158 lines
6.8 KiB
Plaintext
158 lines
6.8 KiB
Plaintext
#include <metal_stdlib>
|
|
|
|
using namespace metal;
|
|
|
|
template<typename T, size_t N>
|
|
struct tint_array {
|
|
const constant T& operator[](size_t i) const constant { return elements[i]; }
|
|
device T& operator[](size_t i) device { return elements[i]; }
|
|
const device T& operator[](size_t i) const device { return elements[i]; }
|
|
thread T& operator[](size_t i) thread { return elements[i]; }
|
|
const thread T& operator[](size_t i) const thread { return elements[i]; }
|
|
threadgroup T& operator[](size_t i) threadgroup { return elements[i]; }
|
|
const threadgroup T& operator[](size_t i) const threadgroup { return elements[i]; }
|
|
T elements[N];
|
|
};
|
|
|
|
struct Inner {
|
|
/* 0x0000 */ float scalar_f32;
|
|
/* 0x0004 */ int scalar_i32;
|
|
/* 0x0008 */ uint scalar_u32;
|
|
/* 0x000c */ half scalar_f16;
|
|
/* 0x000e */ tint_array<int8_t, 2> tint_pad;
|
|
/* 0x0010 */ float2 vec2_f32;
|
|
/* 0x0018 */ int2 vec2_i32;
|
|
/* 0x0020 */ uint2 vec2_u32;
|
|
/* 0x0028 */ half2 vec2_f16;
|
|
/* 0x002c */ tint_array<int8_t, 4> tint_pad_1;
|
|
/* 0x0030 */ packed_float3 vec3_f32;
|
|
/* 0x003c */ tint_array<int8_t, 4> tint_pad_2;
|
|
/* 0x0040 */ packed_int3 vec3_i32;
|
|
/* 0x004c */ tint_array<int8_t, 4> tint_pad_3;
|
|
/* 0x0050 */ packed_uint3 vec3_u32;
|
|
/* 0x005c */ tint_array<int8_t, 4> tint_pad_4;
|
|
/* 0x0060 */ packed_half3 vec3_f16;
|
|
/* 0x0066 */ tint_array<int8_t, 10> tint_pad_5;
|
|
/* 0x0070 */ float4 vec4_f32;
|
|
/* 0x0080 */ int4 vec4_i32;
|
|
/* 0x0090 */ uint4 vec4_u32;
|
|
/* 0x00a0 */ half4 vec4_f16;
|
|
/* 0x00a8 */ float2x2 mat2x2_f32;
|
|
/* 0x00b8 */ tint_array<int8_t, 8> tint_pad_6;
|
|
/* 0x00c0 */ float2x3 mat2x3_f32;
|
|
/* 0x00e0 */ float2x4 mat2x4_f32;
|
|
/* 0x0100 */ float3x2 mat3x2_f32;
|
|
/* 0x0118 */ tint_array<int8_t, 8> tint_pad_7;
|
|
/* 0x0120 */ float3x3 mat3x3_f32;
|
|
/* 0x0150 */ float3x4 mat3x4_f32;
|
|
/* 0x0180 */ float4x2 mat4x2_f32;
|
|
/* 0x01a0 */ float4x3 mat4x3_f32;
|
|
/* 0x01e0 */ float4x4 mat4x4_f32;
|
|
/* 0x0220 */ half2x2 mat2x2_f16;
|
|
/* 0x0228 */ half2x3 mat2x3_f16;
|
|
/* 0x0238 */ half2x4 mat2x4_f16;
|
|
/* 0x0248 */ half3x2 mat3x2_f16;
|
|
/* 0x0254 */ tint_array<int8_t, 4> tint_pad_8;
|
|
/* 0x0258 */ half3x3 mat3x3_f16;
|
|
/* 0x0270 */ half3x4 mat3x4_f16;
|
|
/* 0x0288 */ half4x2 mat4x2_f16;
|
|
/* 0x0298 */ half4x3 mat4x3_f16;
|
|
/* 0x02b8 */ half4x4 mat4x4_f16;
