main sync #3

Merged
Patrick Busch merged 318 commits from blender/blender:main into main 2023-03-17 15:52:21 +01:00
4 changed files with 11 additions and 7 deletions
Showing only changes of commit 089e8a1887 - Show all commits

View File

@ -567,6 +567,10 @@ void MetalDevice::compile_and_load(int device_id, MetalPipelineType pso_type)
thread_scoped_lock lock(existing_devices_mutex);
if (MetalDevice *instance = get_device_by_ID(device_id, lock)) {
if (mtlLibrary) {
if (error && [error localizedDescription]) {
VLOG_WARNING << "MSL compilation messages: " << [[error localizedDescription] UTF8String];
}
instance->mtlLibrary[pso_type] = mtlLibrary;
starttime = time_dt();

View File

@ -715,7 +715,7 @@ void MetalKernelPipeline::compile()
}
}
};
if (computePipelineStateDescriptor.linkedFunctions) {
if (linked_functions) {
addComputePipelineFunctionsWithDescriptor();
}
@ -748,7 +748,7 @@ void MetalKernelPipeline::compile()
}
/* Add pipeline into the new archive (unless we did it earlier). */
if (pipeline && !computePipelineStateDescriptor.linkedFunctions) {
if (pipeline && !linked_functions) {
addComputePipelineFunctionsWithDescriptor();
}
}

View File

@ -38,7 +38,7 @@ ccl_device_inline void gpu_parallel_sort_bucket_pass(const uint num_states,
ccl_gpu_shared int *buckets,
const ushort local_id,
const ushort local_size,
const ushort grid_id)
const uint grid_id)
{
/* Zero the bucket sizes. */
if (local_id < max_shaders) {
@ -89,7 +89,7 @@ ccl_device_inline void gpu_parallel_sort_write_pass(const uint num_states,
ccl_gpu_shared int *local_offset,
const ushort local_id,
const ushort local_size,
const ushort grid_id)
const uint grid_id)
{
/* Calculate each partition's global offset from the prefix sum of the active state counts per
* partition. */

View File

@ -109,7 +109,7 @@ struct kernel_gpu_##name \
const uint metal_global_id, \
const ushort metal_local_id, \
const ushort metal_local_size, \
const ushort metal_grid_id, \
const uint metal_grid_id, \
uint simdgroup_size, \
uint simd_lane_index, \
uint simd_group_index, \
@ -122,7 +122,7 @@ kernel void cycles_metal_##name(device const kernel_gpu_##name *params_struct, \
const uint metal_global_id [[thread_position_in_grid]], \
const ushort metal_local_id [[thread_position_in_threadgroup]], \
const ushort metal_local_size [[threads_per_threadgroup]], \
const ushort metal_grid_id [[threadgroup_position_in_grid]], \
const uint metal_grid_id [[threadgroup_position_in_grid]], \
uint simdgroup_size [[threads_per_simdgroup]], \
uint simd_lane_index [[thread_index_in_simdgroup]], \
uint simd_group_index [[simdgroup_index_in_threadgroup]], \
@ -135,7 +135,7 @@ void kernel_gpu_##name::run(thread MetalKernelContext& context, \
const uint metal_global_id, \
const ushort metal_local_id, \
const ushort metal_local_size, \
const ushort metal_grid_id, \
const uint metal_grid_id, \
uint simdgroup_size, \
uint simd_lane_index, \
uint simd_group_index, \