youmuliu/stgsharp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

use modular initing system rather than calling Init() method.

Browse Source
This commit is contained in:
Youmu Liu
2025-11-11 18:22:28 +08:00
parent 678a846e54
commit 44bf78b937
65 changed files with 1328 additions and 1814 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.Mathematics/Numeric/MatrixParallelHandle.cs
+134 -162
View File
@@ -1,9 +1,9 @@
//-----------------------------------------------------------------------
// -----------------------------------------------------------------------
// file="MatrixParallelHandle"
// file="MatrixParallelPool"
// Project: StgSharp
// AuthorGroup: Nitload
// Copyright (c) Nitload. All rights reserved.
// AuthorGroup: Nitload Space
// Copyright (c) Nitload Space. All rights reserved.
//
// Permission is hereby granted, free of charge, to any person obtaining a copy
// of this software and associated documentation files (the "Software"), to deal
@@ -25,181 +25,153 @@
//
// -----------------------------------------------------------------------
// -----------------------------------------------------------------------
using StgSharp.HighPerformance;
using StgSharp.Collections;
using System;
using System.Numerics;
using System.Collections.Generic;
using System.Diagnostics.CodeAnalysis;
using System.Linq;
using System.Runtime.CompilerServices;
using System.Runtime.InteropServices;
using System.Text;
using System.Threading.Tasks;
namespace StgSharp.Mathematics.Numeric
{
[StructLayout(LayoutKind.Sequential)]
internal unsafe struct MatrixParallelTaskPackageNonGeneric
public struct MatrixParallelHandle : IDisposable
{
[MethodImpl(MethodImplOptions.AggressiveInlining)]
public static unsafe void Copy(
MatrixParallelTaskPackageNonGeneric* source,
MatrixParallelTaskPackageNonGeneric* target)
private static readonly (int a, int b)[] LookupTable = [ (1, 1), // 7
(0, 2), // 8 (no non-zero solution)
(3, 0), // 9 (no non-zero solution)
(2, 1), // 10
(1, 2), // 11
(0, 3), // 12 (no non-zero solution)
(3, 1), // 13
(2, 2), // 14
(1, 3), // 15
(4, 1), // 16
(3, 2), // 17
(2, 3), // 18
(1, 4), // 19
(4, 2), // 20
];
private MatrixParallelThread[] _threads;
private bool _disposed = false;
private CapacityFixedStack<MatrixParallelWrap> _wraps;
internal MatrixParallelHandle(MatrixParallelThread[] threadArray)
{
Buffer.MemoryCopy(source, target, sizeof(MatrixParallelTaskPackageNonGeneric),
sizeof(MatrixParallelTaskPackageNonGeneric));
_threads = threadArray;
IsMultithread = threadArray.Length > 1;
SliceToWraps();
}
#region source
// Group 1: Left/Right matrix kernel base address
public IntPtr Left; // 8
public IntPtr Right; // 8
#endregion
#region ans
// Group 2: Result kernel base address + reserved
public IntPtr Result; // 8
public int ElementSize;
public int ReservedPtr0; // 4 (extension/reserved)
#endregion
#region left enum
public int LeftPrimOffset; // 4
public int LeftPrimStride; // 4
public int LeftSecOffset; // 4
public int LeftSecStride; // 4
#endregion
#region right enum
public int RightPrimOffset; // 4
public int RightPrimStride; // 4
public int RightSecOffset; // 4
public int RightSecStride; // 4
#endregion
#region ans enum
public int ResultPrimOffset; // 4
public int ResultPrimStride; // 4
public int ResultSecOffset; // 4
public int ResultSecStride; // 4
#endregion
#region scalar/compute
public ScalarPacket* Scalar; // 8
public IntPtr ComputeHandle; // 8
#endregion
#region global profile
public int PrimCount; // 4
public int ComputeMode; // 4
public int SecCount; // 4
private int ReservedInt0; // 4
#endregion
#region count offset
public int PrimTileOffset; // 4
public int SecTileOffset; // 4
public int PrimCountInTile; // 4
public int SecCountInTile; // 4
#endregion
}
[StructLayout(LayoutKind.Sequential)]
internal unsafe struct MatrixParallelTaskPackage<T> where T: unmanaged, INumber<T>
{
public MatrixParallelTaskPackage()
public ReadOnlySpan<MatrixParallelThread> Threads
{
Unsafe.SkipInit(out this);
get
{
if (_disposed) {
throw new ObjectDisposedException(typeof(MatrixParallelHandle).Name);
}
return _threads;
}
}
internal IEnumerable<MatrixParallelWrap> Wraps
{
get
{
if (_disposed) {
throw new ObjectDisposedException(typeof(MatrixParallelHandle).Name);
}
return _wraps;
}
}
internal ReadOnlySpan<MatrixParallelThread> ThreadsUnsafe
{
[MethodImpl(MethodImplOptions.AggressiveInlining)]
get => _threads;
}
private bool IsMultithread { get; init; }
public void Dispose()
{
if (_disposed) {
return;
}
MatrixParallel.ReturnParallelResources(ref _threads);
_disposed = true;
}
public MatrixParallelThread GetLeader()
{
Threads[0].Role = 2;
return Threads[0];
}
[MethodImpl(MethodImplOptions.AggressiveInlining)]
public static unsafe void Copy(MatrixParallelTaskPackage<T>* source, MatrixParallelTaskPackage<T>* target)
private static (int count_3, int count_4) FastDecompose(int c)
{
Buffer.MemoryCopy(source, target, sizeof(MatrixParallelTaskPackage<T>),
sizeof(MatrixParallelTaskPackage<T>));
/*
* Thread wrapping to accelerate matrix operations by dividing threads into groups (pool).
