void thread_shader(DeviceTask& task)
{
KernelGlobals kg = kernel_globals;
#ifdef WITH_OSL
OSLShader::thread_init(&kg, &kernel_globals, &osl_globals);
#endif
#ifdef WITH_OPTIMIZED_KERNEL
if(system_cpu_support_optimized()) {
for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
kernel_cpu_optimized_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x);
if(task_pool.cancelled())
break;
}
}
else
#endif
{
for(int x = task.shader_x; x < task.shader_x + task.shader_w; x++) {
kernel_cpu_shader(&kg, (uint4*)task.shader_input, (float4*)task.shader_output, task.shader_eval_type, x);
if(task_pool.cancelled())
break;
}
}
#ifdef WITH_OSL
OSLShader::thread_free(&kg);
#endif
}
void thread_path_trace(DeviceTask& task)
{
if(task_pool.cancelled()) {
if(task.need_finish_queue == false)
return;
}
KernelGlobals kg = kernel_globals;
#ifdef WITH_OSL
OSLShader::thread_init(&kg, &kernel_globals, &osl_globals);
#endif
RenderTile tile;
while(task.acquire_tile(this, tile)) {
float *render_buffer = (float*)tile.buffer;
uint *rng_state = (uint*)tile.rng_state;
int start_sample = tile.start_sample;
int end_sample = tile.start_sample + tile.num_samples;
#ifdef WITH_OPTIMIZED_KERNEL
if(system_cpu_support_sse3()) {
for(int sample = start_sample; sample < end_sample; sample++) {
if (task.get_cancel() || task_pool.cancelled()) {
if(task.need_finish_queue == false)
break;
}
for(int y = tile.y; y < tile.y + tile.h; y++) {
for(int x = tile.x; x < tile.x + tile.w; x++) {
kernel_cpu_sse3_path_trace(&kg, render_buffer, rng_state,
sample, x, y, tile.offset, tile.stride);
}
}
tile.sample = sample + 1;
task.update_progress(tile);
}
}
else if(system_cpu_support_sse2()) {
for(int sample = start_sample; sample < end_sample; sample++) {
if (task.get_cancel() || task_pool.cancelled()) {
if(task.need_finish_queue == false)
break;
}
for(int y = tile.y; y < tile.y + tile.h; y++) {
for(int x = tile.x; x < tile.x + tile.w; x++) {
kernel_cpu_sse2_path_trace(&kg, render_buffer, rng_state,
sample, x, y, tile.offset, tile.stride);
}
}
tile.sample = sample + 1;
task.update_progress(tile);
}
}
else
#endif
{
for(int sample = start_sample; sample < end_sample; sample++) {
if (task.get_cancel() || task_pool.cancelled()) {
if(task.need_finish_queue == false)
break;
}
for(int y = tile.y; y < tile.y + tile.h; y++) {
for(int x = tile.x; x < tile.x + tile.w; x++) {
kernel_cpu_path_trace(&kg, render_buffer, rng_state,
sample, x, y, tile.offset, tile.stride);
}
}
tile.sample = sample + 1;
task.update_progress(tile);
}
}
task.release_tile(tile);
if(task_pool.cancelled()) {
if(task.need_finish_queue == false)
break;
}
}
#ifdef WITH_OSL
OSLShader::thread_free(&kg);
#endif
}