From bef001c1515a13a9b8e361a5a7d372fc2e01094c Mon Sep 17 00:00:00 2001 From: Takoroka <137927095+Takoroka@users.noreply.github.com> Date: Fri, 30 Oct 2026 15:39:08 -0400 Subject: nx2: read command lists the GPU wrote from the GPU's copy Mario Kart World runs small command lists (20-byte entries) that its culling dispatches fill in each frame. Accurate GPU mode reads every command list through the caches; Fast mode reads guest memory directly (on the split's front thread too), where those entries were still the words of an older frame, so the draws behind them were left out: the trees around the courses vanished in Fast mode only. The buffer cache now also marks the device pages the GPU writes (shader stores, copies of GPU-written data) in a bitmap the split's front can read. A command list on such a page waits for the back to run what was queued before it and is flushed when its words are still only in host buffers. In a race that is ~5% of command lists (13,500 a second of 262,000) and ~400 flushes a second. --- src/video_core/buffer_cache/buffer_cache.h | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'src/video_core/buffer_cache/buffer_cache.h') diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 42d7ba066d..a8f1edc68d 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -17,6 +17,7 @@ #include "video_core/buffer_cache/buffer_cache_base.h" #include "video_core/guest_memory.h" #include "video_core/host1x/gpu_device_memory_manager.h" +#include "video_core/nx2.h" #include "video_core/nx2_profile.h" #include "video_core/texture_cache/util.h" @@ -289,6 +290,9 @@ bool BufferCache

::DMACopy(GPUVAddr src_address, GPUVAddr dest_address, u64 am const bool has_new_downloads = tmp_intervals.size() != 0; for (const auto& pair : tmp_intervals) { gpu_modified_ranges.Add(pair.first, pair.second); + if (Tegra::g_nx2_ampere.load(std::memory_order_relaxed)) { + Tegra::Nx2MarkGpuWritten(pair.first, pair.second); + } } const auto& copy = copies[0]; src_buffer.MarkUsage(copy.src_offset, copy.size); @@ -1606,6 +1610,9 @@ void BufferCache

::MarkWrittenBuffer(BufferId buffer_id, DAddr device_addr, u3 if (AsyncBufferDownloads()) { uncommitted_gpu_modified_ranges.Add(device_addr, size); } + if (Tegra::g_nx2_ampere.load(std::memory_order_relaxed)) { + Tegra::Nx2MarkGpuWritten(device_addr, size); + } } template -- cgit v1.2.3 From 594e07ee5302098e417858608dd03c46336274f5 Mon Sep 17 00:00:00 2001 From: Takoroka <137927095+Takoroka@users.noreply.github.com> Date: Sun, 1 Nov 2026 01:11:09 -0400 Subject: buffer_cache: collect Switch 2 buffers from 512 MiB The collector ran on Switch 2 titles only once their buffers reached 1 GiB (aggressively at 2 GiB), so geometry of scenes long gone stayed in VRAM: Mario Kart World's title screen grew from 0.2 to 0.9 GB of buffers in ten minutes, most of a ~1.4 GB climb in VRAM in use. It now starts at 512 MiB (aggressive at 1 GiB); it only evicts buffers unused for 120 frames or more. (cherry picked from commit fb5d797636f2269591b68fda7c8d65787a9e3a9f) --- src/video_core/buffer_cache/buffer_cache.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) (limited to 'src/video_core/buffer_cache/buffer_cache.h') diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index e90af3214d..53874e75cc 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -141,9 +141,10 @@ void BufferCache

