This is the first of a sequence of changes to support compiling Cycles kernels as MSL (Metal Shading Language) in preparation for a Metal GPU device implementation. MSL requires that all pointer types be declared with explicit address space attributes (device, thread, etc...). There is already precedent for this with Cycles' address space macros (ccl_global, ccl_private, etc...), therefore the first step of MSL-enablement is to apply these consistently. Line-for-line this represents the largest change required to enable MSL. Applying this change first will simplify future patches as well as offering the emergent benefit of enhanced descriptiveness. The vast majority of deltas in this patch fall into one of two cases: - Ensuring ccl_private is specified for thread-local pointer types - Ensuring ccl_global is specified for device-wide pointer types Additionally, the ccl_addr_space qualifier can be removed. Prior to Cycles X, ccl_addr_space was used as a context-dependent address space qualifier, but now it is either redundant (e.g. in struct typedefs), or can be replaced by ccl_global in the case of pointer types. Associated function variants (e.g. lcg_step_float_addrspace) are also redundant. In cases where address space qualifiers are chained with "const", this patch places the address space qualifier first. The rationale for this is that the choice of address space is likely to have the greater impact on runtime performance and overall architecture. The final part of this patch is the addition of a metal/compat.h header. This is partially complete and will be extended in future patches, paving the way for the full Metal implementation. Ref T92212 Reviewed By: brecht Maniphest Tasks: T92212 Differential Revision: https://developer.blender.org/D12864
256 lines
8.1 KiB
C
256 lines
8.1 KiB
C
/*
|
|
* Copyright 2011-2013 Blender Foundation
|
|
*
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*/
|
|
|
|
CCL_NAMESPACE_BEGIN
|
|
|
|
ccl_device float4
|
|
svm_image_texture(ccl_global const KernelGlobals *kg, int id, float x, float y, uint flags)
|
|
{
|
|
if (id == -1) {
|
|
return make_float4(
|
|
TEX_IMAGE_MISSING_R, TEX_IMAGE_MISSING_G, TEX_IMAGE_MISSING_B, TEX_IMAGE_MISSING_A);
|
|
}
|
|
|
|
float4 r = kernel_tex_image_interp(kg, id, x, y);
|
|
const float alpha = r.w;
|
|
|
|
if ((flags & NODE_IMAGE_ALPHA_UNASSOCIATE) && alpha != 1.0f && alpha != 0.0f) {
|
|
r /= alpha;
|
|
r.w = alpha;
|
|
}
|
|
|
|
if (flags & NODE_IMAGE_COMPRESS_AS_SRGB) {
|
|
r = color_srgb_to_linear_v4(r);
|
|
}
|
|
|
|
return r;
|
|
}
|
|
|
|
/* Remap coordinate from 0..1 box to -1..-1 */
|
|
ccl_device_inline float3 texco_remap_square(float3 co)
|
|
{
|
|
return (co - make_float3(0.5f, 0.5f, 0.5f)) * 2.0f;
|
|
}
|
|
|
|
ccl_device_noinline int svm_node_tex_image(ccl_global const KernelGlobals *kg,
|
|
ccl_private ShaderData *sd,
|
|
ccl_private float *stack,
|
|
uint4 node,
|
|
int offset)
|
|
{
|
|
uint co_offset, out_offset, alpha_offset, flags;
|
|
|
|
svm_unpack_node_uchar4(node.z, &co_offset, &out_offset, &alpha_offset, &flags);
|
|
|
|
float3 co = stack_load_float3(stack, co_offset);
|
|
float2 tex_co;
|
|
if (node.w == NODE_IMAGE_PROJ_SPHERE) {
|
|
co = texco_remap_square(co);
|
|
tex_co = map_to_sphere(co);
|
|
}
|
|
else if (node.w == NODE_IMAGE_PROJ_TUBE) {
|
|
co = texco_remap_square(co);
|
|
tex_co = map_to_tube(co);
|
|
}
|
|
else {
|
|
tex_co = make_float2(co.x, co.y);
|
|
}
|
|
|
|
/* TODO(lukas): Consider moving tile information out of the SVM node.