* For M pool, each containing an average of N threads, we satisfy the condition M * N = T - 4,
* where T is the total thread _count of the CPU. Typically, for modern consumer CPUs, T is around 16.
* For simplicity, let M * N = 16.
*
* The total time complexity for publishing tasks is O(M + N):
* - O(M) accounts for distributing tasks across M pool.
* - O(N) accounts for threads within each wrap fetching tasks from the wrap queue.
*
* Using the inequality Sqrt(M * N) <= (M + N) / 2 (derived from the arithmetic-geometric mean inequality),
* the optimal configuration occurs when M = N. For M * N = 16, this gives M = N = 4.
*
* To ensure compatibility across a wide range of CPU architectures,
* we allow each wrap to contain either 3 or 4 threads, providing flexibility while maintaining efficiency.
*/
if (c < 7)
{
return (0, 0);
}
int t = c - 7;
int k = t / 14;
int d = t % 14;
(int a0, int b0) = LookupTable[d];
return (a0 + (2 * k), b0 + (2 * k));
}
#region source
private void SliceToWraps()
{
if (_disposed) {
throw new ObjectDisposedException(typeof(MatrixParallelHandle).Name);
}
if (_wraps is not null) {
return;
}
int count = _threads.Length;
if (count < 6)
{
MatrixParallelWrap wrap = new MatrixParallelWrap(_threads, 0, count);
_wraps = [ wrap ];
}
(int c3,int c4) = FastDecompose(_threads.Length);
_wraps = new(c3 + c4);
int begin = 0;
for (int i = 0; i < c3; i++)
{
_wraps.Push(new MatrixParallelWrap(_threads, begin, 3));
begin += 3;
}
for (int i = 0; i < c4; i++)
{
_wraps.Push(new MatrixParallelWrap(_threads, begin, 4));
begin += 4;
}
}
// Group 1: Left/Right matrix kernel base address
public MatrixKernel<T>* Left; // 8
public MatrixKernel<T>* Right; // 8
#endregion
#region ans
// Group 2: Result kernel base address + reserved
public MatrixKernel<T>* Result; // 8
public int ElementSize = sizeof(T);
public int ReservedPtr0; // 4 (extension/reserved)
#endregion
#region left enum
public int LeftPrimOffset; // 4
public int LeftPrimStride; // 4
public int LeftSecOffset; // 4
public int LeftSecStride; // 4
#endregion
#region right enum
public int RightPrimOffset; // 4
public int RightPrimStride; // 4
public int RightSecOffset; // 4
public int RightSecStride; // 4
#endregion
#region ans enum
public int ResultPrimOffset; // 4
public int ResultPrimStride; // 4
public int ResultSecOffset; // 4
public int ResultSecStride; // 4
#endregion
#region scalar/compute
public ScalarPacket* Scalar; // 8
public IntPtr ComputeHandle; // 8
#endregion
#region global profile
public int PrimCount; // 4
public int ComputeMode; // 4
public int SecCount; // 4
private readonly int ReservedInt0; // 4
#endregion
#region count offset
public int PrimTileOffset; // 4
public int SecTileOffset; // 4
public int PrimColumnCountInTile; // 4
public int SecColumnCountInTile; // 4
#endregion
}
}
}