From e12c08e8d170b7ca40f204a5b0423c23a9fbc2c1 Mon Sep 17 00:00:00 2001 From: Campbell Barton Date: Wed, 17 Apr 2019 06:17:24 +0200 Subject: ClangFormat: apply to source, most of intern Apply clang format as proposed in T53211. For details on usage and instructions for migrating branches without conflicts, see: https://wiki.blender.org/wiki/Tools/ClangFormat --- intern/cycles/kernel/split/kernel_shader_sort.h | 134 ++++++++++++------------ 1 file changed, 67 insertions(+), 67 deletions(-) (limited to 'intern/cycles/kernel/split/kernel_shader_sort.h') diff --git a/intern/cycles/kernel/split/kernel_shader_sort.h b/intern/cycles/kernel/split/kernel_shader_sort.h index 666355de334..95d33a42014 100644 --- a/intern/cycles/kernel/split/kernel_shader_sort.h +++ b/intern/cycles/kernel/split/kernel_shader_sort.h @@ -16,82 +16,82 @@ CCL_NAMESPACE_BEGIN - -ccl_device void kernel_shader_sort(KernelGlobals *kg, - ccl_local_param ShaderSortLocals *locals) +ccl_device void kernel_shader_sort(KernelGlobals *kg, ccl_local_param ShaderSortLocals *locals) { #ifndef __KERNEL_CUDA__ - int tid = ccl_global_id(1) * ccl_global_size(0) + ccl_global_id(0); - uint qsize = kernel_split_params.queue_index[QUEUE_ACTIVE_AND_REGENERATED_RAYS]; - if(tid == 0) { - kernel_split_params.queue_index[QUEUE_SHADER_SORTED_RAYS] = qsize; - } + int tid = ccl_global_id(1) * ccl_global_size(0) + ccl_global_id(0); + uint qsize = kernel_split_params.queue_index[QUEUE_ACTIVE_AND_REGENERATED_RAYS]; + if (tid == 0) { + kernel_split_params.queue_index[QUEUE_SHADER_SORTED_RAYS] = qsize; + } - uint offset = (tid/SHADER_SORT_LOCAL_SIZE)*SHADER_SORT_BLOCK_SIZE; - if(offset >= qsize) { - return; - } + uint offset = (tid / SHADER_SORT_LOCAL_SIZE) * SHADER_SORT_BLOCK_SIZE; + if (offset >= qsize) { + return; + } - int lid = ccl_local_id(1) * ccl_local_size(0) + ccl_local_id(0); - uint input = QUEUE_ACTIVE_AND_REGENERATED_RAYS * (kernel_split_params.queue_size); - uint output = QUEUE_SHADER_SORTED_RAYS * (kernel_split_params.queue_size); - ccl_local uint *local_value = &locals->local_value[0]; - ccl_local ushort *local_index = &locals->local_index[0]; + int lid = ccl_local_id(1) * ccl_local_size(0) + ccl_local_id(0); + uint input = QUEUE_ACTIVE_AND_REGENERATED_RAYS * (kernel_split_params.queue_size); + uint output = QUEUE_SHADER_SORTED_RAYS * (kernel_split_params.queue_size); + ccl_local uint *local_value = &locals->local_value[0]; + ccl_local ushort *local_index = &locals->local_index[0]; - /* copy to local memory */ - for(uint i = 0; i < SHADER_SORT_BLOCK_SIZE; i += SHADER_SORT_LOCAL_SIZE) { - uint idx = offset + i + lid; - uint add = input + idx; - uint value = (~0); - if(idx < qsize) { - int ray_index = kernel_split_state.queue_data[add]; - bool valid = (ray_index != QUEUE_EMPTY_SLOT) && IS_STATE(kernel_split_state.ray_state, ray_index, RAY_ACTIVE); - if(valid) { - value = kernel_split_sd(sd, ray_index)->shader & SHADER_MASK; - } - } - local_value[i + lid] = value; - local_index[i + lid] = i + lid; - } - ccl_barrier(CCL_LOCAL_MEM_FENCE); + /* copy to local memory */ + for (uint i = 0; i < SHADER_SORT_BLOCK_SIZE; i += SHADER_SORT_LOCAL_SIZE) { + uint idx = offset + i + lid; + uint add = input + idx; + uint value = (~0); + if (idx < qsize) { + int ray_index = kernel_split_state.queue_data[add]; + bool valid = (ray_index != QUEUE_EMPTY_SLOT) && + IS_STATE(kernel_split_state.ray_state, ray_index, RAY_ACTIVE); + if (valid) { + value = kernel_split_sd(sd, ray_index)->shader & SHADER_MASK; + } + } + local_value[i + lid] = value; + local_index[i + lid] = i + lid; + } + ccl_barrier(CCL_LOCAL_MEM_FENCE); - /* skip sorting for cpu split kernel */ + /* skip sorting for cpu split kernel */ # ifdef __KERNEL_OPENCL__ - /* bitonic sort */ - for(uint length = 1; length < SHADER_SORT_BLOCK_SIZE; length <<= 1) { - for(uint inc = length; inc > 0; inc >>= 1) { - for(uint ii = 0; ii < SHADER_SORT_BLOCK_SIZE; ii += SHADER_SORT_LOCAL_SIZE) { - uint i = lid + ii; - bool direction = ((i & (length << 1)) != 0); - uint j = i ^ inc; - ushort ioff = local_index[i]; - ushort joff = local_index[j]; - uint iKey = local_value[ioff]; - uint jKey = local_value[joff]; - bool smaller = (jKey < iKey) || (jKey == iKey && j < i); - bool swap = smaller ^ (j < i) ^ direction; - ccl_barrier(CCL_LOCAL_MEM_FENCE); - local_index[i] = (swap) ? joff : ioff; - local_index[j] = (swap) ? ioff : joff; - ccl_barrier(CCL_LOCAL_MEM_FENCE); - } - } - } -# endif /* __KERNEL_OPENCL__ */ + /* bitonic sort */ + for (uint length = 1; length < SHADER_SORT_BLOCK_SIZE; length <<= 1) { + for (uint inc = length; inc > 0; inc >>= 1) { + for (uint ii = 0; ii < SHADER_SORT_BLOCK_SIZE; ii += SHADER_SORT_LOCAL_SIZE) { + uint i = lid + ii; + bool direction = ((i & (length << 1)) != 0); + uint j = i ^ inc; + ushort ioff = local_index[i]; + ushort joff = local_index[j]; + uint iKey = local_value[ioff]; + uint jKey = local_value[joff]; + bool smaller = (jKey < iKey) || (jKey == iKey && j < i); + bool swap = smaller ^ (j < i) ^ direction; + ccl_barrier(CCL_LOCAL_MEM_FENCE); + local_index[i] = (swap) ? joff : ioff; + local_index[j] = (swap) ? ioff : joff; + ccl_barrier(CCL_LOCAL_MEM_FENCE); + } + } + } +# endif /* __KERNEL_OPENCL__ */ - /* copy to destination */ - for(uint i = 0; i < SHADER_SORT_BLOCK_SIZE; i += SHADER_SORT_LOCAL_SIZE) { - uint idx = offset + i + lid; - uint lidx = local_index[i + lid]; - uint outi = output + idx; - uint ini = input + offset + lidx; - uint value = local_value[lidx]; - if(idx < qsize) { - kernel_split_state.queue_data[outi] = (value == (~0)) ? QUEUE_EMPTY_SLOT : kernel_split_state.queue_data[ini]; - } - } -#endif /* __KERNEL_CUDA__ */ + /* copy to destination */ + for (uint i = 0; i < SHADER_SORT_BLOCK_SIZE; i += SHADER_SORT_LOCAL_SIZE) { + uint idx = offset + i + lid; + uint lidx = local_index[i + lid]; + uint outi = output + idx; + uint ini = input + offset + lidx; + uint value = local_value[lidx]; + if (idx < qsize) { + kernel_split_state.queue_data[outi] = (value == (~0)) ? QUEUE_EMPTY_SLOT : + kernel_split_state.queue_data[ini]; + } + } +#endif /* __KERNEL_CUDA__ */ } CCL_NAMESPACE_END -- cgit v1.2.3