MSL: Support std140 half matrices and arrays.
Super awkward since there is no clean way to express this.
This commit is contained in:
Hans-Kristian Arntzen
parent
42299f92ef
commit
57dbfa0400
@ -0,0 +1,66 @@
|
||||
#pragma clang diagnostic ignored "-Wmissing-prototypes"
|
||||
|
||||
#include <metal_stdlib>
|
||||
#include <simd/simd.h>
|
||||
|
||||
using namespace metal;
|
||||
|
||||
struct half8 { alignas(16) half4 data; half4 padding_for_std140_fix_your_shader; };
|
||||
using half2x8 = half8[2];
|
||||
using half3x8 = half8[3];
|
||||
using half4x8 = half8[4];
|
||||
struct ushort8 { alignas(16) ushort4 data; ushort4 padding_for_std140_fix_your_shader; };
|
||||
struct short8 { alignas(16) short4 data; short4 padding_for_std140_fix_your_shader; };
|
||||
|
||||
struct Foo
|
||||
{
|
||||
half2x8 c23;
|
||||
half3x8 c32;
|
||||
half3x8 r23;
|
||||
half2x8 r32;
|
||||
half8 h1[6];
|
||||
half8 h2[6];
|
||||
half8 h3[6];
|
||||
half8 h4[6];
|
||||
};
|
||||
|
||||
struct main0_out
|
||||
{
|
||||
float4 FragColor [[color(0)]];
|
||||
};
|
||||
|
||||
fragment main0_out main0(device Foo& _20 [[buffer(0)]])
|
||||
{
|
||||
main0_out out = {};
|
||||
((device half*)&_20.c23[1].data)[2u] = half(1.0);
|
||||
((device half*)&_20.c32[2].data)[1u] = half(2.0);
|
||||
((device half*)&_20.r23[2u])[1] = half(3.0);
|
||||
((device half*)&_20.r32[1u])[2] = half(4.0);
|
||||
_20.c23[1].data = half3(half(0.0), half(1.0), half(2.0));
|
||||
_20.c32[1].data = half2(half(0.0), half(1.0));
|
||||
((device half*)&_20.r23[0])[1] = half3(half(0.0), half(1.0), half(2.0)).x;
|
||||
((device half*)&_20.r23[1])[1] = half3(half(0.0), half(1.0), half(2.0)).y;
|
||||
((device half*)&_20.r23[2])[1] = half3(half(0.0), half(1.0), half(2.0)).z;
|
||||
((device half*)&_20.r32[0])[1] = half2(half(0.0), half(1.0)).x;
|
||||
((device half*)&_20.r32[1])[1] = half2(half(0.0), half(1.0)).y;
|
||||
(device half3&)_20.c23[0] = half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[0];
|
||||
(device half3&)_20.c23[1] = half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[1];
|
||||
(device half2&)_20.c32[0] = half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[0];
|
||||
(device half2&)_20.c32[1] = half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[1];
|
||||
(device half2&)_20.c32[2] = half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[2];
|
||||
(device half2&)_20.r23[0] = half2(half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[0][0], half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[1][0]);
|
||||
(device half2&)_20.r23[1] = half2(half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[0][1], half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[1][1]);
|
||||
(device half2&)_20.r23[2] = half2(half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[0][2], half2x3(half3(half(1.0), half(2.0), half(3.0)), half3(half(4.0), half(5.0), half(6.0)))[1][2]);
|
||||
(device half3&)_20.r32[0] = half3(half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[0][0], half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[1][0], half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[2][0]);
|
||||