|
|
* TextureInfo seems a reasonable candidate. */
|
|
int id = -1;
|
|
int num_nodes = (int)node.y;
|
|
if (num_nodes > 0) {
|
|
/* Remember the offset of the node following the tile nodes. */
|
|
int next_offset = offset + num_nodes;
|
|
|
|
/* Find the tile that the UV lies in. */
|
|
int tx = (int)tex_co.x;
|
|
int ty = (int)tex_co.y;
|
|
|
|
/* Check that we're within a legitimate tile. */
|
|
if (tx >= 0 && ty >= 0 && tx < 10) {
|
|
int tile = 1001 + 10 * ty + tx;
|
|
|
|
/* Find the index of the tile. */
|
|
for (int i = 0; i < num_nodes; i++) {
|
|
uint4 tile_node = read_node(kg, &offset);
|
|
if (tile_node.x == tile) {
|
|
id = tile_node.y;
|
|
break;
|
|
}
|
|
if (tile_node.z == tile) {
|
|
id = tile_node.w;
|
|
break;
|
|
}
|
|
}
|
|
|
|
/* If we found the tile, offset the UVs to be relative to it. */
|
|
if (id != -1) {
|
|
tex_co.x -= tx;
|
|
tex_co.y -= ty;
|
|
}
|
|
}
|
|
|
|
/* Skip over the remaining nodes. */
|
|
offset = next_offset;
|
|
}
|
|
else {
|
|
id = -num_nodes;
|
|
}
|
|
|
|
float4 f = svm_image_texture(kg, id, tex_co.x, tex_co.y, flags);
|
|
|
|
if (stack_valid(out_offset))
|
|
stack_store_float3(stack, out_offset, make_float3(f.x, f.y, f.z));
|
|
if (stack_valid(alpha_offset))
|
|
stack_store_float(stack, alpha_offset, f.w);
|
|
return offset;
|
|
}
|
|
|
|
ccl_device_noinline void svm_node_tex_image_box(ccl_global const KernelGlobals *kg,
|
|
ccl_private ShaderData *sd,
|
|
ccl_private float *stack,
|
|
uint4 node)
|
|
{
|
|
/* get object space normal */
|
|
float3 N = sd->N;
|
|
|
|
N = sd->N;
|
|
object_inverse_normal_transform(kg, sd, &N);
|
|
|
|
/* project from direction vector to barycentric coordinates in triangles */
|
|
float3 signed_N = N;
|
|
|
|
N.x = fabsf(N.x);
|
|
N.y = fabsf(N.y);
|
|
N.z = fabsf(N.z);
|
|
|
|
N /= (N.x + N.y + N.z);
|
|
|
|
/* basic idea is to think of this as a triangle, each corner representing
|
|
* one of the 3 faces of the cube. in the corners we have single textures,
|
|
* in between we blend between two textures, and in the middle we a blend
|
|
* between three textures.
|
|
*
|
|
* The `Nxyz` values are the barycentric coordinates in an equilateral
|
|
* triangle, which in case of blending, in the middle has a smaller
|
|
* equilateral triangle where 3 textures blend. this divides things into
|
|
* 7 zones, with an if() test for each zone. */
|
|
|
|
float3 weight = make_float3(0.0f, 0.0f, 0.0f);
|
|
float blend = __int_as_float(node.w);
|
|
float limit = 0.5f * (1.0f + blend);
|
|
|
|
/* first test for corners with single texture */
|
|
if (N.x > limit * (N.x + N.y) && N.x > limit * (N.x + N.z)) {
|
|
weight.x = 1.0f;
|
|
}
|
|
else if (N.y > limit * (N.x + N.y) && N.y > limit * (N.y + N.z)) {
|
|
weight.y = 1.0f;
|
|
}
|
|
else if (N.z > limit * (N.x + N.z) && N.z > limit * (N.y + N.z)) {
|
|
weight.z = 1.0f;
|
|
}
|
|
else if (blend > 0.0f) {
|
|
/* in case of blending, test for mixes between two textures */
|
|
if (N.z < (1.0f - limit) * (N.y + N.x)) {