::TickFrame() { // title 1's buffers were ~200 MiB of an 8 GB GPU's ~6.8 GB in use, yet this collector // ran aggressively every frame on the device-wide figure, evicting up to 64 buffers (and // downloading the GPU-written ones, ~5% of the GPU thread) that were recreated later. - // Collect when the buffers themselves are large. - if (buffer_bytes >= 1_GiB) { - RunGarbageCollector(buffer_bytes >= 2_GiB); + // Collect when the buffers themselves are large. At 1 GiB it held geometry of scenes + // long gone: title 1's title screen grew to ~0.9 GB of buffers in ten minutes. + if (buffer_bytes >= 512_MiB) { + RunGarbageCollector(buffer_bytes >= 1_GiB); } } else if (total_used_memory >= minimum_memory) { RunGarbageCollector(total_used_memory >= critical_memory); -- cgit v1.2.3 From 86968cb5ce3195cb131eeb0ef68d2d8a4860f7df Mon Sep 17 00:00:00 2001 From: Takoroka <137927095+Takoroka@users.noreply.github.com> Date: Sun, 1 Nov 2026 08:42:26 -0500 Subject: buffer_cache: mark guest memory large buffer uploads came from cold Like the texture cache's uploads: guest ranges of 256 KiB or more uploaded to the host GPU are mostly geometry the guest rarely reads again, so every 600 frames they are marked cold and the OS pages them out first when it needs RAM. (cherry picked from commit 5c942a5158a0f9769553eeba21b6121f277d0dc5) --- src/video_core/buffer_cache/buffer_cache.h | 11 +++++++++++ src/video_core/buffer_cache/buffer_cache_base.h | 2 ++ 2 files changed, 13 insertions(+) (limited to 'src/video_core/buffer_cache/buffer_cache.h') diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 42d7ba066d..633f8760e4 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -100,6 +100,12 @@ void BufferCache

::TickFrame() { return; } runtime.TickFrame(slot_buffers); + if (frame_tick % 600 == 0 && !uploaded_buffer_ranges.empty()) { + for (const auto& [address, size] : uploaded_buffer_ranges) { + device_memory.HintCold(address, size); + } + uploaded_buffer_ranges.clear(); + } if (static u64 frames{}; ++frames % 600 == 0) { u64 total{}; size_t count{}; @@ -1923,6 +1929,11 @@ void BufferCache

