This repository has been archived on 2023-10-09. You can view files and clone it. You cannot open issues or pull requests or push a commit.
Files
blender-archive/intern/cycles/kernel/device/gpu/parallel_prefix_sum.h
Brecht Van Lommel 9cfc7967dd Cycles: use SPDX license headers
* Replace license text in headers with SPDX identifiers.
* Remove specific license info from outdated readme.txt, instead leave details
  to the source files.
* Add list of SPDX license identifiers used, and corresponding license texts.
* Update copyright dates while we're at it.

Ref D14069, T95597
2022-02-11 17:47:34 +01:00

42 lines
1022 B
C++

/* SPDX-License-Identifier: Apache-2.0
* Copyright 2021-2022 Blender Foundation */
#pragma once
CCL_NAMESPACE_BEGIN
/* Parallel prefix sum.
*
* TODO: actually make this work in parallel.
*
* This is used for an array the size of the number of shaders in the scene
* which is not usually huge, so might not be a significant bottleneck. */
#include "util/atomic.h"
#ifdef __HIP__
# define GPU_PARALLEL_PREFIX_SUM_DEFAULT_BLOCK_SIZE 1024
#else
# define GPU_PARALLEL_PREFIX_SUM_DEFAULT_BLOCK_SIZE 512
#endif
__device__ void gpu_parallel_prefix_sum(const int global_id,
ccl_global int *counter,
ccl_global int *prefix_sum,
const int num_values)
{
if (global_id != 0) {
return;
}
int offset = 0;
for (int i = 0; i < num_values; i++) {
const int new_offset = offset + counter[i];
prefix_sum[i] = offset;
counter[i] = 0;
offset = new_offset;
}
}
CCL_NAMESPACE_END