(device half3&)_20.r32[1] = half3(half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[0][1], half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[1][1], half3x2(half2(half(1.0), half(2.0)), half2(half(3.0), half(4.0)), half2(half(5.0), half(6.0)))[2][1]);
|
||||
_20.h1[5].data = half(1.0);
|
||||
_20.h2[5].data = half2(half(1.0), half(2.0));
|
||||
_20.h3[5].data = half3(half(1.0), half(2.0), half(3.0));
|
||||
_20.h4[5].data = half4(half(1.0), half(2.0), half(3.0), half(4.0));
|
||||
((device half*)&_20.h2[5].data)[1u] = half(10.0);
|
||||
((device half*)&_20.h3[5].data)[2u] = half(11.0);
|
||||
((device half*)&_20.h4[5].data)[3u] = half(12.0);
|
||||
out.FragColor = float4(1.0);
|
||||
return out;
|
||||
}
|
||||
|
||||
@ -0,0 +1,100 @@
|
||||
#pragma clang diagnostic ignored "-Wmissing-prototypes"
|
||||
|
||||
#include <metal_stdlib>
|
||||
#include <simd/simd.h>
|
||||
|
||||
using namespace metal;
|
||||
|
||||
struct half8 { alignas(16) half4 data; half4 padding_for_std140_fix_your_shader; };
|
||||
using half2x8 = half8[2];
|
||||
using half3x8 = half8[3];
|
||||
using half4x8 = half8[4];
|
||||
struct ushort8 { alignas(16) ushort4 data; ushort4 padding_for_std140_fix_your_shader; };
|
||||
struct short8 { alignas(16) short4 data; short4 padding_for_std140_fix_your_shader; };
|
||||
|
||||
struct Foo
|
||||
{
|
||||
half2x8 c22;
|
||||
half2x8 c22arr[3];
|
||||
half2x8 c23;
|
||||
half2x8 c24;
|
||||
half3x8 c32;
|
||||
half3x8 c33;
|
||||
half3x8 c34;
|
||||
half4x8 c42;
|
||||
half4x8 c43;
|
||||
half4x8 c44;
|
||||
half2x8 r22;
|
||||
half2x8 r22arr[3];
|
||||
half3x8 r23;
|
||||
half4x8 r24;
|
||||
half2x8 r32;
|
||||
half3x8 r33;
|
||||
half4x8 r34;
|
||||
half2x8 r42;
|
||||
half3x8 r43;
|
||||
half4x8 r44;
|
||||
half8 h1[6];
|
||||
half8 h2[6];
|
||||
half8 h3[6];
|
||||
half8 h4[6];
|
||||
};
|
||||
|
||||
struct main0_out
|
||||
{
|
||||
float4 FragColor [[color(0)]];
|
||||
};
|
||||
|
||||
fragment main0_out main0(constant Foo& u [[buffer(0)]])
|
||||
{
|
||||
main0_out out = {};
|
||||
half2 c2 = u.c22[0].data.xy + u.c22[1].data.xy;
|
||||
c2 = u.c22arr[2][0].data.xy + u.c22arr[2][1].data.xy;
|
||||
half3 c3 = u.c23[0].data.xyz + u.c23[1].data.xyz;
|
||||
half4 c4 = u.c24[0].data + u.c24[1].data;
|
||||
c2 = (u.c32[0].data.xy + u.c32[1].data.xy) + u.c32[2].data.xy;
|
||||
c3 = (u.c33[0].data.xyz + u.c33[1].data.xyz) + u.c33[2].data.xyz;
|
||||
c4 = (u.c34[0].data + u.c34[1].data) + u.c34[2].data;
|
||||
c2 = ((u.c42[0].data.xy + u.c42[1].data.xy) + u.c42[2].data.xy) + u.c42[3].data.xy;
|
||||
c3 = ((u.c43[0].data.xyz + u.c43[1].data.xyz) + u.c43[2].data.xyz) + u.c43[3].data.xyz;
|
||||
c4 = ((u.c44[0].data + u.c44[1].data) + u.c44[2].data) + u.c44[3].data;
|
||||
half c = ((u.c22[0].data.x + u.c22[0].data.y) + u.c22[1].data.x) + u.c22[1].data.y;
|
||||
c = ((u.c22arr[2][0].data.x + u.c22arr[2][0].data.y) + u.c22arr[2][1].data.x) + u.c22arr[2][1].data.y;
|
||||
half2x2 c22 = half2x2(u.c22[0].data.xy, u.c22[1].data.xy);
|
||||
c22 = half2x2(u.c22arr[2][0].data.xy, u.c22arr[2][1].data.xy);
|
||||
half2x3 c23 = half2x3(u.c23[0].data.xyz, u.c23[1].data.xyz);
|
||||
half2x4 c24 = half2x4(u.c24[0].data, u.c24[1].data);
|
||||
half3x2 c32 = half3x2(u.c32[0].data.xy, u.c32[1].data.xy, u.c32[2].data.xy);
|
||||
half3x3 c33 = half3x3(u.c33[0].data.xyz, u.c33[1].data.xyz, u.c33[2].data.xyz);
