diff options
author | Howard Trickey <howard.trickey@gmail.com> | 2021-10-30 22:37:05 +0300 |
---|---|---|
committer | Howard Trickey <howard.trickey@gmail.com> | 2021-10-30 22:37:05 +0300 |
commit | e9bbfd0c8c7a508d220bf355722ff03f91e93183 (patch) | |
tree | 1230f26bc82f24547aeccbaa7fcd6d3db2655fd3 /intern/cycles/kernel/film/write_passes.h | |
parent | 1aa953bd1913c81b22c80a00edbf4ad88a32c52f (diff) | |
parent | 03a962d8cab44221650f59eb223cb0a767e05b2b (diff) |
Merge branch 'master' into soc-2020-io-performancesoc-2020-io-performance
Diffstat (limited to 'intern/cycles/kernel/film/write_passes.h')
-rw-r--r-- | intern/cycles/kernel/film/write_passes.h | 88 |
1 files changed, 88 insertions, 0 deletions
diff --git a/intern/cycles/kernel/film/write_passes.h b/intern/cycles/kernel/film/write_passes.h new file mode 100644 index 00000000000..9d379495629 --- /dev/null +++ b/intern/cycles/kernel/film/write_passes.h @@ -0,0 +1,88 @@ +/* + * 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. + */ + +#pragma once + +#ifdef __KERNEL_GPU__ +# define __ATOMIC_PASS_WRITE__ +#endif + +CCL_NAMESPACE_BEGIN + +ccl_device_inline void kernel_write_pass_float(ccl_global float *ccl_restrict buffer, float value) +{ +#ifdef __ATOMIC_PASS_WRITE__ + atomic_add_and_fetch_float(buffer, value); +#else + *buffer += value; +#endif +} + +ccl_device_inline void kernel_write_pass_float3(ccl_global float *ccl_restrict buffer, + float3 value) +{ +#ifdef __ATOMIC_PASS_WRITE__ + ccl_global float *buf_x = buffer + 0; + ccl_global float *buf_y = buffer + 1; + ccl_global float *buf_z = buffer + 2; + + atomic_add_and_fetch_float(buf_x, value.x); + atomic_add_and_fetch_float(buf_y, value.y); + atomic_add_and_fetch_float(buf_z, value.z); +#else + buffer[0] += value.x; + buffer[1] += value.y; + buffer[2] += value.z; +#endif +} + +ccl_device_inline void kernel_write_pass_float4(ccl_global float *ccl_restrict buffer, + float4 value) +{ +#ifdef __ATOMIC_PASS_WRITE__ + ccl_global float *buf_x = buffer + 0; + ccl_global float *buf_y = buffer + 1; + ccl_global float *buf_z = buffer + 2; + ccl_global float *buf_w = buffer + 3; + + atomic_add_and_fetch_float(buf_x, value.x); + atomic_add_and_fetch_float(buf_y, value.y); + atomic_add_and_fetch_float(buf_z, value.z); + atomic_add_and_fetch_float(buf_w, value.w); +#else + buffer[0] += value.x; + buffer[1] += value.y; + buffer[2] += value.z; + buffer[3] += value.w; +#endif +} + +ccl_device_inline float kernel_read_pass_float(ccl_global float *ccl_restrict buffer) +{ + return *buffer; +} + +ccl_device_inline float3 kernel_read_pass_float3(ccl_global float *ccl_restrict buffer) +{ + return make_float3(buffer[0], buffer[1], buffer[2]); +} + +ccl_device_inline float4 kernel_read_pass_float4(ccl_global float *ccl_restrict buffer) +{ + return make_float4(buffer[0], buffer[1], buffer[2], buffer[3]); +} + +CCL_NAMESPACE_END |