tidus4400/rjx-mirror
1
0
Fork 0
forked from Mirror/Ryujinx
Code Pull requests Activity
rjx-mirror/Ryujinx.Graphics.OpenGL/Program.cs

178 lines
5.1 KiB
C#
Raw Normal View History

Add basic error logging to the GPU
2019-12-28 23:45:33 +00:00
using OpenTK.Graphics.OpenGL;
using Ryujinx.Common.Logging;
Initial work
2019-10-13 06:02:07 +00:00
using Ryujinx.Graphics.GAL;
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
using Ryujinx.Graphics.Shader;
using Ryujinx.Graphics.Shader.Translation;
Initial transform feedback support (#1370) * Initial transform feedback support * Some nits and fixes * Update ReportCounterType and Write method * Can't change shader or TFB bindings while TFB is active * Fix geometry shader input names with new naming
2020-07-15 03:01:10 +00:00
using System;
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
using System.Buffers.Binary;
Initial work
2019-10-13 06:02:07 +00:00
namespace Ryujinx.Graphics.OpenGL
{
class Program : IProgram
{
Log shader compile errors with Warning level (#2617) * Log shader compile errors with Warning level These are infrequent enough that I think it's worth dumping any errors into the log. They also keep causing graphical glitches, and the only indication that anything went wrong is a debug log that is never enabled. * Add maximum length for shader log
2023-02-09 02:50:18 +00:00
private const int MaxShaderLogLength = 2048;
Initial work
2019-10-13 06:02:07 +00:00
public int Handle { get; private set; }
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
public bool IsLinked
{
get
{
if (_status == ProgramLinkStatus.Incomplete)
{
CheckProgramLink(true);
}
return _status == ProgramLinkStatus.Success;
}
}
private ProgramLinkStatus _status = ProgramLinkStatus.Incomplete;
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
private int[] _shaderHandles;
Initial work
2019-10-13 06:02:07 +00:00
Do not allow render targets not explicitly written by the fragment shader to be modified (#3063) * Do not allow render targets not explicitly written by the fragment shader to be modified * Shader cache version bump * Remove blank lines * Avoid redundant color mask updates * HostShaderCacheEntry can be null * Avoid more redundant glColorMask calls * nit: Mask -> Masks * Fix currentComponentMask * More efficient way to update _currentComponentMasks
2022-02-16 22:15:39 +00:00
public bool HasFragmentShader;
public int FragmentOutputMap { get; }
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
public Program(ShaderSource[] shaders, int fragmentOutputMap)
Initial work
2019-10-13 06:02:07 +00:00
{
Handle = GL.CreateProgram();
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
GL.ProgramParameter(Handle, ProgramParameterName.ProgramBinaryRetrievableHint, 1);
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
_shaderHandles = new int[shaders.Length];
Initial work
2019-10-13 06:02:07 +00:00
for (int index = 0; index < shaders.Length; index++)
{
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
ShaderSource shader = shaders[index];
Initial work
2019-10-13 06:02:07 +00:00
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
if (shader.Stage == ShaderStage.Fragment)
Do not allow render targets not explicitly written by the fragment shader to be modified (#3063) * Do not allow render targets not explicitly written by the fragment shader to be modified * Shader cache version bump * Remove blank lines * Avoid redundant color mask updates * HostShaderCacheEntry can be null * Avoid more redundant glColorMask calls * nit: Mask -> Masks * Fix currentComponentMask * More efficient way to update _currentComponentMasks
2022-02-16 22:15:39 +00:00
{
HasFragmentShader = true;
}
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
int shaderHandle = GL.CreateShader(shader.Stage.Convert());
switch (shader.Language)
{
case TargetLanguage.Glsl:
GL.ShaderSource(shaderHandle, shader.Code);
GL.CompileShader(shaderHandle);
break;
case TargetLanguage.Spirv:
GL.ShaderBinary(1, ref shaderHandle, (BinaryFormat)All.ShaderBinaryFormatSpirVArb, shader.BinaryCode, shader.BinaryCode.Length);
GL.SpecializeShader(shaderHandle, "main", 0, (int[])null, (int[])null);
break;
}
GL.AttachShader(Handle, shaderHandle);
_shaderHandles[index] = shaderHandle;