|
||||
half3x4 c34 = half3x4(u.c34[0].data, u.c34[1].data, u.c34[2].data);
|
||||
half4x2 c42 = half4x2(u.c42[0].data.xy, u.c42[1].data.xy, u.c42[2].data.xy, u.c42[3].data.xy);
|
||||
half4x3 c43 = half4x3(u.c43[0].data.xyz, u.c43[1].data.xyz, u.c43[2].data.xyz, u.c43[3].data.xyz);
|
||||
half4x4 c44 = half4x4(u.c44[0].data, u.c44[1].data, u.c44[2].data, u.c44[3].data);
|
||||
half2 r2 = half2(u.r22[0].data[0], u.r22[1].data[0]) + half2(u.r22[0].data[1], u.r22[1].data[1]);
|
||||
r2 = half2(u.r22arr[2][0].data[0], u.r22arr[2][1].data[0]) + half2(u.r22arr[2][0].data[1], u.r22arr[2][1].data[1]);
|
||||
half3 r3 = half3(u.r23[0].data[0], u.r23[1].data[0], u.r23[2].data[0]) + half3(u.r23[0].data[1], u.r23[1].data[1], u.r23[2].data[1]);
|
||||
half4 r4 = half4(u.r24[0].data[0], u.r24[1].data[0], u.r24[2].data[0], u.r24[3].data[0]) + half4(u.r24[0].data[1], u.r24[1].data[1], u.r24[2].data[1], u.r24[3].data[1]);
|
||||
r2 = (half2(u.r32[0].data[0], u.r32[1].data[0]) + half2(u.r32[0].data[1], u.r32[1].data[1])) + half2(u.r32[0].data[2], u.r32[1].data[2]);
|
||||
r3 = (half3(u.r33[0].data[0], u.r33[1].data[0], u.r33[2].data[0]) + half3(u.r33[0].data[1], u.r33[1].data[1], u.r33[2].data[1])) + half3(u.r33[0].data[2], u.r33[1].data[2], u.r33[2].data[2]);
|
||||
r4 = (half4(u.r34[0].data[0], u.r34[1].data[0], u.r34[2].data[0], u.r34[3].data[0]) + half4(u.r34[0].data[1], u.r34[1].data[1], u.r34[2].data[1], u.r34[3].data[1])) + half4(u.r34[0].data[2], u.r34[1].data[2], u.r34[2].data[2], u.r34[3].data[2]);
|
||||
r2 = ((half2(u.r42[0].data[0], u.r42[1].data[0]) + half2(u.r42[0].data[1], u.r42[1].data[1])) + half2(u.r42[0].data[2], u.r42[1].data[2])) + half2(u.r42[0].data[3], u.r42[1].data[3]);
|
||||
r3 = ((half3(u.r43[0].data[0], u.r43[1].data[0], u.r43[2].data[0]) + half3(u.r43[0].data[1], u.r43[1].data[1], u.r43[2].data[1])) + half3(u.r43[0].data[2], u.r43[1].data[2], u.r43[2].data[2])) + half3(u.r43[0].data[3], u.r43[1].data[3], u.r43[2].data[3]);
|
||||
r4 = ((half4(u.r44[0].data[0], u.r44[1].data[0], u.r44[2].data[0], u.r44[3].data[0]) + half4(u.r44[0].data[1], u.r44[1].data[1], u.r44[2].data[1], u.r44[3].data[1])) + half4(u.r44[0].data[2], u.r44[1].data[2], u.r44[2].data[2], u.r44[3].data[2])) + half4(u.r44[0].data[3], u.r44[1].data[3], u.r44[2].data[3], u.r44[3].data[3]);
|
||||
half r = ((u.r22[0u].data[0] + u.r22[1u].data[0]) + u.r22[0u].data[1]) + u.r22[1u].data[1];
|
||||
half2x2 r22 = transpose(half2x2(u.r22[0].data.xy, u.r22[1].data.xy));
|
||||
half2x3 r23 = transpose(half3x2(u.r23[0].data.xy, u.r23[1].data.xy, u.r23[2].data.xy));
|
||||
half2x4 r24 = transpose(half4x2(u.r24[0].data.xy, u.r24[1].data.xy, u.r24[2].data.xy, u.r24[3].data.xy));
|
||||
half3x2 r32 = transpose(half2x3(u.r32[0].data.xyz, u.r32[1].data.xyz));
|
||||
half3x3 r33 = transpose(half3x3(u.r33[0].data.xyz, u.r33[1].data.xyz, u.r33[2].data.xyz));
|
||||
half3x4 r34 = transpose(half4x3(u.r34[0].data.xyz, u.r34[1].data.xyz, u.r34[2].data.xyz, u.r34[3].data.xyz));
|
||||
half4x2 r42 = transpose(half2x4(u.r42[0].data, u.r42[1].data));
|
||||
half4x3 r43 = transpose(half3x4(u.r43[0].data, u.r43[1].data, u.r43[2].data));
|
||||
half4x4 r44 = transpose(half4x4(u.r44[0].data, u.r44[1].data, u.r44[2].data, u.r44[3].data));