::MappedUploadMemory([[maybe_unused]] Buffer& buffer, u8* const src_pointer = staging_pointer.data() + copy.src_offset; const DAddr device_addr = buffer.CpuAddr() + copy.dst_offset; device_memory.ReadBlockUnsafe(device_addr, src_pointer, copy.size); + // Large uploads are mostly geometry the host GPU now has its own copy of: TickFrame + // marks their guest memory cold (see TextureCache::RefreshContents). + if (copy.size >= 256_KiB) { + uploaded_buffer_ranges.emplace_back(device_addr, copy.size); + } // Apply the staging offset copy.src_offset += upload_staging.offset; } diff --git a/src/video_core/buffer_cache/buffer_cache_base.h b/src/video_core/buffer_cache/buffer_cache_base.h index 45a1836341..0db8bdb1eb 100644 --- a/src/video_core/buffer_cache/buffer_cache_base.h +++ b/src/video_core/buffer_cache/buffer_cache_base.h @@ -549,6 +549,8 @@ private: size_t immediate_buffer_capacity = 0; Common::ScratchBuffer immediate_buffer_alloc; + /// Guest memory large uploads came from since the last cold hint (device address, size). + std::vector> uploaded_buffer_ranges; struct LRUItemParams { using ObjectType = BufferId; -- cgit v1.2.3 From 808feaf3bf34b3c108ff30baae1d2261f216f4f3 Mon Sep 17 00:00:00 2001 From: Duck912 <350528965+Duck912@users.noreply.github.com> Date: Sun, 1 Nov 2026 21:42:19 -0500 Subject: video_core: keep root tables and bind table 0 as constant bank 17 The 3D methods 0x504/0x508 write eight 256-byte root tables, which shaders read as constant banks 24-31; NVN2 keeps push constants in table 0 (bank 17 after the Ampere lowering). Fire Emblem: Fortune's Weave keeps its material constants there and never binds bank 17, and the blit constants pass replaced those reads with render-target sizes, flooding its 3D scenes red. Maxwell3D now keeps the tables, the buffer cache streams table 0 as bank 17 when the game has not bound one, and the blit constants pass leaves shaders that read bank 17 outside its parameter block alone. Assisted-by: Claude Opus 5.5 (debugging, implementation and drafting) Signed-off-by: Duck912 <350528965+Duck912@users.noreply.github.com> --- src/shader_recompiler/ir_opt/position_pass.cpp | 21 +++++++++++++++++++++ src/video_core/buffer_cache/buffer_cache.h | 20 ++++++++++++++++++++ src/video_core/engines/maxwell_3d.cpp | 15 +++++++++++++++ src/video_core/engines/maxwell_3d.h | 12 ++++++++++++ 4 files changed, 68 insertions(+) (limited to 'src/video_core/buffer_cache/buffer_cache.h') diff --git a/src/shader_recompiler/ir_opt/position_pass.cpp b/src/shader_recompiler/ir_opt/position_pass.cpp index 4890346944..23b1200cfc 100644 --- a/src/shader_recompiler/ir_opt/position_pass.cpp +++ b/src/shader_recompiler/ir_opt/position_pass.cpp @@ -83,6 +83,27 @@ void BlitConstantsPass(Environment& env, IR::Program& program) { if (env.ShaderStage() != Stage::VertexB && env.ShaderStage() != Stage::Fragment) { return; } + // Bank 24 is root table 0, which games may load themselves (3D methods 0x504/0x508; the buffer + // cache binds it as bank 17): Fire Emblem: Fortune's Weave keeps its material push constants + // there. A shader reading bank 17 outside the blit block reads such a table and keeps its reads. + const auto in_blit_block = [](u32 offset) { + return offset < 0x20 || (offset >= 0x30 && offset < 0x4c) || (offset >= 0x50 && offset < 0x60); + }; + for (IR::Block* const block : program.post_order_blocks) { + for (IR::Inst& inst : block->Instructions()) { + const IR::Opcode opcode{inst.GetOpcode()}; + if ((opcode != IR::Opcode::GetCbufU32 && opcode != IR::Opcode::GetCbufF32 && + opcode != IR::Opcode::GetCbufU32x2) || + !inst.Arg(0).IsImmediate() || inst.Arg(0).U32() != 17 || !inst.Arg(1).IsImmediate()) { + continue; + } + const u32 offset{inst.Arg(1).U32()}; + const bool pair{opcode == IR::Opcode::GetCbufU32x2}; + if (!in_blit_block(offset) || (pair && !in_blit_block(offset + 4))) { + return; + } + } + } bool used{}; for (IR::Block* const block : program.post_order_blocks) { for (IR::Inst& inst : block->Instructions()) { diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index e90af3214d..2fc74cb6c8 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -994,6 +994,26 @@ void BufferCache

