Implemented CLT accumulation for virtual orientation

src/main/java/com/elphel/imagej/gpu/GpuQuad.java

@@ -1114,6 +1114,17 @@ public class GpuQuad{ // quad camera description
 		copyD2H.Height =          img_height; // /4;
 		cuMemcpy2D(copyD2H); // run copy
 	}
+	public int [] getWH(boolean use_ref) {
+		return use_ref ? gpu_clt_ref_wh : gpu_clt_wh;
+	}
+	
+	public int getNumColors() {
+		return num_colors;
+	}
+
+	public int getNumSensors() {
+		return num_cams;
+	}
 	
 	public int getCltSize(boolean use_ref) { // per camera, in floats
 		int [] wh = use_ref ? gpu_clt_ref_wh : gpu_clt_wh;
@@ -1121,7 +1132,16 @@ public class GpuQuad{ // quad camera description
 		int tilesY =  wh[1] / GPUTileProcessor.DTT_SIZE;
 		return tilesY*tilesX*num_colors* 4 * GPUTileProcessor.DTT_SIZE * GPUTileProcessor.DTT_SIZE;
 	}
-	
+	/*
+	public int getCltLength(boolean use_ref) {
+		int [] wh = use_ref ? gpu_clt_ref_wh : gpu_clt_wh;
+		int tilesX =  wh[0] / GPUTileProcessor.DTT_SIZE;
+		int tilesY =  wh[1] / GPUTileProcessor.DTT_SIZE;
+		int tile_size_td = getCltSize(use_ref); // 4 * GPUTileProcessor.DTT_SIZE * GPUTileProcessor.DTT_SIZE;
+		int num_tiles = tilesY*tilesX*num_colors;
+		return num_tiles* tile_size_td;
+	}
+	*/
 	public float [][] getCltData( // only for color=0
 			boolean use_ref){
 		CUdeviceptr [] gpu_sel_clt_h = 	use_ref ? gpu_clt_ref_h : gpu_clt_h;
@@ -1209,24 +1229,36 @@ public class GpuQuad{ // quad camera description
 		return;
 	}
 
-/*
-	public void setBayerImage(
-			float [] bayer_image,
-			int ncam) {
+	public void setCltData( // for testing only
+			int      ncam,
+			float [] fclt, //
+			boolean  use_ref){
+		int clt_size = getCltSize(use_ref);
+		if (fclt.length != clt_size) {
+			System.out.println("getCltData(): wrong array size: got ["+fclt.length+"], "+
+			"should be ["+clt_size+"]");
+			return;
+		}
+		
+		CUdeviceptr [] gpu_sel_clt_h = 	use_ref ? gpu_clt_ref_h : gpu_clt_h;
+		int [] wh = use_ref ? gpu_clt_ref_wh : gpu_clt_wh;
+		int tilesX =  wh[0] / GPUTileProcessor.DTT_SIZE;
+		int tilesY =  wh[1] / GPUTileProcessor.DTT_SIZE;
+		int tile_size_td = 4 * GPUTileProcessor.DTT_SIZE * GPUTileProcessor.DTT_SIZE;
+		int num_tiles = tilesY*tilesX*num_colors;
 		CUDA_MEMCPY2D copyH2D =   new CUDA_MEMCPY2D();
 		copyH2D.srcMemoryType =   CUmemorytype.CU_MEMORYTYPE_HOST;
-		copyH2D.srcHost =         Pointer.to(bayer_image);
-		copyH2D.srcPitch =        img_width*Sizeof.FLOAT; // width_in_bytes;
+		copyH2D.srcHost =         Pointer.to(fclt);
+		copyH2D.srcPitch =        tile_size_td * Sizeof.FLOAT;
 		copyH2D.dstMemoryType =   CUmemorytype.CU_MEMORYTYPE_DEVICE;
-		copyH2D.dstDevice =       gpu_bayer_h[ncam]; // src_dpointer;
-		copyH2D.dstPitch =        mclt_stride *Sizeof.FLOAT; // device_stride[0];
-		copyH2D.WidthInBytes =    img_width*Sizeof.FLOAT; // width_in_bytes;
-		copyH2D.Height =          img_height; // /4;
-		cuMemcpy2D(copyH2D);
+		copyH2D.dstDevice =       gpu_sel_clt_h[ncam];
+		copyH2D.dstPitch =        tile_size_td * Sizeof.FLOAT;
+		copyH2D.WidthInBytes =    tile_size_td * Sizeof.FLOAT;
+		copyH2D.Height =          num_tiles;
+		cuMemcpy2D(copyH2D); // run copy
+		return;
 	}
 	
- */
-	
 	
 	/**
 	 * Copy a set of images to the GPU (if they are new)