|
||||
half h1 = u.h1[5].data.x;
|
||||
half2 h2 = u.h2[5].data.xy;
|
||||
half3 h3 = u.h3[5].data.xyz;
|
||||
half4 h4 = half4(u.h4[5].data);
|
||||
out.FragColor = float4(1.0);
|
||||
return out;
|
||||
}
|
||||
|
||||
@ -0,0 +1,51 @@
|
||||
#version 450
|
||||
#extension GL_EXT_shader_explicit_arithmetic_types : require
|
||||
|
||||
layout(set = 0, binding = 0, std140) buffer Foo
|
||||
{
|
||||
f16mat2x3 c23;
|
||||
f16mat3x2 c32;
|
||||
layout(row_major) f16mat2x3 r23;
|
||||
layout(row_major) f16mat3x2 r32;
|
||||
|
||||
float16_t h1[6];
|
||||
f16vec2 h2[6];
|
||||
f16vec3 h3[6];
|
||||
f16vec4 h4[6];
|
||||
};
|
||||
|
||||
layout(location = 0) out vec4 FragColor;
|
||||
|
||||
void main()
|
||||
{
|
||||
// Store scalar
|
||||
c23[1][2] = 1.0hf;
|
||||
c32[2][1] = 2.0hf;
|
||||
r23[1][2] = 3.0hf;
|
||||
r32[2][1] = 4.0hf;
|
||||
|
||||
// Store vector
|
||||
c23[1] = f16vec3(0, 1, 2);
|
||||
c32[1] = f16vec2(0, 1);
|
||||
r23[1] = f16vec3(0, 1, 2);
|
||||
r32[1] = f16vec2(0, 1);
|
||||
|
||||
// Store matrix
|
||||
c23 = f16mat2x3(1, 2, 3, 4, 5, 6);
|
||||
c32 = f16mat3x2(1, 2, 3, 4, 5, 6);
|
||||
r23 = f16mat2x3(1, 2, 3, 4, 5, 6);
|
||||
r32 = f16mat3x2(1, 2, 3, 4, 5, 6);
|
||||
|
||||
// Store array
|
||||
h1[5] = 1.0hf;
|
||||
h2[5] = f16vec2(1, 2);
|
||||
h3[5] = f16vec3(1, 2, 3);
|
||||
h4[5] = f16vec4(1, 2, 3, 4);
|
||||
|
||||
// Store scalar in array
|
||||
h2[5][1] = 10.0hf;
|
||||
h3[5][2] = 11.0hf;
|
||||
h4[5][3] = 12.0hf;
|
||||
|
||||
FragColor = vec4(1.0);
|
||||
}
|
||||
110
shaders-msl-no-opt/frag/std140-half-matrix-and-array.frag
Normal file
110
shaders-msl-no-opt/frag/std140-half-matrix-and-array.frag
Normal file
@ -0,0 +1,110 @@
|
||||
#version 450
|
||||
#extension GL_EXT_shader_explicit_arithmetic_types : require
|
||||
|
||||
layout(set = 0, binding = 0, std140) uniform Foo
|
||||
{
|
||||
f16mat2x2 c22;
|
||||
f16mat2x2 c22arr[3];
|
||||
f16mat2x3 c23;
|
||||
f16mat2x4 c24;
|
||||
|
||||
f16mat3x2 c32;
|
||||
f16mat3x3 c33;
|
||||
f16mat3x4 c34;
|
||||
|
||||
f16mat4x2 c42;
|
||||
f16mat4x3 c43;
|
||||
f16mat4x4 c44;
|
||||
|
||||
layout(row_major) f16mat2x2 r22;
|
||||
layout(row_major) f16mat2x2 r22arr[3];
|
||||
layout(row_major) f16mat2x3 r23;
|
||||
layout(row_major) f16mat2x4 r24;
|
||||
|
||||
layout(row_major) f16mat3x2 r32;
|
||||
layout(row_major) f16mat3x3 r33;
|
||||
layout(row_major) f16mat3x4 r34;
|
||||
|
||||
layout(row_major) f16mat4x2 r42;
|
||||
layout(row_major) f16mat4x3 r43;
|
||||
layout(row_major) f16mat4x4 r44;
|
||||
|
||||
float16_t h1[6];
|
||||
f16vec2 h2[6];
|
||||
f16vec3 h3[6];
|
||||
f16vec4 h4[6];
|
||||
} u;
|
||||
|
||||
layout(location = 0) out vec4 FragColor;
|
||||
|
||||
void main()
|
||||
{
|
||||
// Load vectors.
|
||||
f16vec2 c2 = u.c22[0] + u.c22[1];
|
||||
c2 = u.c22arr[2][0] + u.c22arr[2][1];
|
||||
f16vec3 c3 = u.c23[0] + u.c23[1];
|
||||
f16vec4 c4 = u.c24[0] + u.c24[1];
|
||||
|
||||
c2 = u.c32[0] + u.c32[1] + u.c32[2];
|
||||
c3 = u.c33[0] + u.c33[1] + u.c33[2];
|
||||
c4 = u.c34[0] + u.c34[1] + u.c34[2];
|
||||
|
||||
c2 = u.c42[0] + u.c42[1] + u.c42[2] + u.c42[3];
|
||||
c3 = u.c43[0] + u.c43[1] + u.c43[2] + u.c43[3];
|
||||
c4 = u.c44[0] + u.c44[1] + u.c44[2] + u.c44[3];
|
||||
|
||||
// Load scalars.