|
|
/* 0x02d8 */ tint_array<int8_t, 8> tint_pad_9;
|
|
/* 0x02e0 */ tint_array<float3, 2> arr2_vec3_f32;
|
|
/* 0x0300 */ tint_array<half4x2, 2> arr2_mat4x2_f16;
|
|
};
|
|
|
|
struct S {
|
|
/* 0x0000 */ tint_array<Inner, 1> arr;
|
|
};
|
|
|
|
void assign_and_preserve_padding(device float2x3* const dest, float2x3 value) {
|
|
(*(dest))[0u] = value[0u];
|
|
(*(dest))[1u] = value[1u];
|
|
}
|
|
|
|
void assign_and_preserve_padding_1(device float3x3* const dest, float3x3 value) {
|
|
(*(dest))[0u] = value[0u];
|
|
(*(dest))[1u] = value[1u];
|
|
(*(dest))[2u] = value[2u];
|
|
}
|
|
|
|
void assign_and_preserve_padding_2(device float4x3* const dest, float4x3 value) {
|
|
(*(dest))[0u] = value[0u];
|
|
(*(dest))[1u] = value[1u];
|
|
(*(dest))[2u] = value[2u];
|
|
(*(dest))[3u] = value[3u];
|
|
}
|
|
|
|
void assign_and_preserve_padding_3(device half2x3* const dest, half2x3 value) {
|
|
(*(dest))[0u] = value[0u];
|
|
(*(dest))[1u] = value[1u];
|
|
}
|
|
|
|
void assign_and_preserve_padding_4(device half3x3* const dest, half3x3 value) {
|
|
(*(dest))[0u] = value[0u];
|
|
(*(dest))[1u] = value[1u];
|
|
(*(dest))[2u] = value[2u];
|
|
}
|
|
|
|
void assign_and_preserve_padding_5(device half4x3* const dest, half4x3 value) {
|
|
(*(dest))[0u] = value[0u];
|
|
(*(dest))[1u] = value[1u];
|
|
(*(dest))[2u] = value[2u];
|
|
(*(dest))[3u] = value[3u];
|
|
}
|
|
|
|
void assign_and_preserve_padding_6(device tint_array<float3, 2>* const dest, tint_array<float3, 2> value) {
|
|
for(uint i = 0u; (i < 2u); i = (i + 1u)) {
|
|
(*(dest))[i] = value[i];
|
|
}
|
|
}
|
|
|
|
void tint_symbol_inner(uint idx, device S* const tint_symbol_3) {
|
|
(*(tint_symbol_3)).arr[idx].scalar_f32 = 0.0f;
|
|
(*(tint_symbol_3)).arr[idx].scalar_i32 = 0;
|
|
(*(tint_symbol_3)).arr[idx].scalar_u32 = 0u;
|
|
(*(tint_symbol_3)).arr[idx].scalar_f16 = 0.0h;
|
|
(*(tint_symbol_3)).arr[idx].vec2_f32 = float2(0.0f);
|
|
(*(tint_symbol_3)).arr[idx].vec2_i32 = int2(0);
|
|
(*(tint_symbol_3)).arr[idx].vec2_u32 = uint2(0u);
|
|
(*(tint_symbol_3)).arr[idx].vec2_f16 = half2(0.0h);
|
|
(*(tint_symbol_3)).arr[idx].vec3_f32 = float3(0.0f);
|
|
(*(tint_symbol_3)).arr[idx].vec3_i32 = int3(0);
|
|
(*(tint_symbol_3)).arr[idx].vec3_u32 = uint3(0u);
|
|
(*(tint_symbol_3)).arr[idx].vec3_f16 = half3(0.0h);
|
|
(*(tint_symbol_3)).arr[idx].vec4_f32 = float4(0.0f);
|
|
(*(tint_symbol_3)).arr[idx].vec4_i32 = int4(0);
|
|
(*(tint_symbol_3)).arr[idx].vec4_u32 = uint4(0u);
|
|
