JinxRyu/Ryujinx.Graphics.Gpu/Shader/HashTable/HashState.cs

114 lines
3.6 KiB
C#
Raw Normal View History

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 System;
using System.Runtime.InteropServices;
namespace Ryujinx.Graphics.Gpu.Shader.HashTable
{
/// <summary>
/// State of a hash calculation.
/// </summary>
struct HashState
{
// This is using a slightly modified implementation of FastHash64.
// Reference: https://github.com/ztanml/fast-hash/blob/master/fasthash.c
private const ulong M = 0x880355f21e6d1965UL;
private ulong _hash;
private int _start;
/// <summary>
/// One shot hash calculation for a given data.
/// </summary>
/// <param name="data">Data to be hashed</param>
/// <returns>Hash of the given data</returns>
public static uint CalcHash(ReadOnlySpan<byte> data)
{
HashState state = new HashState();
state.Initialize();
state.Continue(data);
return state.Finalize(data);
}
/// <summary>
/// Initializes the hash state.
/// </summary>
public void Initialize()
{
_hash = 23;
}
/// <summary>
/// Calculates the hash of the given data.
/// </summary>
/// <remarks>
/// The full data must be passed on <paramref name="data"/>.
/// If this is not the first time the method is called, then <paramref name="data"/> must start with the data passed on the last call.
/// If a smaller slice of the data was already hashed before, only the additional data will be hashed.
/// This can be used for additive hashing of data in chuncks.
/// </remarks>
/// <param name="data">Data to be hashed</param>
public void Continue(ReadOnlySpan<byte> data)
{
ulong h = _hash;
ReadOnlySpan<ulong> dataAsUlong = MemoryMarshal.Cast<byte, ulong>(data.Slice(_start));
for (int i = 0; i < dataAsUlong.Length; i++)
{
ulong value = dataAsUlong[i];
h ^= Mix(value);
h *= M;
}
_hash = h;
_start = data.Length & ~7;
}
/// <summary>
/// Performs the hash finalization step, and returns the calculated hash.
/// </summary>
/// <remarks>
/// The full data must be passed on <paramref name="data"/>.
/// <paramref name="data"/> must start with the data passed on the last call to <see cref="Continue"/>.
/// No internal state is changed, so one can still continue hashing data with <see cref="Continue"/>
/// after calling this method.
/// </remarks>
/// <param name="data">Data to be hashed</param>
/// <returns>Hash of all the data hashed with this <see cref="HashState"/></returns>
public uint Finalize(ReadOnlySpan<byte> data)
{
ulong h = _hash;
int remainder = data.Length & 7;
if (remainder != 0)
{
ulong v = 0;
for (int i = data.Length - remainder; i < data.Length; i++)
{
v |= (ulong)data[i] << ((i - remainder) * 8);
}
h ^= Mix(v);
h *= M;
}
h = Mix(h);
return (uint)(h - (h >> 32));
}
/// <summary>
/// Hash mix function.
/// </summary>
/// <param name="h">Hash to mix</param>
/// <returns>Mixed hash</returns>
private static ulong Mix(ulong h)
{
h ^= h >> 23;
h *= 0x2127599bf4325c37UL;
h ^= h >> 47;
return h;
}
}
}