|
||||
float16_t c = u.c22[0].x + u.c22[0].y + u.c22[1].x + u.c22[1].y;
|
||||
c = u.c22arr[2][0].x + u.c22arr[2][0].y + u.c22arr[2][1].x + u.c22arr[2][1].y;
|
||||
|
||||
// Load full matrix.
|
||||
f16mat2x2 c22 = u.c22;
|
||||
c22 = u.c22arr[2];
|
||||
f16mat2x3 c23 = u.c23;
|
||||
f16mat2x4 c24 = u.c24;
|
||||
|
||||
f16mat3x2 c32 = u.c32;
|
||||
f16mat3x3 c33 = u.c33;
|
||||
f16mat3x4 c34 = u.c34;
|
||||
|
||||
f16mat4x2 c42 = u.c42;
|
||||
f16mat4x3 c43 = u.c43;
|
||||
f16mat4x4 c44 = u.c44;
|
||||
|
||||
// Same, but row-major.
|
||||
f16vec2 r2 = u.r22[0] + u.r22[1];
|
||||
r2 = u.r22arr[2][0] + u.r22arr[2][1];
|
||||
f16vec3 r3 = u.r23[0] + u.r23[1];
|
||||
f16vec4 r4 = u.r24[0] + u.r24[1];
|
||||
|
||||
r2 = u.r32[0] + u.r32[1] + u.r32[2];
|
||||
r3 = u.r33[0] + u.r33[1] + u.r33[2];
|
||||
r4 = u.r34[0] + u.r34[1] + u.r34[2];
|
||||
|
||||
r2 = u.r42[0] + u.r42[1] + u.r42[2] + u.r42[3];
|
||||
r3 = u.r43[0] + u.r43[1] + u.r43[2] + u.r43[3];
|
||||
r4 = u.r44[0] + u.r44[1] + u.r44[2] + u.r44[3];
|
||||
|
||||
// Load scalars.
|
||||
float16_t r = u.r22[0].x + u.r22[0].y + u.r22[1].x + u.r22[1].y;
|
||||
|
||||
// Load full matrix.
|
||||
f16mat2x2 r22 = u.r22;
|
||||
f16mat2x3 r23 = u.r23;
|
||||
f16mat2x4 r24 = u.r24;
|
||||
|
||||
f16mat3x2 r32 = u.r32;
|
||||
f16mat3x3 r33 = u.r33;
|
||||
f16mat3x4 r34 = u.r34;
|
||||
|
||||
f16mat4x2 r42 = u.r42;
|
||||
f16mat4x3 r43 = u.r43;
|
||||
f16mat4x4 r44 = u.r44;
|
||||
|
||||
float16_t h1 = u.h1[5];
|
||||
f16vec2 h2 = u.h2[5];
|
||||
f16vec3 h3 = u.h3[5];
|
||||
f16vec4 h4 = u.h4[5];
|
||||
|
||||
FragColor = vec4(1.0);
|
||||
}
|
||||
@ -10176,6 +10176,16 @@ string CompilerGLSL::access_chain_internal(uint32_t base, const uint32_t *indice
|
||||
type_id = type->parent_type;
|
||||
type = &get<SPIRType>(type_id);
|
||||
|
||||
// If the physical type has an unnatural vecsize,
|
||||
// we must assume it's a faked struct where the .data member
|
||||
// is used for the real payload.
|
||||
if (physical_type && (is_vector(*type) || is_scalar(*type)))
|
||||
{
|
||||
auto &phys = get<SPIRType>(physical_type);
|
||||
if (phys.vecsize > 4)
|
||||
expr += ".data";
|
||||
}
|
||||
|
||||
access_chain_is_arrayed = true;
|
||||
}
|
||||
// For structs, the index refers to a constant, which indexes into the members, possibly through a redirection mapping.
|
||||
@ -10261,6 +10271,16 @@ string CompilerGLSL::access_chain_internal(uint32_t base, const uint32_t *indice
|
||||
expr += to_unpacked_expression(index, register_expression_read);
|
||||
expr += "]";
|
||||
|
||||
// If the physical type has an unnatural vecsize,
|
||||
// we must assume it's a faked struct where the .data member
|
||||
// is used for the real payload.