::BindHostGraphicsUniformBuffers(size_t stage) { template void BufferCache

::BindHostGraphicsUniformBuffer(size_t stage, u32 index, u32 binding_index, bool needs_bind) { ++channel_state->uniform_cache_shots[0]; + // Bank 24 (root table 0, NVN2's push constants) is read from bank 17, which such games do not + // bind: stream the table's current words as that buffer. + if (index == 17 && maxwell3d->root_tables.loaded && channel_state->uniform_buffers[stage][index].size == 0) { + const auto& words = maxwell3d->root_tables.tables[0]; + const u32 size = (*channel_state->uniform_buffer_sizes)[stage][index]; + const u32 copied = (std::min)(size, static_cast(sizeof(words))); + if constexpr (IS_OPENGL) { + std::array data{}; + std::memcpy(data.data(), words.data(), copied); + runtime.BindFastUniformBuffer(stage, binding_index, copied); + runtime.PushFastUniformBuffer(stage, binding_index, std::span(data.data(), copied)); + } else { + const std::span span = runtime.BindMappedUniformBuffer(stage, binding_index, size); + std::memset(span.data(), 0, span.size()); + std::memcpy(span.data(), words.data(), copied); + } + channel_state->fast_bound_uniform_buffers[stage] |= 1u << binding_index; + channel_state->uniform_buffer_binding_sizes[stage][binding_index] = size; + return; + } const Binding& binding = channel_state->uniform_buffers[stage][index]; const DAddr device_addr = binding.device_addr; const u32 size = (std::min)(binding.size, (*channel_state->uniform_buffer_sizes)[stage][index]); diff --git a/src/video_core/engines/maxwell_3d.cpp b/src/video_core/engines/maxwell_3d.cpp index 6adfcf89b8..4001284127 100644 --- a/src/video_core/engines/maxwell_3d.cpp +++ b/src/video_core/engines/maxwell_3d.cpp @@ -164,6 +164,8 @@ bool Maxwell3D::IsMethodExecutable(u32 method) { case MAXWELL3D_REG_INDEX(vertex_array_instance_first): case MAXWELL3D_REG_INDEX(vertex_array_instance_subsequent): case MAXWELL3D_REG_INDEX(draw_texture.src_y0): + case RootTables::SelectMethod: + case RootTables::LoadMethod: case MAXWELL3D_REG_INDEX(wait_for_idle): case MAXWELL3D_REG_INDEX(shadow_ram_control): case MAXWELL3D_REG_INDEX(load_mme.instruction_ptr): @@ -346,6 +348,17 @@ void Maxwell3D::ProcessMethodCall(u32 method, u32 argument, u32 nonshadow_argume LOG_DEBUG(HW_GPU, "Native GPU surface clear: value={:#x}", argument); } switch (method) { + case RootTables::SelectMethod: + root_tables.selected = argument & 7; + root_tables.offset = (argument >> 8) & 0xff; + return; + case RootTables::LoadMethod: + if (root_tables.offset < 256 && root_tables.offset % 4 == 0) { + root_tables.tables[root_tables.selected][root_tables.offset / 4] = argument; + root_tables.loaded = true; + } + root_tables.offset += 4; + return; case MAXWELL3D_REG_INDEX(wait_for_idle): return rasterizer->WaitForIdle(); case MAXWELL3D_REG_INDEX(shadow_ram_control): @@ -819,6 +832,8 @@ bool FrontHandledOutsideDrawManager(u32 method) { return true; } switch (method) { + case Maxwell3D::RootTables::SelectMethod: + case Maxwell3D::RootTables::LoadMethod: case MAXWELL3D_REG_INDEX(wait_for_idle): case MAXWELL3D_REG_INDEX(shadow_ram_control): case MAXWELL3D_REG_INDEX(load_mme.instruction_ptr): diff --git a/src/video_core/engines/maxwell_3d.h b/src/video_core/engines/maxwell_3d.h index 0121f1a01c..a32b31fd1e 100644 --- a/src/video_core/engines/maxwell_3d.h +++ b/src/video_core/engines/maxwell_3d.h @@ -3160,6 +3160,18 @@ public: std::array& GetNativeDataRam() { return native_data_ram; } std::span GetNativeMacroPositions() { return native_macro_context ? std::span{native_macro_context->positions} : std::span{macro_positions}; } + /// Root tables: eight 256-byte tables written inline (method 0x504 selects a table, bits 0-2, + /// and a byte offset, bits 8-15; each 0x508 writes the next word) that shaders read as constant + /// banks 24-31. NVN2 keeps push constants in bank 24, which the Ampere lowering reads as bank 17. + struct RootTables { + static constexpr u32 SelectMethod = 0x504 / 4; + static constexpr u32 LoadMethod = 0x508 / 4; + std::array, 8> tables{}; + u32 selected{}; + u32 offset{}; + bool loaded{}; + } root_tables; + /// Write the value to the register identified by method. void CallMethod(Core::System& system, u32 method, u32 method_argument, bool is_last_call) override; -- cgit v1.2.3 From 46b2577bf78870ed85cbb790a9dc3c82cc77f4a7 Mon Sep 17 00:00:00 2001 From: ymatrix1 <10088631+ymatrix1@users.noreply.github.com> Date: Mon, 2 Nov 2026 04:41:15 -0300 Subject: video_core/buffer_cache: fast-guard stage buffer updates and uniform buffer disable (#264) --- src/video_core/buffer_cache/buffer_cache.h | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) (limited to 'src/video_core/buffer_cache/buffer_cache.h') diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index 2fc74cb6c8..e09a435dcc 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -379,6 +379,10 @@ template void BufferCache

