mirror of
https://github.com/blender/blender
synced 2026-09-29 04:37:17 +03:00
5fc85d1cb7
The new intersect_mnee kernel runs before shade_surface, and shade_surface_mnee is eliminated. That large kernel was causing problems for some GPU compilers. MNEE state is packed into a shadow path state to avoid significantly increasing the path state size. This shadow state is then either turned into an actual shadow ray state or discarded in shade_surface. MNEE was re-enabled on HIP RDNA2 as it works again now. Texture cache misses now also work correctly with MNEE. This adds some extra code to the regular shade_surface kernel even when MNEE is not used, to use the MNEE sampled point instead of sampling a light. But there seems to be no significant performance impact. Co-authored-by: Sergey Sharybin <sergey@blender.org> Pull Request: https://projects.blender.org/blender/blender/pulls/158698
38 lines
844 B
C++
38 lines
844 B
C++
/* SPDX-FileCopyrightText: 2011-2022 Blender Foundation
|
|
*
|
|
* SPDX-License-Identifier: Apache-2.0 */
|
|
|
|
#pragma once
|
|
|
|
#ifndef __KERNEL_ONEAPI__
|
|
# include "kernel/types.h"
|
|
|
|
# include "util/string.h"
|
|
|
|
# include <bitset>
|
|
# include <iosfwd>
|
|
#endif
|
|
|
|
CCL_NAMESPACE_BEGIN
|
|
|
|
/* DeviceKernel */
|
|
|
|
bool device_kernel_has_shading(DeviceKernel kernel);
|
|
bool device_kernel_has_intersection(DeviceKernel kernel);
|
|
bool device_kernel_has_gpu_function(DeviceKernel kernel);
|
|
|
|
const char *device_kernel_as_string(DeviceKernel kernel);
|
|
|
|
#ifndef __KERNEL_ONEAPI__
|
|
std::ostream &operator<<(std::ostream &os, DeviceKernel kernel);
|
|
|
|
/* DeviceKernelMask */
|
|
|
|
struct DeviceKernelMask : public std::bitset<DEVICE_KERNEL_NUM> {
|
|
bool operator<(const DeviceKernelMask &other) const;
|
|
};
|
|
|
|
string device_kernel_mask_as_string(DeviceKernelMask mask);
|
|
#endif
|
|
|
|
CCL_NAMESPACE_END
|