|
||||
if (physical_type)
|
||||
{
|
||||
auto &phys = get<SPIRType>(physical_type);
|
||||
if (phys.vecsize > 4 || phys.columns > 4)
|
||||
expr += ".data";
|
||||
}
|
||||
|
||||
type_id = type->parent_type;
|
||||
type = &get<SPIRType>(type_id);
|
||||
}
|
||||
@ -10275,6 +10295,18 @@ string CompilerGLSL::access_chain_internal(uint32_t base, const uint32_t *indice
|
||||
if (column_index != string::npos)
|
||||
{
|
||||
deferred_index = expr.substr(column_index);
|
||||
|
||||
auto end_deferred_index = deferred_index.find_last_of(']');
|
||||
if (end_deferred_index != string::npos && end_deferred_index + 1 != deferred_index.size())
|
||||
{
|
||||
// If we have any data member fixups, it must be transposed so that it refers to this index.
|
||||
// E.g. [0].data followed by [1] would be shuffled to [1][0].data which is wrong,
|
||||
// and needs to be [1].data[0] instead.
|
||||
end_deferred_index++;
|
||||
deferred_index = deferred_index.substr(end_deferred_index) +
|
||||
deferred_index.substr(0, end_deferred_index);
|
||||
}
|
||||
|
||||
expr.resize(column_index);
|
||||
}
|
||||
}
|
||||
@ -10353,8 +10385,14 @@ string CompilerGLSL::access_chain_internal(uint32_t base, const uint32_t *indice
|
||||
|
||||
if (row_major_matrix_needs_conversion && !ignore_potential_sliced_writes)
|
||||
{
|
||||
prepare_access_chain_for_scalar_access(expr, get<SPIRType>(type->parent_type), effective_storage,
|
||||
is_packed);
|
||||
if (prepare_access_chain_for_scalar_access(expr, get<SPIRType>(type->parent_type), effective_storage,
|
||||
is_packed))
|
||||
{
|
||||
// We're in a pointer context now, so just remove any member dereference.
|
||||
auto first_index = deferred_index.find_first_of('[');
|
||||
if (first_index != string::npos && first_index != 0)
|
||||
deferred_index = deferred_index.substr(first_index);
|
||||
}
|
||||
}
|
||||
|
||||
if (access_meshlet_position_y)
|
||||
@ -10413,8 +10451,9 @@ void CompilerGLSL::check_physical_type_cast(std::string &, const SPIRType *, uin
|
||||
{
|
||||
}
|
||||
|
||||
void CompilerGLSL::prepare_access_chain_for_scalar_access(std::string &, const SPIRType &, spv::StorageClass, bool &)
|
||||
bool CompilerGLSL::prepare_access_chain_for_scalar_access(std::string &, const SPIRType &, spv::StorageClass, bool &)
|
||||
{
|
||||
return false;
|
||||
}
|
||||
|
||||
string CompilerGLSL::to_flattened_struct_member(const string &basename, const SPIRType &type, uint32_t index)
|
||||
@ -14957,6 +14996,17 @@ string CompilerGLSL::convert_row_major_matrix(string exp_str, const SPIRType &ex
|
||||
auto column_expr = exp_str.substr(column_index);
|
||||
exp_str.resize(column_index);
|
||||
|
||||
auto end_deferred_index = column_expr.find_last_of(']');
|
||||
if (end_deferred_index != string::npos && end_deferred_index + 1 != column_expr.size())
|
||||
{
|
||||
// If we have any data member fixups, it must be transposed so that it refers to this index.
|
||||
// E.g. [0].data followed by [1] would be shuffled to [1][0].data which is wrong,
|
||||
// and needs to be [1].data[0] instead.
|
||||
end_deferred_index++;
|
||||
column_expr = column_expr.substr(end_deferred_index) +
|
||||
column_expr.substr(0, end_deferred_index);
|
||||
}
|
||||
|
||||
auto transposed_expr = type_to_glsl_constructor(exp_type) + "(";
|
||||
|
||||
// Loading a column from a row-major matrix. Unroll the load.
