tidying up

2025-12-16 18:14:25 +00:00 · 2010-10-23 18:29:06 +00:00
parent 349123ec19
commit 3ccf418f6c
2 changed files with 207 additions and 213 deletions
--- a/CUETools.Codecs.FLACCL/FLACCLWriter.cs
+++ b/CUETools.Codecs.FLACCL/FLACCLWriter.cs
@@ -1336,8 +1336,8 @@ namespace CUETools.Codecs.FLACCL
 			frame_count += nFrames;
 			frame_pos += nFrames * blocksize;
 			task.openCLCQ.EnqueueWriteBuffer(task.clSamplesBytes, false, 0, sizeof(short) * channels * blocksize * nFrames, task.clSamplesBytes.HostPtr);
-			//task.openCLCQ.EnqueueUnmapMemObject(task.cudaSamplesBytes, task.cudaSamplesBytes.HostPtr);
-			//task.openCLCQ.EnqueueMapBuffer(task.cudaSamplesBytes, true, MapFlags.WRITE, 0, task.samplesBufferLen / 2);
+			//task.openCLCQ.EnqueueUnmapMemObject(task.clSamplesBytes, task.clSamplesBytes.HostPtr);
+			//task.openCLCQ.EnqueueMapBuffer(task.clSamplesBytes, true, MapFlags.WRITE, 0, task.samplesBufferLen / 2);
 		}

 		unsafe void run_GPU_task(FLACCLTask task)
@@ -1467,6 +1467,9 @@ namespace CUETools.Codecs.FLACCL
 				OCLMan.Defines =
 					"#define MAX_ORDER " + eparams.max_prediction_order.ToString() + "\n" +
 					"#define GROUP_SIZE " + groupSize.ToString() + "\n" +
+#if DEBUG
+					"#define DEBUG\n" +
+#endif
 					_settings.Defines + "\n";
 				// The BuildOptions string is passed directly to clBuild and can be used to do debug builds etc
 				OCLMan.BuildOptions = "";
@@ -2230,8 +2233,7 @@ namespace CUETools.Codecs.FLACCL
 			int riceParamsLen = sizeof(int) * (4 << 8) * channels * FLACCLWriter.maxFrames;
 			int lpcDataLen = sizeof(float) * 32 * 33 * lpc.MAX_LPC_WINDOWS * channelsCount * FLACCLWriter.maxFrames;

-			clSamplesBytes = openCLProgram.Context.CreateBuffer(MemFlags.READ_WRITE | MemFlags.ALLOC_HOST_PTR, (uint)samplesBufferLen / 2);
-			//openCLCQ.EnqueueMapBuffer(cudaSamplesBytes, true, MapFlags.WRITE, 0, samplesBufferLen / 2);
+			clSamplesBytes = openCLProgram.Context.CreateBuffer(MemFlags.READ_ONLY | MemFlags.ALLOC_HOST_PTR, (uint)samplesBufferLen / 2);
 			clSamples = openCLProgram.Context.CreateBuffer(MemFlags.READ_WRITE, samplesBufferLen);
 			clResidual = openCLProgram.Context.CreateBuffer(MemFlags.READ_WRITE | MemFlags.ALLOC_HOST_PTR, samplesBufferLen);
 			clLPCData = openCLProgram.Context.CreateBuffer(MemFlags.READ_WRITE, lpcDataLen);
@@ -2244,24 +2246,26 @@ namespace CUETools.Codecs.FLACCL
 			clResidualOutput = openCLProgram.Context.CreateBuffer(MemFlags.READ_WRITE, sizeof(int) * channelsCount * (lpc.MAX_LPC_WINDOWS * lpc.MAX_LPC_ORDER + 8) * 64 /*FLACCLWriter.maxResidualParts*/ * FLACCLWriter.maxFrames);
 			clWindowFunctions = openCLProgram.Context.CreateBuffer(MemFlags.READ_ONLY | MemFlags.ALLOC_HOST_PTR, sizeof(float) * FLACCLWriter.MAX_BLOCKSIZE /** 2*/ * lpc.MAX_LPC_WINDOWS);