|
|
weight.x = N.x / (N.x + N.y);
|
|
weight.x = saturate((weight.x - 0.5f * (1.0f - blend)) / blend);
|
|
weight.y = 1.0f - weight.x;
|
|
}
|
|
else if (N.x < (1.0f - limit) * (N.y + N.z)) {
|
|
weight.y = N.y / (N.y + N.z);
|
|
weight.y = saturate((weight.y - 0.5f * (1.0f - blend)) / blend);
|
|
weight.z = 1.0f - weight.y;
|
|
}
|
|
else if (N.y < (1.0f - limit) * (N.x + N.z)) {
|
|
weight.x = N.x / (N.x + N.z);
|
|
weight.x = saturate((weight.x - 0.5f * (1.0f - blend)) / blend);
|
|
weight.z = 1.0f - weight.x;
|
|
}
|
|
else {
|
|
/* last case, we have a mix between three */
|
|
weight.x = ((2.0f - limit) * N.x + (limit - 1.0f)) / (2.0f * limit - 1.0f);
|
|
weight.y = ((2.0f - limit) * N.y + (limit - 1.0f)) / (2.0f * limit - 1.0f);
|
|
weight.z = ((2.0f - limit) * N.z + (limit - 1.0f)) / (2.0f * limit - 1.0f);
|
|
}
|
|
}
|
|
else {
|
|
/* Desperate mode, no valid choice anyway, fallback to one side. */
|
|
weight.x = 1.0f;
|
|
}
|
|
|
|
/* now fetch textures */
|
|
uint co_offset, out_offset, alpha_offset, flags;
|
|
svm_unpack_node_uchar4(node.z, &co_offset, &out_offset, &alpha_offset, &flags);
|
|
|
|
float3 co = stack_load_float3(stack, co_offset);
|
|
uint id = node.y;
|
|
|
|
float4 f = make_float4(0.0f, 0.0f, 0.0f, 0.0f);
|
|
|
|
/* Map so that no textures are flipped, rotation is somewhat arbitrary. */
|
|
if (weight.x > 0.0f) {
|
|
float2 uv = make_float2((signed_N.x < 0.0f) ? 1.0f - co.y : co.y, co.z);
|
|
f += weight.x * svm_image_texture(kg, id, uv.x, uv.y, flags);
|
|
}
|
|
if (weight.y > 0.0f) {
|
|
float2 uv = make_float2((signed_N.y > 0.0f) ? 1.0f - co.x : co.x, co.z);
|
|
f += weight.y * svm_image_texture(kg, id, uv.x, uv.y, flags);
|
|
}
|
|
if (weight.z > 0.0f) {
|
|
float2 uv = make_float2((signed_N.z > 0.0f) ? 1.0f - co.y : co.y, co.x);
|
|
f += weight.z * svm_image_texture(kg, id, uv.x, uv.y, flags);
|
|
}
|
|
|
|
if (stack_valid(out_offset))
|
|
stack_store_float3(stack, out_offset, make_float3(f.x, f.y, f.z));
|
|
if (stack_valid(alpha_offset))
|
|
stack_store_float(stack, alpha_offset, f.w);
|
|
}
|
|
|
|
ccl_device_noinline void svm_node_tex_environment(ccl_global const KernelGlobals *kg,
|
|
ccl_private ShaderData *sd,
|
|
ccl_private float *stack,
|
|
uint4 node)
|
|
{
|
|
uint id = node.y;
|
|
uint co_offset, out_offset, alpha_offset, flags;
|
|
uint projection = node.w;
|
|
|
|
svm_unpack_node_uchar4(node.z, &co_offset, &out_offset, &alpha_offset, &flags);
|
|
|
|
float3 co = stack_load_float3(stack, co_offset);
|
|
float2 uv;
|
|
|
|
co = safe_normalize(co);
|
|
|
|
if (projection == 0)
|
|
uv = direction_to_equirectangular(co);
|
|
else
|
|
uv = direction_to_mirrorball(co);
|
|
|
|
float4 f = svm_image_texture(kg, id, uv.x, uv.y, flags);
|
|
|
|
if (stack_valid(out_offset))
|
|
stack_store_float3(stack, out_offset, make_float3(f.x, f.y, f.z));
|
|
if (stack_valid(alpha_offset))
|
|
stack_store_float(stack, alpha_offset, f.w);
|
|
}
|
|
|
|
CCL_NAMESPACE_END
|