|
||||
|
||||
@ -757,7 +757,7 @@ protected:
|
||||
virtual bool access_chain_needs_stage_io_builtin_translation(uint32_t base);
|
||||
|
||||
virtual void check_physical_type_cast(std::string &expr, const SPIRType *type, uint32_t physical_type);
|
||||
virtual void prepare_access_chain_for_scalar_access(std::string &expr, const SPIRType &type,
|
||||
virtual bool prepare_access_chain_for_scalar_access(std::string &expr, const SPIRType &type,
|
||||
spv::StorageClass storage, bool &is_packed);
|
||||
|
||||
std::string access_chain(uint32_t base, const uint32_t *indices, uint32_t count, const SPIRType &target_type,
|
||||
|
||||
@ -4775,9 +4775,17 @@ void CompilerMSL::ensure_member_packing_rules_msl(SPIRType &ib_type, uint32_t in
|
||||
|
||||
if (elems_per_stride == 3)
|
||||
SPIRV_CROSS_THROW("Cannot use ArrayStride of 3 elements in remapping scenarios.");
|
||||
else if (elems_per_stride > 4)
|
||||
else if (elems_per_stride > 4 && elems_per_stride != 8)
|
||||
SPIRV_CROSS_THROW("Cannot represent vectors with more than 4 elements in MSL.");
|
||||
|
||||
if (elems_per_stride == 8)
|
||||
{
|
||||
if (mbr_type.width == 16)
|
||||
add_spv_func_and_recompile(SPVFuncImplHalfStd140);
|
||||
else
|
||||
SPIRV_CROSS_THROW("Unexpected type in std140 wide array resolve.");
|
||||
}
|
||||
|
||||
auto physical_type = mbr_type;
|
||||
physical_type.vecsize = elems_per_stride;
|
||||
physical_type.parent_type = 0;
|
||||
@ -4809,13 +4817,20 @@ void CompilerMSL::ensure_member_packing_rules_msl(SPIRType &ib_type, uint32_t in
|
||||
|
||||
if (elems_per_stride == 3)
|
||||
SPIRV_CROSS_THROW("Cannot use ArrayStride of 3 elements in remapping scenarios.");
|
||||
else if (elems_per_stride > 4)
|
||||
else if (elems_per_stride > 4 && elems_per_stride != 8)
|
||||
SPIRV_CROSS_THROW("Cannot represent vectors with more than 4 elements in MSL.");
|
||||
|
||||
bool row_major = has_member_decoration(ib_type.self, index, DecorationRowMajor);
|
||||
if (elems_per_stride == 8)
|
||||
{
|
||||
if (mbr_type.basetype != SPIRType::Half)
|
||||
SPIRV_CROSS_THROW("Unexpected type in std140 wide matrix stride resolve.");
|
||||
add_spv_func_and_recompile(SPVFuncImplHalfStd140);
|
||||
}
|
||||
|
||||
bool row_major = has_member_decoration(ib_type.self, index, DecorationRowMajor);
|
||||
auto physical_type = mbr_type;
|
||||
physical_type.parent_type = 0;
|
||||
|
||||
if (row_major)
|
||||
physical_type.columns = elems_per_stride;
|
||||
else
|
||||
@ -5114,6 +5129,13 @@ void CompilerMSL::emit_store_statement(uint32_t lhs_expression, uint32_t rhs_exp
|
||||
{
|
||||
auto lhs_expr = to_enclosed_expression(lhs_expression);
|
||||
auto column_index = lhs_expr.find_last_of('[');
|
||||
|
||||
// Get rid of any ".data" half8 handling here, we're casting to scalar anyway.
|
||||
auto end_column_index = lhs_expr.find_last_of(']');
|
||||
auto end_dot_index = lhs_expr.find_last_of('.');
|
||||
if (end_dot_index != string::npos && end_dot_index > end_column_index)
|
||||
lhs_expr.resize(end_dot_index);
|
||||
|
||||
if (column_index != string::npos)
|
||||
{
|
||||
statement("((", cast_addr_space, " ", type_to_glsl(write_type), "*)&",
|
||||
@ -5124,7 +5146,9 @@ void CompilerMSL::emit_store_statement(uint32_t lhs_expression, uint32_t rhs_exp
|
||||
|
||||
lhs_e->need_transpose = true;
|
||||
}
|
||||
else if ((is_matrix(physical_type) || is_array(physical_type)) && physical_type.vecsize > type.vecsize)
|
||||
else if ((is_matrix(physical_type) || is_array(physical_type)) &&
|
||||
physical_type.vecsize <= 4 &&
|
||||
physical_type.vecsize > type.vecsize)
|
||||
{
|
||||
assert(type.vecsize >= 1 && type.vecsize <= 3);
|
||||
|
||||
@ -5181,6 +5205,7 @@ string CompilerMSL::unpack_expression_type(string expr_str, const SPIRType &type
|
||||
".x",
|
||||
".xy",
|
||||
".xyz",
|
||||
"",
|
||||
};
|
||||
|
||||
if (physical_type && is_vector(*physical_type) && is_array(*physical_type) &&
|
||||
@ -5193,7 +5218,7 @@ string CompilerMSL::unpack_expression_type(string expr_str, const SPIRType &type
|
||||
else if (physical_type && is_matrix(*physical_type) && is_vector(type) && physical_type->vecsize > type.vecsize)
|
||||
{
|
||||
// Extract column from padded matrix.