Initial work
2019-10-13 06:02:07 +00:00
}
GL.LinkProgram(Handle);
Do not allow render targets not explicitly written by the fragment shader to be modified (#3063) * Do not allow render targets not explicitly written by the fragment shader to be modified * Shader cache version bump * Remove blank lines * Avoid redundant color mask updates * HostShaderCacheEntry can be null * Avoid more redundant glColorMask calls * nit: Mask -> Masks * Fix currentComponentMask * More efficient way to update _currentComponentMasks
2022-02-16 22:15:39 +00:00
FragmentOutputMap = fragmentOutputMap;
Initial work
2019-10-13 06:02:07 +00:00
}
Do not allow render targets not explicitly written by the fragment shader to be modified (#3063) * Do not allow render targets not explicitly written by the fragment shader to be modified * Shader cache version bump * Remove blank lines * Avoid redundant color mask updates * HostShaderCacheEntry can be null * Avoid more redundant glColorMask calls * nit: Mask -> Masks * Fix currentComponentMask * More efficient way to update _currentComponentMasks
2022-02-16 22:15:39 +00:00
public Program(ReadOnlySpan<byte> code, bool hasFragmentShader, int fragmentOutputMap)
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
{
Handle = GL.CreateProgram();
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
if (code.Length >= 4)
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
{
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
BinaryFormat binaryFormat = (BinaryFormat)BinaryPrimitives.ReadInt32LittleEndian(code.Slice(code.Length - 4, 4));
unsafe
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
{
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
fixed (byte* ptr = code)
{
GL.ProgramBinary(Handle, binaryFormat, (IntPtr)ptr, code.Length - 4);
}
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
}
}
Do not allow render targets not explicitly written by the fragment shader to be modified (#3063) * Do not allow render targets not explicitly written by the fragment shader to be modified * Shader cache version bump * Remove blank lines * Avoid redundant color mask updates * HostShaderCacheEntry can be null * Avoid more redundant glColorMask calls * nit: Mask -> Masks * Fix currentComponentMask * More efficient way to update _currentComponentMasks
2022-02-16 22:15:39 +00:00
HasFragmentShader = hasFragmentShader;
FragmentOutputMap = fragmentOutputMap;
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
}
Initial work
2019-10-13 06:02:07 +00:00
public void Bind()
{
GL.UseProgram(Handle);
}
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
public ProgramLinkStatus CheckProgramLink(bool blocking)
Initial work
2019-10-13 06:02:07 +00:00
{
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
if (!blocking && HwCapabilities.SupportsParallelShaderCompile)
{
GL.GetProgram(Handle, (GetProgramParameterName)ArbParallelShaderCompile.CompletionStatusArb, out int completed);
if (completed == 0)
{
return ProgramLinkStatus.Incomplete;
}
}
Add basic error logging to the GPU
2019-12-28 23:45:33 +00:00
GL.GetProgram(Handle, GetProgramParameterName.LinkStatus, out int status);
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
DeleteShaders();
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
Initial work
2019-10-13 06:02:07 +00:00
if (status == 0)
{
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
_status = ProgramLinkStatus.Failure;
Log shader compile errors with Warning level (#2617) * Log shader compile errors with Warning level These are infrequent enough that I think it's worth dumping any errors into the log. They also keep causing graphical glitches, and the only indication that anything went wrong is a debug log that is never enabled. * Add maximum length for shader log
2023-02-09 02:50:18 +00:00
string log = GL.GetProgramInfoLog(Handle);
if (log.Length > MaxShaderLogLength)
{
log = log.Substring(0, MaxShaderLogLength) + "...";
}
Logger.Warning?.Print(LogClass.Gpu, $"Shader linking failed: \n{log}");
Initial work
2019-10-13 06:02:07 +00:00
}
else
{
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
_status = ProgramLinkStatus.Success;
Initial work
2019-10-13 06:02:07 +00:00
}
Implement parallel host shader cache compilation.
