver_uuid.data());
-
-    for (const VkPhysicalDevice raw_physical_device : instance.EnumeratePhysicalDevices()) {
-        VkPhysicalDeviceIDProperties device_id_properties{};
-        device_id_properties.sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_ID_PROPERTIES;
-
-        VkPhysicalDeviceProperties2KHR properties{
-            .sType = VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR,
-            .pNext = &device_id_properties,
-            .properties{},
-        };
-        vk::PhysicalDevice physical_device(raw_physical_device, instance.Dispatch());
-        physical_device.GetProperties2KHR(properties);
-        if (!std::ranges::equal(device_id_properties.driverUUID, driver_uuid)) {
-            continue;
-        }
-        const auto it =
-            std::ranges::find_if(device_uuids, [&device_id_properties, driver_uuid](UUID uuid) {
-                return std::ranges::equal(device_id_properties.deviceUUID, uuid);
-            });
-        if (it != device_uuids.end()) {
-            return physical_device;
-        }
-    }
-    throw vk::Exception(VK_ERROR_INCOMPATIBLE_DRIVER);
-}
 } // Anonymous namespace
 
-struct VulkanObjects {
-    static std::unique_ptr<VulkanObjects> TryCreate() {
-        if (!GLAD_GL_EXT_memory_object) {
-            // Interop is not present
-            return nullptr;
-        }
-        const std::string_view vendor{reinterpret_cast<const char*>(glGetString(GL_VENDOR))};
-        if (vendor == "ATI Technologies Inc.") {
-            // Avoid using GL_EXT_memory_object on AMD, as it makes the GL driver crash
-            return nullptr;
-        }
-        if (!Settings::values.use_assembly_shaders.GetValue()) {
-            // We only need interop when assembly shaders are enabled
-            return nullptr;
-        }
-#ifdef __linux__
-        LOG_WARNING(Render_OpenGL, "Interop doesn't work on Linux at the moment");
-        return nullptr;
-#endif
-        try {
-            return std::make_unique<VulkanObjects>();
-        } catch (const Vulkan::vk::Exception& exception) {
-            LOG_ERROR(Render_OpenGL, "Failed to initialize Vulkan objects with error: {}",
-                      exception.what());
-            return nullptr;
-        }
-    }
-
-    Common::DynamicLibrary library{Vulkan::OpenLibrary()};
-    Vulkan::vk::InstanceDispatch dld;
-    Vulkan::vk::Instance instance{Vulkan::CreateInstance(library, dld, VK_API_VERSION_1_1)};
-    Vulkan::Device device{*instance, FindPhysicalDevice(instance), nullptr, dld};
-    Vulkan::MemoryAllocator memory_allocator{device, true};
-};
-
 RendererOpenGL::RendererOpenGL(Core::TelemetrySession& telemetry_session_,
                                Core::Frontend::EmuWindow& emu_window_,
                                Core::Memory::Memory& cpu_memory_, Tegra::GPU& gpu_,
                                std::unique_ptr<Core::Frontend::GraphicsContext> context_)
     : RendererBase{emu_window_, std::move(context_)}, telemetry_session{telemetry_session_},
-      emu_window{emu_window_}, cpu_memory{cpu_memory_}, gpu{gpu_},
-      vulkan_objects{VulkanObjects::TryCreate()}, device{vulkan_objects != nullptr},
-      state_tracker{gpu}, program_manager{device},
-      rasterizer(emu_window, gpu, cpu_memory, device,
-                 vulkan_objects ? &vulkan_objects->device : nullptr,
-                 vulkan_objects ? &vulkan_objects->memory_allocator : nullptr, screen_info,
-                 program_manager, state_tracker) {
+      emu_window{emu_window_}, cpu_memory{cpu_memory_}, gpu{gpu_}, state_tracker{gpu},
+      program_manager{device},
+      rasterizer(emu_window, gpu, cpu_memory, device, screen_info, program_manager, state_tracker) {
     if (Settings::values.renderer_debug && GLAD_GL_KHR_debug) {
         glEnable(GL_DEBUG_OUTPUT);
         glEnable(GL_DEBUG_OUTPUT_SYNCHRONOUS);
diff --git a/src/video_core/renderer_opengl/renderer_opengl.h b/src/video_core/renderer_opengl/renderer_opengl.h
index f210190dd..cc19a110f 100755
--- a/src/video_core/renderer_opengl/renderer_opengl.h
+++ b/src/video_core/renderer_opengl/renderer_opengl.h
@@ -38,8 +38,6 @@ class GPU;
 
 namespace OpenGL {
 
-struct VulkanObjects;
-
 /// Structure used for storing information about the textures for the Switch screen
 struct TextureInfo {
     OGLTexture resource;
@@ -101,7 +99,6 @@ private:
     Core::Memory::Memory& cpu_memory;
     Tegra::GPU& gpu;
 
-    std::unique_ptr<VulkanObjects> vulkan_objects;
     Device device;
     StateTracker state_tracker;
     ProgramManager program_manager;
diff --git a/src/video_core/vulkan_common/vulkan_wrapper.cpp b/src/video_core/vulkan_common/vulkan_wrapper.cpp
index 311d3d22d..2aa0ffbe6 100755
--- a/src/video_core/vulkan_common/vulkan_wrapper.cpp
+++ b/src/video_core/vulkan_common/vulkan_wrapper.cpp
@@ -787,7 +787,7 @@ DeviceMemory Device::AllocateMemory(const VkMemoryAllocateInfo& ai) const {
 }
 
 VkMemoryRequirements Device::GetBufferMemoryRequirements(VkBuffer buffer,
-                                                          void* pnext) const noexcept {
+                                                         void* pnext) const noexcept {
     const VkBufferMemoryRequirementsInfo2 info{
         .sType = VK_STRUCTURE_TYPE_BUFFER_MEMORY_REQUIREMENTS_INFO_2,
         .pNext = nullptr,
diff --git a/src/yuzu/configuration/config.cpp b/src/yuzu/configuration/config.cpp
index 5e4d22299..72e69cdfe 100755
--- a/src/yuzu/configuration/config.cpp
+++ b/src/yuzu/configuration/config.cpp
@@ -785,7 +785,7 @@ void Config::ReadRendererValues() {
                       true);
     ReadSettingGlobal(Settings::values.use_vsync, QStringLiteral("use_vsync"), true);
     ReadSettingGlobal(Settings::values.use_assembly_shaders, QStringLiteral("use_assembly_shaders"),
-                      true);
+                      false);
     ReadSettingGlobal(Settings::values.use_asynchronous_shaders,
                       QStringLiteral("use_asynchronous_shaders"), false);
     ReadSettingGlobal(Settings::values.use_fast_gpu_time, QStringLiteral("use_fast_gpu_time"),