Archived
1
0
Fork 0
forked from Mirror/Ryujinx
This repository has been archived on 2024-10-11. You can view files and clone it, but cannot push or open issues or pull requests.
jinx/Ryujinx.Graphics.Gpu/Engine/Compute.cs

155 lines
5.8 KiB
C#
Raw Normal View History

2019-12-29 17:41:50 +00:00
using Ryujinx.Graphics.GAL;
using Ryujinx.Graphics.Gpu.Image;
using Ryujinx.Graphics.Gpu.Shader;
2019-10-13 06:02:07 +00:00
using Ryujinx.Graphics.Gpu.State;
using Ryujinx.Graphics.Shader;
using System;
namespace Ryujinx.Graphics.Gpu.Engine
{
partial class Methods
{
/// <summary>
/// Dispatches compute work.
/// </summary>
/// <param name="state">Current GPU state</param>
/// <param name="argument">Method call argument</param>
2019-11-22 02:46:14 +00:00
public void Dispatch(GpuState state, int argument)
2019-10-13 06:02:07 +00:00
{
uint qmdAddress = (uint)state.Get<int>(MethodOffset.DispatchParamsAddress);
2019-10-13 06:02:07 +00:00
var qmd = _context.MemoryManager.Read<ComputeQmd>((ulong)qmdAddress << 8);
2019-10-13 06:02:07 +00:00
2019-11-22 02:46:14 +00:00
GpuVa shaderBaseAddress = state.Get<GpuVa>(MethodOffset.ShaderBaseAddress);
2019-10-13 06:02:07 +00:00
ulong shaderGpuVa = shaderBaseAddress.Pack() + (uint)qmd.ProgramOffset;
2019-10-13 06:02:07 +00:00
int localMemorySize = qmd.ShaderLocalMemoryLowSize + qmd.ShaderLocalMemoryHighSize;
int sharedMemorySize = Math.Min(qmd.SharedMemorySize, _context.Capabilities.MaximumComputeSharedMemorySize);
for (int index = 0; index < Constants.TotalCpUniformBuffers; index++)
{
if (!qmd.ConstantBufferValid(index))
{
continue;
}
ulong gpuVa = (uint)qmd.ConstantBufferAddrLower(index) | (ulong)qmd.ConstantBufferAddrUpper(index) << 32;
ulong size = (ulong)qmd.ConstantBufferSize(index);
BufferManager.SetComputeUniformBuffer(index, gpuVa, size);
}
ShaderBundle cs = ShaderCache.GetComputeShader(
state,
2019-10-13 06:02:07 +00:00
shaderGpuVa,
qmd.CtaThreadDimension0,
qmd.CtaThreadDimension1,
qmd.CtaThreadDimension2,
localMemorySize,
sharedMemorySize);
2019-10-13 06:02:07 +00:00
2019-12-29 17:41:50 +00:00
_context.Renderer.Pipeline.SetProgram(cs.HostProgram);
2019-11-22 02:46:14 +00:00
var samplerPool = state.Get<PoolState>(MethodOffset.SamplerPoolState);
var texturePool = state.Get<PoolState>(MethodOffset.TexturePoolState);
TextureManager.SetComputeSamplerPool(samplerPool.Address.Pack(), samplerPool.MaximumId, qmd.SamplerIndex);
2019-12-29 17:41:50 +00:00
TextureManager.SetComputeTexturePool(texturePool.Address.Pack(), texturePool.MaximumId);
TextureManager.SetComputeTextureBufferIndex(state.Get<int>(MethodOffset.TextureBufferIndex));
2019-10-13 06:02:07 +00:00
ShaderProgramInfo info = cs.Shaders[0].Info;
2019-10-13 06:02:07 +00:00
for (int index = 0; index < info.CBuffers.Count; index++)
{
BufferDescriptor cb = info.CBuffers[index];
// NVN uses the "hardware" constant buffer for anything that is less than 8,
// and those are already bound above.
// Anything greater than or equal to 8 uses the emulated constant buffers.