2021-03-29 20:52:25 +00:00
return _status;
Initial work
2019-10-13 06:02:07 +00:00
}
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
public byte[] GetBinary()
{
GL.GetProgram(Handle, (GetProgramParameterName)All.ProgramBinaryLength, out int size);
byte[] data = new byte[size + 4];
GL.GetProgramBinary(Handle, size, out _, out BinaryFormat binFormat, data);
Collapse AsSpan().Slice(..) calls into AsSpan(..) (#3145) * Collapse AsSpan().Slice(..) calls into AsSpan(..) Less code and a bit faster * Collapse an Array.Clear(array, 0, array.Length) call to Array.Clear(array)
2022-02-22 13:32:10 +00:00
BinaryPrimitives.WriteInt32LittleEndian(data.AsSpan(size, 4), (int)binFormat);
Salieri: shader cache (#1701) Here come Salieri, my implementation of a disk shader cache! "I'm sure you know why I named it that." "It doesn't really mean anything." This implementation collects shaders at runtime and cache them to be later compiled when starting a game.
2020-11-12 23:15:34 +00:00
return data;
}
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
private void DeleteShaders()
{
if (_shaderHandles != null)
{
foreach (int shaderHandle in _shaderHandles)
{
GL.DetachShader(Handle, shaderHandle);
GL.DeleteShader(shaderHandle);
}
_shaderHandles = null;
}
}
Initial work
2019-10-13 06:02:07 +00:00
public void Dispose()
{
if (Handle != 0)
{
New shader cache implementation (#3194) * New shader cache implementation * Remove some debug code * Take transform feedback varying count into account * Create shader cache directory if it does not exist + fragment output map related fixes * Remove debug code * Only check texture descriptors if the constant buffer is bound * Also check CPU VA on GetSpanMapped * Remove more unused code and move cache related code * XML docs + remove more unused methods * Better codegen for TransformFeedbackDescriptor.AsSpan * Support migration from old cache format, remove more unused code Shader cache rebuild now also rewrites the shared toc and data files * Fix migration error with BRX shaders * Add a limit to the async translation queue Avoid async translation threads not being able to keep up and the queue growing very large * Re-create specialization state on recompile This might be required if a new version of the shader translator requires more or less state, or if there is a bug related to the GPU state access * Make shader cache more error resilient * Add some missing XML docs and move GpuAccessor docs to the interface/use inheritdoc * Address early PR feedback * Fix rebase * Remove IRenderer.CompileShader and IShader interface, replace with new ShaderSource struct passed to CreateProgram directly * Handle some missing exceptions * Make shader cache purge delete both old and new shader caches * Register textures on new specialization state * Translate and compile shaders in forward order (eliminates diffs due to different binding numbers) * Limit in-flight shader compilation to the maximum number of compilation threads * Replace ParallelDiskCacheLoader state changed event with a callback function * Better handling for invalid constant buffer 1 data length * Do not create the old cache directory structure if the old cache does not exist * Constant buffer use should be per-stage. This change will invalidate existing new caches (file format version was incremented) * Replace rectangle texture with just coordinate normalization * Skip incompatible shaders that are missing texture information, instead of crashing This is required if we, for example, support new texture instruction to the shader translator, and then they allow access to textures that were not accessed before. In this scenario, the old cache entry is no longer usable * Fix coordinates normalization on cubemap textures * Check if title ID is null before combining shader cache path * More robust constant buffer address validation on spec state * More robust constant buffer address validation on spec state (2) * Regenerate shader cache with one stream, rather than one per shader. * Only create shader cache directory during initialization * Logging improvements * Proper shader program disposal * PR feedback, and add a comment on serialized structs * XML docs for RegisterTexture Co-authored-by: riperiperi <rhy3756547@hotmail.com>
2022-04-10 13:49:44 +00:00
DeleteShaders();
Initial work
2019-10-13 06:02:07 +00:00
GL.DeleteProgram(Handle);
Handle = 0;
}
}
}
}
Copy permalink