From 9764c13d6d2977903f407761b27d847c0056e1c4 Mon Sep 17 00:00:00 2001 From: ReinUsesLisp Date: Wed, 30 Dec 2020 02:25:23 -0300 Subject: video_core: Rewrite the texture cache The current texture cache has several points that hurt maintainability and performance. It's easy to break unrelated parts of the cache when doing minor changes. The cache can easily forget valuable information about the cached textures by CPU writes or simply by its normal usage.The current texture cache has several points that hurt maintainability and performance. It's easy to break unrelated parts of the cache when doing minor changes. The cache can easily forget valuable information about the cached textures by CPU writes or simply by its normal usage. This commit aims to address those issues. --- src/video_core/texture_cache/util.h | 107 ++++++++++++++++++++++++++++++++++++ 1 file changed, 107 insertions(+) create mode 100644 src/video_core/texture_cache/util.h (limited to 'src/video_core/texture_cache/util.h') diff --git a/src/video_core/texture_cache/util.h b/src/video_core/texture_cache/util.h new file mode 100644 index 000000000..dbbbd33cd --- /dev/null +++ b/src/video_core/texture_cache/util.h @@ -0,0 +1,107 @@ +// Copyright 2020 yuzu Emulator Project +// Licensed under GPLv2 or any later version +// Refer to the license.txt file included. + +#pragma once + +#include +#include + +#include "common/common_types.h" + +#include "video_core/engines/maxwell_3d.h" +#include "video_core/surface.h" +#include "video_core/texture_cache/image_base.h" +#include "video_core/texture_cache/image_view_base.h" +#include "video_core/texture_cache/types.h" +#include "video_core/textures/texture.h" + +namespace VideoCommon { + +using Tegra::Texture::TICEntry; + +struct OverlapResult { + GPUVAddr gpu_addr; + VAddr cpu_addr; + SubresourceExtent resources; +}; + +[[nodiscard]] u32 CalculateGuestSizeInBytes(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateUnswizzledSizeBytes(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateConvertedSizeBytes(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateLayerStride(const ImageInfo& info) noexcept; + +[[nodiscard]] u32 CalculateLayerSize(const ImageInfo& info) noexcept; + +[[nodiscard]] std::array CalculateMipLevelOffsets( + const ImageInfo& info) noexcept; + +[[nodiscard]] std::vector CalculateSliceOffsets(const ImageInfo& info); + +[[nodiscard]] std::vector CalculateSliceSubresources(const ImageInfo& info); + +[[nodiscard]] u32 CalculateLevelStrideAlignment(const ImageInfo& info, u32 level); + +[[nodiscard]] VideoCore::Surface::PixelFormat PixelFormatFromTIC( + const Tegra::Texture::TICEntry& config) noexcept; + +[[nodiscard]] ImageViewType RenderTargetImageViewType(const ImageInfo& info) noexcept; + +[[nodiscard]] std::vector MakeShrinkImageCopies(const ImageInfo& dst, + const ImageInfo& src, + SubresourceBase base); + +[[nodiscard]] bool IsValidAddress(const Tegra::MemoryManager& gpu_memory, const TICEntry& config); + +[[nodiscard]] std::vector UnswizzleImage(Tegra::MemoryManager& gpu_memory, + GPUVAddr gpu_addr, const ImageInfo& info, + std::span output); + +[[nodiscard]] BufferCopy UploadBufferCopy(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, + const ImageBase& image, std::span output); + +void ConvertImage(std::span input, const ImageInfo& info, std::span output, + std::span copies); + +[[nodiscard]] std::vector FullDownloadCopies(const ImageInfo& info); + +[[nodiscard]] Extent3D MipSize(Extent3D size, u32 level); + +[[nodiscard]] Extent3D MipBlockSize(const ImageInfo& info, u32 level); + +[[nodiscard]] std::vector FullUploadSwizzles(const ImageInfo& info); + +void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const ImageInfo& info, + std::span copies, std::span memory); + +[[nodiscard]] bool IsBlockLinearSizeCompatible(const ImageInfo& new_info, + const ImageInfo& overlap_info, u32 new_level, + u32 overlap_level, bool strict_size) noexcept; + +[[nodiscard]] bool IsPitchLinearSameSize(const ImageInfo& lhs, const ImageInfo& rhs, + bool strict_size) noexcept; + +[[nodiscard]] std::optional ResolveOverlap(const ImageInfo& new_info, + GPUVAddr gpu_addr, VAddr cpu_addr, + const ImageBase& overlap, + bool strict_size); + +[[nodiscard]] bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs); + +[[nodiscard]] std::optional FindSubresource(const ImageInfo& candidate, + const ImageBase& image, + GPUVAddr candidate_addr, + RelaxedOptions options); + +[[nodiscard]] bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, + GPUVAddr candidate_addr, RelaxedOptions options); + +void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst, + const ImageBase* src); + +[[nodiscard]] u32 MapSizeBytes(const ImageBase& image); + +} // namespace VideoCommon -- cgit v1.2.3 From 7d904fef2e6ac9ce8a3df71e758a36d39b8f69e5 Mon Sep 17 00:00:00 2001 From: ReinUsesLisp Date: Mon, 4 Jan 2021 01:56:44 -0300 Subject: gl_texture_cache: Avoid format views on Intel and AMD Intel and AMD proprietary drivers are incapable of rendering to texture views of different formats than the original texture. Avoid creating these at a cache level. This will consume more memory, emulating them with copies. --- src/video_core/texture_cache/util.h | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'src/video_core/texture_cache/util.h') diff --git a/src/video_core/texture_cache/util.h b/src/video_core/texture_cache/util.h index dbbbd33cd..52a9207d6 100644 --- a/src/video_core/texture_cache/util.h +++ b/src/video_core/texture_cache/util.h @@ -87,17 +87,19 @@ void SwizzleImage(Tegra::MemoryManager& gpu_memory, GPUVAddr gpu_addr, const Ima [[nodiscard]] std::optional ResolveOverlap(const ImageInfo& new_info, GPUVAddr gpu_addr, VAddr cpu_addr, const ImageBase& overlap, - bool strict_size); + bool strict_size, bool broken_views); [[nodiscard]] bool IsLayerStrideCompatible(const ImageInfo& lhs, const ImageInfo& rhs); [[nodiscard]] std::optional FindSubresource(const ImageInfo& candidate, const ImageBase& image, GPUVAddr candidate_addr, - RelaxedOptions options); + RelaxedOptions options, + bool broken_views); [[nodiscard]] bool IsSubresource(const ImageInfo& candidate, const ImageBase& image, - GPUVAddr candidate_addr, RelaxedOptions options); + GPUVAddr candidate_addr, RelaxedOptions options, + bool broken_views); void DeduceBlitImages(ImageInfo& dst_info, ImageInfo& src_info, const ImageBase* dst, const ImageBase* src); -- cgit v1.2.3