commit 8b5b3c1e2ec0f944e2ed8333634bf4dada2e006c
Author: Brecht Van Lommel <brechtvanlommel@gmail.com>
Date:   Tue Jun 4 13:14:13 2013 +0200

    Tmp.

diff --git a/intern/cycles/device/device_cuda.cpp b/intern/cycles/device/device_cuda.cpp
index f32c6dd..27978b9 100644
--- a/intern/cycles/device/device_cuda.cpp
+++ b/intern/cycles/device/device_cuda.cpp
@@ -46,6 +46,7 @@ public:
 	map<device_ptr, bool> tex_interp_map;
 	int cuDevId;
 	bool first_error;
+	vector<CUstream> cuStreams;
 
 	struct PixelMem {
 		GLuint cuPBO;
@@ -205,6 +206,12 @@ public:
 		if(cuda_error_(result, "cuCtxCreate"))
 			return;
 
+		const int num_streams = 8;
+		cuStreams.resize(num_streams);
+
+		for(int i = 0; i < num_streams; i++)
+			cuStreamCreate(&cuStreams[i], 0);
+
 		cuda_pop_context();
 	}
 
@@ -212,6 +219,9 @@ public:
 	{
 		task_pool.stop();
 
+		for(int i = 0; i < cuStreams.size(); i++)
+			cuStreamDestroy(cuStreams[i]);
+
 		cuda_push_context();
 		cuda_assert(cuCtxDetach(cuContext))
 	}
@@ -514,7 +524,7 @@ public:
 		}
 	}
 
-	void path_trace(RenderTile& rtile, int sample)
+	void path_trace(RenderTile& rtile, int sample, CUstream stream)
 	{
 		if(have_error())
 			return;
@@ -575,9 +585,9 @@ public:
 
 		cuda_assert(cuFuncSetCacheConfig(cuPathTrace, CU_FUNC_CACHE_PREFER_L1))
 		cuda_assert(cuFuncSetBlockShape(cuPathTrace, xthreads, ythreads, 1))
-		cuda_assert(cuLaunchGrid(cuPathTrace, xblocks, yblocks))
+		cuda_assert(cuLaunchGridAsync(cuPathTrace, xblocks, yblocks, stream))
 
-		cuda_assert(cuCtxSynchronize())
+		//cuda_assert(cuCtxSynchronize())
 
 		cuda_pop_context();
 	}
@@ -882,12 +892,35 @@ public:
 	void thread_run(DeviceTask *task)
 	{
 		if(task->type == DeviceTask::PATH_TRACE) {
-			RenderTile tile;
+			vector<RenderTile> concurrent_tiles(cuStreams.size());
+			vector<bool> have_tile(cuStreams.size());
 			
 			/* keep rendering tiles until done */
-			while(task->acquire_tile(this, tile)) {
-				int start_sample = tile.start_sample;
-				int end_sample = tile.start_sample + tile.num_samples;
+			while(1) {
+				int start_sample = -1;
+				int end_sample = -1;
+
+				for(int i = 0; i < concurrent_tiles.size(); i++) {
+					RenderTile& tile = concurrent_tiles[i];
+
+					if(task->acquire_tile(this, tile)) {
+						have_tile[i] = true;
+
+						if(start_sample == -1) {
+							start_sample = tile.start_sample;
+							end_sample = tile.start_sample + tile.num_samples;
+						}
+						else {
+							start_sample = min(start_sample, tile.start_sample);
+							end_sample = max(end_sample, tile.start_sample + tile.num_samples);
+						}
+					}
+					else
+						have_tile[i] = false;
+				}
+
+				if(start_sample == -1)
+					break;
 
 				for(int sample = start_sample; sample < end_sample; sample++) {
 					if (task->get_cancel()) {
@@ -895,21 +928,35 @@ public:
 							break;
 					}
 
-					path_trace(tile, sample);
+					for(int i = 0; i < concurrent_tiles.size(); i++) {
+						if(have_tile[i]) {
+							RenderTile& tile = concurrent_tiles[i];
+							int tile_end_sample = tile.start_sample + tile.num_samples;
 
-					tile.sample = sample + 1;
+							if(sample > tile.start_sample && sample < tile_end_sample) {
+								path_trace(tile, sample, cuStreams[i]);
+								tile.sample = sample + 1;
 
-					task->update_progress(tile);
+								if(i == 0)
+									task->update_progress(tile);
+							}
+						}
+					}
 				}
 
-				task->release_tile(tile);
+				for(int i = 0; i < concurrent_tiles.size(); i++) {
+					if(have_tile[i]) {
+						RenderTile& tile = concurrent_tiles[i];
+						task->release_tile(tile);
+					}
+				}
 			}
 		}
 		else if(task->type == DeviceTask::SHADER) {
 			shader(*task);
 
 			cuda_push_context();
-			cuda_assert(cuCtxSynchronize())
+			//cuda_assert(cuCtxSynchronize())
 			cuda_pop_context();
 		}
 	}
@@ -930,7 +977,7 @@ public:
 			tonemap(task, task.buffer, task.rgba);
 
 			cuda_push_context();
-			cuda_assert(cuCtxSynchronize())
+			//cuda_assert(cuCtxSynchronize())
 			cuda_pop_context();
 		}
 		else {