::BindGraphicsUniformBuffer(size_t stage, u32 index, GPUVAddr gpu_addr, u32 size) { const std::optional device_addr = gpu_memory->GpuToCpuAddress(gpu_addr); + if (!device_addr) { + channel_state->uniform_buffers[stage][index] = NULL_BINDING; + return; + } const Binding binding{ .device_addr = *device_addr, .size = size, @@ -389,7 +393,11 @@ void BufferCache

::BindGraphicsUniformBuffer(size_t stage, u32 index, GPUVAddr template void BufferCache

::DisableGraphicsUniformBuffer(size_t stage, u32 index) { - channel_state->uniform_buffers[stage][index] = NULL_BINDING; + auto& current = channel_state->uniform_buffers[stage][index]; + if (current.device_addr == 0 && current.size == 0 && !current.buffer_id) [[likely]] { + return; + } + current = NULL_BINDING; } template @@ -1514,6 +1522,9 @@ void BufferCache

::UpdateDrawIndirect() { template void BufferCache

::UpdateUniformBuffers(size_t stage) { + if (channel_state->enabled_uniform_buffer_masks[stage] == 0) [[likely]] { + return; + } ForEachEnabledBit(channel_state->enabled_uniform_buffer_masks[stage], [&](u32 index) { Binding& binding = channel_state->uniform_buffers[stage][index]; if (binding.buffer_id) { @@ -1531,6 +1542,9 @@ void BufferCache

::UpdateUniformBuffers(size_t stage) { template void BufferCache

::UpdateStorageBuffers(size_t stage) { + if (channel_state->enabled_storage_buffers[stage] == 0) [[likely]] { + return; + } ForEachEnabledBit(channel_state->enabled_storage_buffers[stage], [&](u32 index) { // Resolve buffer Binding& binding = channel_state->storage_buffers[stage][index]; @@ -1543,6 +1557,9 @@ void BufferCache