|
||||
assert(type.vecsize >= 1 && type.vecsize <= 3);
|
||||
assert(type.vecsize >= 1 && type.vecsize <= 4);
|
||||
return enclose_expression(expr_str) + swizzle_lut[type.vecsize - 1];
|
||||
}
|
||||
else if (is_matrix(type))
|
||||
@ -5215,6 +5240,7 @@ string CompilerMSL::unpack_expression_type(string expr_str, const SPIRType &type
|
||||
string unpack_expr = join(base_type, columns, "x", vecsize, "(");
|
||||
|
||||
const char *load_swiz = "";
|
||||
const char *data_swiz = physical_vecsize > 4 ? ".data" : "";
|
||||
|
||||
if (physical_vecsize != vecsize)
|
||||
load_swiz = swizzle_lut[vecsize - 1];
|
||||
@ -5227,7 +5253,7 @@ string CompilerMSL::unpack_expression_type(string expr_str, const SPIRType &type
|
||||
if (packed)
|
||||
unpack_expr += join(base_type, physical_vecsize, "(", expr_str, "[", i, "]", ")", load_swiz);
|
||||
else
|
||||
unpack_expr += join(expr_str, "[", i, "]", load_swiz);
|
||||
unpack_expr += join(expr_str, "[", i, "]", data_swiz, load_swiz);
|
||||
}
|
||||
|
||||
unpack_expr += ")";
|
||||
@ -7335,6 +7361,18 @@ void CompilerMSL::emit_custom_functions()
|
||||
}
|
||||
break;
|
||||
|
||||
case SPVFuncImplHalfStd140:
|
||||
// .data is used in access chain.
|
||||
statement("struct half8 { alignas(16) half4 data; half4 padding_for_std140_fix_your_shader; };");
|
||||
// Physical type remapping is used to load/store full matrices anyway.
|
||||
statement("using half2x8 = half8[2];");
|
||||
statement("using half3x8 = half8[3];");
|
||||
statement("using half4x8 = half8[4];");
|
||||
statement("struct ushort8 { alignas(16) ushort4 data; ushort4 padding_for_std140_fix_your_shader; };");
|
||||
statement("struct short8 { alignas(16) short4 data; short4 padding_for_std140_fix_your_shader; };");
|
||||
statement("");
|
||||
break;
|
||||
|
||||
default:
|
||||
break;
|
||||
}
|
||||
@ -8338,7 +8376,7 @@ bool CompilerMSL::is_out_of_bounds_tessellation_level(uint32_t id_lhs)
|
||||
(builtin == BuiltInTessLevelOuter && c->scalar() == 3);
|
||||
}
|
||||
|
||||
void CompilerMSL::prepare_access_chain_for_scalar_access(std::string &expr, const SPIRType &type,
|
||||
bool CompilerMSL::prepare_access_chain_for_scalar_access(std::string &expr, const SPIRType &type,
|
||||
spv::StorageClass storage, bool &is_packed)
|
||||
{
|
||||
// If there is any risk of writes happening with the access chain in question,
|
||||
@ -8352,7 +8390,10 @@ void CompilerMSL::prepare_access_chain_for_scalar_access(std::string &expr, cons
|
||||
|
||||
// Further indexing should happen with packed rules (array index, not swizzle).
|
||||
is_packed = true;
|
||||
return true;
|
||||
}
|
||||
else
|
||||
return false;
|
||||
}
|
||||
|
||||
bool CompilerMSL::access_chain_needs_stage_io_builtin_translation(uint32_t base)
|
||||
|
||||
@ -815,6 +815,7 @@ protected:
|
||||
SPVFuncImplVariableDescriptor,
|
||||
SPVFuncImplVariableSizedDescriptor,
|
||||
SPVFuncImplVariableDescriptorArray,
|
||||
SPVFuncImplHalfStd140
|
||||
};
|
||||
|
||||
// If the underlying resource has been used for comparison then duplicate loads of that resource must be too
|
||||
@ -1097,7 +1098,7 @@ protected:
|
||||
void analyze_sampled_image_usage();
|
||||
|
||||
bool access_chain_needs_stage_io_builtin_translation(uint32_t base) override;
|
||||
void prepare_access_chain_for_scalar_access(std::string &expr, const SPIRType &type, spv::StorageClass storage,
|
||||
bool prepare_access_chain_for_scalar_access(std::string &expr, const SPIRType &type, spv::StorageClass storage,
|
||||
bool &is_packed) override;
|
||||
void fix_up_interpolant_access_chain(const uint32_t *ops, uint32_t length);
|
||||
void check_physical_type_cast(std::string &expr, const SPIRType *type, uint32_t physical_type) override;
|
||||
|
||||
Loading…
Reference in New Issue
Block a user