Files
blender/intern/cycles/kernel/integrator/volume_stack.h

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

213 lines
6.7 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: Apache-2.0
* Copyright 2011-2022 Blender Foundation */
#pragma once
CCL_NAMESPACE_BEGIN
/* Volumetric read/write lambda functions - default implementations */
#ifndef VOLUME_READ_LAMBDA
# define VOLUME_READ_LAMBDA(function_call) \
auto volume_read_lambda_pass = [=](const int i) { return function_call; };
# define VOLUME_WRITE_LAMBDA(function_call) \
auto volume_write_lambda_pass = [=](const int i, VolumeStack entry) { function_call; };
#endif
/* Volume Stack
*
* This is an array of object/shared ID's that the current segment of the path
* is inside of. */
template<typename StackReadOp, typename StackWriteOp>
ccl_device void volume_stack_enter_exit(KernelGlobals kg,
Cycles: Kernel address space changes for MSL 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
2021-10-14 13:53:40 +01:00
ccl_private const ShaderData *sd,
StackReadOp stack_read,
StackWriteOp stack_write)
{
/* todo: we should have some way for objects to indicate if they want the
* world shader to work inside them. excluding it by default is problematic
* because non-volume objects can't be assumed to be closed manifolds */
if (!(sd->flag & SD_HAS_VOLUME)) {
return;
}
if (sd->flag & SD_BACKFACING) {
/* Exit volume object: remove from stack. */
for (int i = 0;; i++) {
VolumeStack entry = stack_read(i);
if (entry.shader == SHADER_NONE) {
break;
}
if (entry.object == sd->object) {
/* Shift back next stack entries. */
do {
entry = stack_read(i + 1);
stack_write(i, entry);
i++;
} while (entry.shader != SHADER_NONE);
return;
}
}
}
else {
/* Enter volume object: add to stack. */
int i;
for (i = 0;; i++) {
VolumeStack entry = stack_read(i);
if (entry.shader == SHADER_NONE) {
break;
}
/* Already in the stack? then we have nothing to do. */
if (entry.object == sd->object) {
return;
}
}
/* If we exceed the stack limit, ignore. */
if (i >= kernel_data.volume_stack_size - 1) {
return;
}
/* Add to the end of the stack. */
const VolumeStack new_entry = {sd->object, sd->shader};
const VolumeStack empty_entry = {OBJECT_NONE, SHADER_NONE};
stack_write(i, new_entry);
stack_write(i + 1, empty_entry);
}
}
ccl_device void volume_stack_enter_exit(KernelGlobals kg,
IntegratorState state,
ccl_private const ShaderData *sd)
{
VOLUME_READ_LAMBDA(integrator_state_read_volume_stack(state, i))
VOLUME_WRITE_LAMBDA(integrator_state_write_volume_stack(state, i, entry))
volume_stack_enter_exit(kg, sd, volume_read_lambda_pass, volume_write_lambda_pass);
}
ccl_device void shadow_volume_stack_enter_exit(KernelGlobals kg,
IntegratorShadowState state,
Cycles: Kernel address space changes for MSL 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
2021-10-14 13:53:40 +01:00
ccl_private const ShaderData *sd)
{
VOLUME_READ_LAMBDA(integrator_state_read_shadow_volume_stack(state, i))
VOLUME_WRITE_LAMBDA(integrator_state_write_shadow_volume_stack(state, i, entry))
volume_stack_enter_exit(kg, sd, volume_read_lambda_pass, volume_write_lambda_pass);
}
/* Clean stack after the last bounce.
*
* It is expected that all volumes are closed manifolds, so at the time when ray
* hits nothing (for example, it is a last bounce which goes to environment) the
* only expected volume in the stack is the world's one. All the rest volume
* entries should have been exited already.
*
* This isn't always true because of ray intersection precision issues, which
* could lead us to an infinite non-world volume in the stack, causing render
* artifacts.
*
* Use this function after the last bounce to get rid of all volumes apart from
* the world's one after the last bounce to avoid render artifacts.
*/
ccl_device_inline void volume_stack_clean(KernelGlobals kg, IntegratorState state)
{
if (kernel_data.background.volume_shader != SHADER_NONE) {
/* Keep the world's volume in stack. */
INTEGRATOR_STATE_ARRAY_WRITE(state, volume_stack, 1, shader) = SHADER_NONE;
}
else {
INTEGRATOR_STATE_ARRAY_WRITE(state, volume_stack, 0, shader) = SHADER_NONE;
}
}
template<typename StackReadOp>
ccl_device float volume_stack_step_size(KernelGlobals kg, StackReadOp stack_read)
{
float step_size = FLT_MAX;
for (int i = 0;; i++) {
VolumeStack entry = stack_read(i);
if (entry.shader == SHADER_NONE) {
break;
}
int shader_flag = kernel_tex_fetch(__shaders, (entry.shader & SHADER_MASK)).flags;
bool heterogeneous = false;
if (shader_flag & SD_HETEROGENEOUS_VOLUME) {
heterogeneous = true;
}
else if (shader_flag & SD_NEED_VOLUME_ATTRIBUTES) {
/* We want to render world or objects without any volume grids
* as homogeneous, but can only verify this at run-time since other
* heterogeneous volume objects may be using the same shader. */
int object = entry.object;
if (object != OBJECT_NONE) {
int object_flag = kernel_tex_fetch(__object_flag, object);
if (object_flag & SD_OBJECT_HAS_VOLUME_ATTRIBUTES) {
heterogeneous = true;
}
}
}
if (heterogeneous) {
float object_step_size = object_volume_step_size(kg, entry.object);
object_step_size *= kernel_data.integrator.volume_step_rate;
step_size = fminf(object_step_size, step_size);
}
}
return step_size;
}
typedef enum VolumeSampleMethod {
VOLUME_SAMPLE_NONE = 0,
VOLUME_SAMPLE_DISTANCE = (1 << 0),
VOLUME_SAMPLE_EQUIANGULAR = (1 << 1),
VOLUME_SAMPLE_MIS = (VOLUME_SAMPLE_DISTANCE | VOLUME_SAMPLE_EQUIANGULAR),
} VolumeSampleMethod;
ccl_device VolumeSampleMethod volume_stack_sample_method(KernelGlobals kg, IntegratorState state)
{
VolumeSampleMethod method = VOLUME_SAMPLE_NONE;
for (int i = 0;; i++) {
VolumeStack entry = integrator_state_read_volume_stack(state, i);
if (entry.shader == SHADER_NONE) {
break;
}
int shader_flag = kernel_tex_fetch(__shaders, (entry.shader & SHADER_MASK)).flags;
if (shader_flag & SD_VOLUME_MIS) {
/* Multiple importance sampling. */
return VOLUME_SAMPLE_MIS;
}
else if (shader_flag & SD_VOLUME_EQUIANGULAR) {
/* Distance + equiangular sampling -> multiple importance sampling. */
if (method == VOLUME_SAMPLE_DISTANCE) {
return VOLUME_SAMPLE_MIS;
}
/* Only equiangular sampling. */
method = VOLUME_SAMPLE_EQUIANGULAR;
}
else {
/* Distance + equiangular sampling -> multiple importance sampling. */
if (method == VOLUME_SAMPLE_EQUIANGULAR) {
return VOLUME_SAMPLE_MIS;
}
/* Distance sampling only. */
method = VOLUME_SAMPLE_DISTANCE;
}
}
return method;
}
CCL_NAMESPACE_END