::UpdateStorageBuffers(size_t stage) { template void BufferCache

::UpdateTextureBuffers(size_t stage) { + if (channel_state->enabled_texture_buffers[stage] == 0) [[likely]] { + return; + } ForEachEnabledBit(channel_state->enabled_texture_buffers[stage], [&](u32 index) { Binding& binding = channel_state->texture_buffers[stage][index]; binding.buffer_id = FindBuffer(binding.device_addr, binding.size, false); -- cgit v1.2.3 From a407703f6d29a5ab73328b4ededb12cf29b2758d Mon Sep 17 00:00:00 2001 From: Whovian64 <86082882+whovian64@users.noreply.github.com> Date: Mon, 2 Nov 2026 02:45:39 -0500 Subject: Linux and AMD: BC3 flashing on RADV, ASTC decode stalls, lower RAM usage, a buffer cache crash (#260) * host_memory: punch holes in the backing on Linux instead of zero filling freed memory Freeing guest memory zero filled the memfd backing with memset, which commits every page it touches. fallocate(PUNCH_HOLE) hands the pages back to the kernel instead, and they read as zero afterwards. Only the partial pages at the ends are still cleared by hand. Mario Kart World on Linux (RX 9070 XT, 31 GB): peak RSS 23.1 -> 15.9 GB. EDEN_PUNCH_HOLE=0 restores the memset. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) * fiber: don't zero fill the stacks, untouched pages stay unbacked The value initialised stack arrays were zero filled when each fiber was made, which commits the whole stack. Left uninitialised, the pages a fiber never touches stay unbacked. Together with the descriptor ring change, about 1.2 GB less RSS in Mario Kart World. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) * vulkan: don't zero fill the descriptor update ring Same as the fiber stacks: make_unique value initialised the whole ring for every frame in flight. Entries are always written before they are read, so allocate it without initialising it. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) * vulkan: wait for the GPU after an ASTC transcode decode like the other ASTC decodes TranscodeAstcToBcn (astc_recompression Bc1/Bc3 with GPU decode) was the only ASTC decode recorded with wait = false. On RADV (RX 9070 XT, Mesa) Mario Kart World with BC3 flashed heavily all over the screen, while the uncompressed GPU decode, which waits, did not. CPU decode with BC3 does not flash either. Waiting after the transcode removes the flashing; full pipeline barriers around its copy did not, so the wait hides a race somewhere else rather than fixing one in the transcode itself. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) * vulkan: submit after an ASTC decode instead of waiting for the GPU by default Every ASTC decode ended with scheduler.Finish(). Over a 19 minute Mario Kart World Grand Prix that stalled the GPU thread 30.7 s in total, up to 1.2 s out of every 5 s while new textures stream in (track sections, race starts), which is where the hitches were. Flushing keeps the ordering that stops the BC3 flashing on RADV without the CPU waiting: no flashing in play, and frames over 100 ms went from dozens to 3 in about 7 minutes. EDEN_NX2_ASTC_FLUSH=0 restores the wait. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) * texture_cache: over the expected use, evict costly images unused for 1800 frames first The collector never deletes CostlyLoad images outside aggressive mode, so once a cup moves on, the previous tracks' textures stay resident and the cache sits over its expected use deleting the cheap images still in use. When over the expected use, costly images unused for 1800 frames (half a minute) now go first. Mario Kart World Grand Prix on an RX 9070 XT: frames over 20 ms 8.8% -> 4.0%, 58.7 fps average. EDEN_NX2_TC_OLD_AGE= sets the age, 0 turns it off. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) * buffer_cache: no buffer for device addresses past the page table FindBuffer indexed page_table, which covers 2^36 bytes, with any address. Mario Kart World crashed after about 4.5 minutes with a uniform buffer at 0x72253b2f2890 (FindBuffer <- UpdateUniformBuffers <- DrawEnd from a macro), reading far past the table. Treat such an address as no buffer. Where that address comes from is still unknown; this only stops the crash. Assisted-by: Claude Opus 5.5 (debugging, profiling and code drafts) --- src/common/fiber.cpp | 4 ++-- src/common/host_memory.cpp | 22 ++++++++++++++++++++++ src/video_core/buffer_cache/buffer_cache.h | 4 ++-- src/video_core/renderer_vulkan/vk_compute_pass.cpp | 6 +++--- .../renderer_vulkan/vk_texture_cache.cpp | 2 +- .../renderer_vulkan/vk_update_descriptor.cpp | 8 ++++---- .../renderer_vulkan/vk_update_descriptor.h | 3 ++- src/video_core/texture_cache/texture_cache.h | 16 ++++++++++++++-- 8 files changed, 50 insertions(+), 15 deletions(-) (limited to 'src/video_core/buffer_cache/buffer_cache.h') diff --git a/src/common/fiber.cpp b/src/common/fiber.cpp index 0e7f28c5b9..08558a0026 100644 --- a/src/common/fiber.cpp +++ b/src/common/fiber.cpp @@ -25,8 +25,8 @@ struct Fiber::FiberImpl { FiberImpl() {} u32 canary_1 = CANARY_VALUE; - std::array stack{}; - std::array rewind_stack{}; + std::array stack; + std::array rewind_stack; u32 canary_2 = CANARY_VALUE; boost::context::detail::fcontext_t context{}; diff --git a/src/common/host_memory.cpp b/src/common/host_memory.cpp index 75a9da80e9..147eba8477 100644 --- a/src/common/host_memory.cpp +++ b/src/common/host_memory.cpp @@ -51,6 +51,7 @@ #endif // ^^^ POSIX ^^^ +#include #include #include @@ -517,6 +518,14 @@ public: , virtual_size{virtual_size_} {} +#ifdef __linux__ + bool PunchHole(size_t offset, size_t length) { + return fd >= 0 && + fallocate(fd, FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE, static_cast(offset), + static_cast(length)) == 0; + } +#endif + bool Init() { long page_size = sysconf(_SC_PAGESIZE); ASSERT_MSG(page_size == 0x1000, "page size {:#x} is incompatible with 4K paging", page_size); @@ -791,6 +800,19 @@ void HostMemory::Protect(size_t virtual_offset, size_t length, MemoryPermission } void HostMemory::ClearBackingRegion(size_t physical_offset, size_t length, u32 fill_value) { +#ifdef __linux__ + const size_t first = AlignUp(physical_offset, PageAlignment); + const size_t last = AlignDown(physical_offset + length, PageAlignment); + static const bool punch = [] { + const char* value = std::getenv("EDEN_PUNCH_HOLE"); + return value == nullptr || value[0] != '0'; + }(); + if (punch && fill_value == 0 && impl && last > first && impl->PunchHole(first, last - first)) { + std::memset(backing_base + physical_offset, 0, first - physical_offset); + std::memset(backing_base + last, 0, physical_offset + length - last); + return; + } +#endif std::memset(backing_base + physical_offset, fill_value, length); } diff --git a/src/video_core/buffer_cache/buffer_cache.h b/src/video_core/buffer_cache/buffer_cache.h index e09a435dcc..a3358048dc 100644 --- a/src/video_core/buffer_cache/buffer_cache.h +++ b/src/video_core/buffer_cache/buffer_cache.h @@ -1651,10 +1651,10 @@ void BufferCache