(*(tint_symbol_3)).arr[idx].vec4_f16 = half4(0.0h);
|
|
(*(tint_symbol_3)).arr[idx].mat2x2_f32 = float2x2(float2(0.0f), float2(0.0f));
|
|
assign_and_preserve_padding(&((*(tint_symbol_3)).arr[idx].mat2x3_f32), float2x3(float3(0.0f), float3(0.0f)));
|
|
(*(tint_symbol_3)).arr[idx].mat2x4_f32 = float2x4(float4(0.0f), float4(0.0f));
|
|
(*(tint_symbol_3)).arr[idx].mat3x2_f32 = float3x2(float2(0.0f), float2(0.0f), float2(0.0f));
|
|
assign_and_preserve_padding_1(&((*(tint_symbol_3)).arr[idx].mat3x3_f32), float3x3(float3(0.0f), float3(0.0f), float3(0.0f)));
|
|
(*(tint_symbol_3)).arr[idx].mat3x4_f32 = float3x4(float4(0.0f), float4(0.0f), float4(0.0f));
|
|
(*(tint_symbol_3)).arr[idx].mat4x2_f32 = float4x2(float2(0.0f), float2(0.0f), float2(0.0f), float2(0.0f));
|
|
assign_and_preserve_padding_2(&((*(tint_symbol_3)).arr[idx].mat4x3_f32), float4x3(float3(0.0f), float3(0.0f), float3(0.0f), float3(0.0f)));
|
|
(*(tint_symbol_3)).arr[idx].mat4x4_f32 = float4x4(float4(0.0f), float4(0.0f), float4(0.0f), float4(0.0f));
|
|
(*(tint_symbol_3)).arr[idx].mat2x2_f16 = half2x2(half2(0.0h), half2(0.0h));
|
|
assign_and_preserve_padding_3(&((*(tint_symbol_3)).arr[idx].mat2x3_f16), half2x3(half3(0.0h), half3(0.0h)));
|
|
(*(tint_symbol_3)).arr[idx].mat2x4_f16 = half2x4(half4(0.0h), half4(0.0h));
|
|
(*(tint_symbol_3)).arr[idx].mat3x2_f16 = half3x2(half2(0.0h), half2(0.0h), half2(0.0h));
|
|
assign_and_preserve_padding_4(&((*(tint_symbol_3)).arr[idx].mat3x3_f16), half3x3(half3(0.0h), half3(0.0h), half3(0.0h)));
|
|
(*(tint_symbol_3)).arr[idx].mat3x4_f16 = half3x4(half4(0.0h), half4(0.0h), half4(0.0h));
|
|
(*(tint_symbol_3)).arr[idx].mat4x2_f16 = half4x2(half2(0.0h), half2(0.0h), half2(0.0h), half2(0.0h));
|
|
assign_and_preserve_padding_5(&((*(tint_symbol_3)).arr[idx].mat4x3_f16), half4x3(half3(0.0h), half3(0.0h), half3(0.0h), half3(0.0h)));
|
|
(*(tint_symbol_3)).arr[idx].mat4x4_f16 = half4x4(half4(0.0h), half4(0.0h), half4(0.0h), half4(0.0h));
|
|
tint_array<float3, 2> const tint_symbol_1 = tint_array<float3, 2>{};
|
|
assign_and_preserve_padding_6(&((*(tint_symbol_3)).arr[idx].arr2_vec3_f32), tint_symbol_1);
|
|
tint_array<half4x2, 2> const tint_symbol_2 = tint_array<half4x2, 2>{};
|
|
(*(tint_symbol_3)).arr[idx].arr2_mat4x2_f16 = tint_symbol_2;
|
|
}
|
|
|
|
kernel void tint_symbol(device S* tint_symbol_4 [[buffer(0)]], uint idx [[thread_index_in_threadgroup]]) {
|
|
tint_symbol_inner(idx, tint_symbol_4);
|
|
return;
|
|
}
|
|
|