// They are emulated using global memory loads.
if (cb.Slot < 8)
{
continue;
}
ulong cbDescAddress = BufferManager.GetComputeUniformBufferAddress(0);
int cbDescOffset = 0x260 + (cb.Slot - 8) * 0x10;
cbDescAddress += (ulong)cbDescOffset;
SbDescriptor cbDescriptor = _context.PhysicalMemory.Read<SbDescriptor>(cbDescAddress);
BufferManager.SetComputeUniformBuffer(cb.Slot, cbDescriptor.PackAddress(), (uint)cbDescriptor.Size);
}
2019-10-13 06:02:07 +00:00
for (int index = 0; index < info.SBuffers.Count; index++)
{
BufferDescriptor sb = info.SBuffers[index];
2019-12-29 17:41:50 +00:00
ulong sbDescAddress = BufferManager.GetComputeUniformBufferAddress(0);
2019-10-13 06:02:07 +00:00
int sbDescOffset = 0x310 + sb.Slot * 0x10;
sbDescAddress += (ulong)sbDescOffset;
SbDescriptor sbDescriptor = _context.PhysicalMemory.Read<SbDescriptor>(sbDescAddress);
2019-10-13 06:02:07 +00:00
Implement lazy flush-on-read for Buffers (SSBO/Copy) (#1790) * Initial implementation of buffer flush (VERY WIP) * Host shaders need to be rebuilt for the SSBO write flag. * New approach with reserved regions and gl sync * Fix a ton of buffer issues. * Remove unused buffer unmapped behaviour * Revert "Remove unused buffer unmapped behaviour" This reverts commit f1700e52fb8760180ac5e0987a07d409d1e70ece. * Delete modified ranges on unmap Fixes potential crashes in Super Smash Bros, where a previously modified range could lie on either side of an unmap. * Cache some more delegates. * Dispose Sync on Close * Also create host sync for GPFifo syncpoint increment. * Copy buffer optimization, add docs * Fix race condition with OpenGL Sync * Enable read tracking on CommandBuffer, insert syncpoint on WaitForIdle * Performance: Only flush individual pages of SSBO at a time This avoids flushing large amounts of data when only a small amount is actually used. * Signal Modified rather than flushing after clear * Fix some docs and code style. * Introduce a new test for tracking memory protection. Sucessfully demonstrates that the bug causing write protection to be cleared by a read action has been fixed. (these tests fail on master) * Address Comments * Add host sync for SetReference This ensures that any indirect draws will correctly flush any related buffer data written before them. Fixes some flashing and misplaced world geometry in MH rise. * Make PageAlign static * Re-enable read tracking, for reads.
2021-01-17 20:08:06 +00:00
BufferManager.SetComputeStorageBuffer(sb.Slot, sbDescriptor.PackAddress(), (uint)sbDescriptor.Size, sb.Flags);
2019-10-13 06:02:07 +00:00
}
BufferManager.SetComputeStorageBufferBindings(info.SBuffers);
BufferManager.SetComputeUniformBufferBindings(info.CBuffers);
2019-10-13 06:02:07 +00:00
var textureBindings = new TextureBindingInfo[info.Textures.Count];
for (int index = 0; index < info.Textures.Count; index++)
{
var descriptor = info.Textures[index];
Target target = ShaderTexture.GetTarget(descriptor.Type);
textureBindings[index] = new TextureBindingInfo(
target,
descriptor.Binding,
descriptor.CbufSlot,
descriptor.HandleIndex,
descriptor.Flags);
}
2019-12-29 17:41:50 +00:00
TextureManager.SetComputeTextures(textureBindings);
var imageBindings = new TextureBindingInfo[info.Images.Count];
for (int index = 0; index < info.Images.Count; index++)
{
var descriptor = info.Images[index];
Target target = ShaderTexture.GetTarget(descriptor.Type);
Format format = ShaderTexture.GetFormat(descriptor.Format);
imageBindings[index] = new TextureBindingInfo(
target,
format,
descriptor.Binding,
descriptor.CbufSlot,
descriptor.HandleIndex,
descriptor.Flags);
}
2019-12-29 17:41:50 +00:00
TextureManager.SetComputeImages(imageBindings);
2019-12-29 17:41:50 +00:00
BufferManager.CommitComputeBindings();
TextureManager.CommitComputeBindings();
2019-10-13 06:02:07 +00:00
2019-12-29 17:41:50 +00:00
_context.Renderer.Pipeline.DispatchCompute(
qmd.CtaRasterWidth,
qmd.CtaRasterHeight,
qmd.CtaRasterDepth);
_forceShaderUpdate = true;
2019-10-13 06:02:07 +00:00
}
}
}