::MarkWrittenBuffer(BufferId buffer_id, DAddr device_addr, u3 template BufferId BufferCache

::FindBuffer(DAddr device_addr, u32 size, bool sparse_compatible) { - if (device_addr == 0) { + const u64 page = device_addr >> CACHING_PAGEBITS; + if (device_addr == 0 || page >= page_table.size()) { return NULL_BUFFER_ID; } - const u64 page = device_addr >> CACHING_PAGEBITS; const BufferId buffer_id = page_table[page]; if (buffer_id) { Buffer& buffer = slot_buffers[buffer_id]; diff --git a/src/video_core/renderer_vulkan/vk_compute_pass.cpp b/src/video_core/renderer_vulkan/vk_compute_pass.cpp index 691be728d2..cb8dcde42f 100644 --- a/src/video_core/renderer_vulkan/vk_compute_pass.cpp +++ b/src/video_core/renderer_vulkan/vk_compute_pass.cpp @@ -801,11 +801,11 @@ void ASTCDecoderPass::AssembleInto(const VideoCommon::ImageInfo& info, u32 skip, cmdbuf.PipelineBarrier(VK_PIPELINE_STAGE_COMPUTE_SHADER_BIT, vk::PIPELINE_STAGE_GRAPHICS_COMPUTE, 0, image_barrier); }); - // Each decode ends with the GPU thread waiting for the host GPU to go idle. Measured here - // (count and time per 5 s); EDEN_NX2_ASTC_FLUSH=1 submits without waiting instead. + // Each decode ends with a submit to the host GPU. Measured here + // (count and time per 5 s); EDEN_NX2_ASTC_FLUSH=0 waits for the host GPU to go idle instead. static const bool flush_only = [] { const char* value = std::getenv("EDEN_NX2_ASTC_FLUSH"); - return value != nullptr && value[0] == '1'; + return value == nullptr || value[0] != '0'; }(); if (!target.wait) { return; diff --git a/src/video_core/renderer_vulkan/vk_texture_cache.cpp b/src/video_core/renderer_vulkan/vk_texture_cache.cpp index 4e428229f6..b0e5eacb32 100644 --- a/src/video_core/renderer_vulkan/vk_texture_cache.cpp +++ b/src/video_core/renderer_vulkan/vk_texture_cache.cpp @@ -4267,7 +4267,7 @@ void TextureCacheRuntime::TranscodeAstcToBcn( .aspect_mask = VK_IMAGE_ASPECT_COLOR_BIT, .initialized = false, .level_view = [&pending](s32 level) { return *pending.views[level]; }, - .wait = false, + .wait = true, }, map, swizzles); diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp index 742a8d9969..29f3ee71a1 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.cpp +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.cpp @@ -20,10 +20,10 @@ UpdateDescriptorQueue::UpdateDescriptorQueue(const Device& device_, size_t frame bool supports_descriptor_buffer_) : device{device_}, frame_payload_size{frame_payload_size_}, supports_descriptor_buffer{supports_descriptor_buffer_}, - payload(frame_payload_size_ * FRAMES_IN_FLIGHT) + payload(std::make_unique_for_overwrite(frame_payload_size_ * FRAMES_IN_FLIGHT)) { - payload_start = payload.data(); - payload_cursor = payload.data(); + payload_start = payload.get(); + payload_cursor = payload.get(); } UpdateDescriptorQueue::~UpdateDescriptorQueue() = default; @@ -32,7 +32,7 @@ void UpdateDescriptorQueue::TickFrame() { if (++frame_index >= FRAMES_IN_FLIGHT) { frame_index = 0; } - payload_start = payload.data() + frame_index * frame_payload_size; + payload_start = payload.get() + frame_index * frame_payload_size; payload_cursor = payload_start; } diff --git a/src/video_core/renderer_vulkan/vk_update_descriptor.h b/src/video_core/renderer_vulkan/vk_update_descriptor.h index 9f30e65dbd..4d393376b4 100644 --- a/src/video_core/renderer_vulkan/vk_update_descriptor.h +++ b/src/video_core/renderer_vulkan/vk_update_descriptor.h @@ -6,6 +6,7 @@ #pragma once +#include #include #include #include "video_core/vulkan_common/vulkan_wrapper.h" @@ -122,7 +123,7 @@ private: DescriptorUpdateEntry* payload_cursor = nullptr; DescriptorUpdateEntry* payload_start = nullptr; const DescriptorUpdateEntry* upload_start = nullptr; - std::vector payload; + std::unique_ptr payload; }; // TODO: should these be separate classes instead? diff --git a/src/video_core/texture_cache/texture_cache.h b/src/video_core/texture_cache/texture_cache.h index 32775b8719..55ec4c64df 100644 --- a/src/video_core/texture_cache/texture_cache.h +++ b/src/video_core/texture_cache/texture_cache.h @@ -125,6 +125,7 @@ template void TextureCache

