threadgroup_bitonic_sort_hlsl/fbuild.bff at main · pm4rtx/threadgroup_bitonic_sort_hlsl · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
//------------------------------------------------------------------------------
//  FASTBuild Configuration
//
//  Compiles 60 HLSL compute shader permutations using DXC in parallel:
//      - 30 (KernelSizeLog2, TGroupSizeLog2) combinations
//      - x2 variants: with/without wave intrinsics
//
//  Usage:
//      tools\FBuild.exe all
//------------------------------------------------------------------------------

Settings
{
    .CachePath = ''
}

//------------------------------------------------------------------------------
//  DXC Compiler (imported from DXC environment variable, set by build.bat)
//------------------------------------------------------------------------------
#import DXC

//------------------------------------------------------------------------------
//  Common flags (matching build.bat dxc_flags_opt + dxc_compile_cs)
//------------------------------------------------------------------------------
.DxcCommonFlags = '-E main -T cs_6_0 -O3 -HV 2021 -Zs -Qstrip_debug'

.OutputDir = '.build_shaders'

//------------------------------------------------------------------------------
//  Header dependencies to make sure re-compilation triggers when those are
//  modified
//------------------------------------------------------------------------------
.ShaderDeps = {
    'threadgroup_bitonic_sort.hlsl'
    'threadgroup_bitonic_sort.hlsli'
}

//------------------------------------------------------------------------------
//  Shader permutations -- parallel arrays for (KernelSizeLog2, TGroupSizeLog2)
//  Generated from: for i in 1..10, j in 0..2: K=i+j, T=i
//------------------------------------------------------------------------------
.KernelSizes = {
    '1','2','3',    // T=1
    '2','3','4',    // T=2
    '3','4','5',    // T=3
    '4','5','6',    // T=4
    '5','6','7',    // T=5
    '6','7','8',    // T=6
    '7','8','9',    // T=7
    '8','9','10',   // T=8
    '9','10','11',  // T=9
    '10','11','12'  // T=10
}
.TGroupSizes = {
    '1','1','1',
    '2','2','2',
    '3','3','3',
    '4','4','4',
    '5','5','5',
    '6','6','6',
    '7','7','7',
    '8','8','8',
    '9','9','9',
    '10','10','10'
}

//------------------------------------------------------------------------------
//  Wave intrinsic variants
//------------------------------------------------------------------------------
.WaveIds     = { '0', '1' }
.WaveDefines = { '',  ' -DNO_WAVE_INTRINSICS=1' }

//------------------------------------------------------------------------------
//  Generate Exec targets: 2 variants per (K, T) combination
//  Outer loop: (KernelSizeLog2, TGroupSizeLog2) — 30 combinations
//  Inner loop: wave/no-wave — 2 variants
//------------------------------------------------------------------------------
.AllShaderTargets = {}

ForEach( .K in .KernelSizes,
         .T in .TGroupSizes )
{
    .Defines  = '-DKERNEL_SIZE_LOG2=$K$ -DTGROUP_SIZE_LOG2=$T$'
    .BaseName = 'threadgroup_bitonic_sort_$K$_$T$'

    ForEach( .W  in .WaveIds,
             .WD in .WaveDefines )
    {
        Exec( 'Shader-$BaseName$_$W$' )
        {
            .ExecExecutable = .DXC
            .ExecArguments  = '$DxcCommonFlags$ $Defines$$WD$'
                            + ' -Fc "$OutputDir$/shader_$BaseName$_$W$.asm"'
                            + ' -Fh "$OutputDir$/shader_$BaseName$_$W$.h"'
                            + ' -Vn shader_$BaseName$_$W$'
                            + ' -Fd "$OutputDir$/shader_$BaseName$_$W$.pdb"'
                            + ' "threadgroup_bitonic_sort.hlsl"'
            .ExecOutput     = '$OutputDir$/shader_$BaseName$_$W$.h'
            .ExecInput      = .ShaderDeps
        }

        ^AllShaderTargets + {
            'Shader-$BaseName$_$W$'
        }
    }
}

//------------------------------------------------------------------------------
//  Setup an alias to build/rebuild all targets
//------------------------------------------------------------------------------
Alias( 'all' )
{
    .Targets = .AllShaderTargets
}