::RunGarbageCollector() { bool high_priority_mode = false; bool aggressive_mode = false; + bool evict_old_costly = false; u64 ticks_to_destroy = 0; size_t num_iterations = 0; const auto Configure = [&](bool allow_aggressive) { @@ -139,7 +140,8 @@ void TextureCache

::RunGarbageCollector() { ticks_to_destroy = (std::max)(ticks_to_destroy, min_age); num_iterations = aggressive_mode ? 40 : (high_priority_mode ? 20 : 10); }; - const auto Cleanup = [this, &num_iterations, &high_priority_mode, &aggressive_mode](ImageId image_id) { + const auto Cleanup = [this, &num_iterations, &high_priority_mode, &aggressive_mode, + &evict_old_costly](ImageId image_id) { if (num_iterations == 0) { return true; } @@ -149,7 +151,8 @@ void TextureCache

::RunGarbageCollector() { return false; } const bool must_download = IsDownloadable(image) && False(image.flags & ImageFlagBits::BadOverlap); - if ((!aggressive_mode && True(image.flags & ImageFlagBits::CostlyLoad)) || (!high_priority_mode && must_download)) { + if ((!aggressive_mode && !evict_old_costly && True(image.flags & ImageFlagBits::CostlyLoad)) || + (!high_priority_mode && must_download)) { return false; } if (must_download) { @@ -176,6 +179,15 @@ void TextureCache

::RunGarbageCollector() { return false; }; Configure(false); + static const u64 old_age = [] { + const char* value = std::getenv("EDEN_NX2_TC_OLD_AGE"); + return value != nullptr ? std::strtoull(value, nullptr, 10) : 1800ULL; + }(); + if (high_priority_mode && old_age != 0 && frame_tick > old_age) { + evict_old_costly = true; + lru_cache.ForEachItemBelow(frame_tick - old_age, Cleanup); + evict_old_costly = false; + } lru_cache.ForEachItemBelow(frame_tick - ticks_to_destroy, Cleanup); if (total_used_memory >= critical_memory) { Configure(true); -- cgit v1.2.3