-			clComputeAutocor = openCLProgram.CreateKernel("cudaComputeAutocor");
-			clStereoDecorr = openCLProgram.CreateKernel("cudaStereoDecorr");
-			//cudaChannelDecorr = openCLProgram.CreateKernel("cudaChannelDecorr");
-			clChannelDecorr2 = openCLProgram.CreateKernel("cudaChannelDecorr2");
-			clFindWastedBits = openCLProgram.CreateKernel("cudaFindWastedBits");
-			clComputeLPC = openCLProgram.CreateKernel("cudaComputeLPC");
-			clQuantizeLPC = openCLProgram.CreateKernel("cudaQuantizeLPC");
-			//cudaComputeLPCLattice = openCLProgram.CreateKernel("cudaComputeLPCLattice");
-			clEstimateResidual = openCLProgram.CreateKernel("cudaEstimateResidual");
-			clChooseBestMethod = openCLProgram.CreateKernel("cudaChooseBestMethod");
-			clCopyBestMethod = openCLProgram.CreateKernel("cudaCopyBestMethod");
-			clCopyBestMethodStereo = openCLProgram.CreateKernel("cudaCopyBestMethodStereo");
-			clEncodeResidual = openCLProgram.CreateKernel("cudaEncodeResidual");
-			clCalcPartition = openCLProgram.CreateKernel("cudaCalcPartition");
-			clCalcPartition16 = openCLProgram.CreateKernel("cudaCalcPartition16");
-			clSumPartition = openCLProgram.CreateKernel("cudaSumPartition");
-			clFindRiceParameter = openCLProgram.CreateKernel("cudaFindRiceParameter");
-			clFindPartitionOrder = openCLProgram.CreateKernel("cudaFindPartitionOrder");
+			//openCLCQ.EnqueueMapBuffer(clSamplesBytes, true, MapFlags.WRITE, 0, samplesBufferLen / 2);
+
+			clComputeAutocor = openCLProgram.CreateKernel("clComputeAutocor");
+			clStereoDecorr = openCLProgram.CreateKernel("clStereoDecorr");
+			//cudaChannelDecorr = openCLProgram.CreateKernel("clChannelDecorr");
+			clChannelDecorr2 = openCLProgram.CreateKernel("clChannelDecorr2");
+			clFindWastedBits = openCLProgram.CreateKernel("clFindWastedBits");
+			clComputeLPC = openCLProgram.CreateKernel("clComputeLPC");
+			clQuantizeLPC = openCLProgram.CreateKernel("clQuantizeLPC");
+			//cudaComputeLPCLattice = openCLProgram.CreateKernel("clComputeLPCLattice");
+			clEstimateResidual = openCLProgram.CreateKernel("clEstimateResidual");
+			clChooseBestMethod = openCLProgram.CreateKernel("clChooseBestMethod");
+			clCopyBestMethod = openCLProgram.CreateKernel("clCopyBestMethod");
+			clCopyBestMethodStereo = openCLProgram.CreateKernel("clCopyBestMethodStereo");
+			clEncodeResidual = openCLProgram.CreateKernel("clEncodeResidual");
+			clCalcPartition = openCLProgram.CreateKernel("clCalcPartition");
+			clCalcPartition16 = openCLProgram.CreateKernel("clCalcPartition16");
+			clSumPartition = openCLProgram.CreateKernel("clSumPartition");
+			clFindRiceParameter = openCLProgram.CreateKernel("clFindRiceParameter");
+			clFindPartitionOrder = openCLProgram.CreateKernel("clFindPartitionOrder");

 			samplesBuffer = new int[FLACCLWriter.MAX_BLOCKSIZE * channelsCount];
 			outputBuffer = new byte[max_frame_size * FLACCLWriter.maxFrames + 1];
@@ -2377,14 +2381,13 @@ namespace CUETools.Codecs.FLACCL
 				clSamples,
 				clWindowFunctions,
 				clResidualTasks,
-				nWindowFunctions - 1,
 				nResidualTasksPerChannel);

 			openCLCQ.EnqueueNDRangeKernel(
 				clComputeAutocor,
 				groupSize, 1,
-				eparams.max_prediction_order / 4 + 1,
-				nWindowFunctions * channelsCount * frameCount);
+				channelsCount * frameCount,
+				nWindowFunctions);

 			clComputeLPC.SetArgs(
 				clResidualTasks,
@@ -2491,7 +2494,7 @@ namespace CUETools.Codecs.FLACCL
 					openCLCQ.EnqueueNDRangeKernel(
 						clCalcPartition,
 						groupSize, 1,
-						1 << max_porder,
+						1 + ((1 << max_porder) - 1) / (groupSize / 16),
 						channels * frameCount);
 				}

@@ -2516,7 +2519,7 @@ namespace CUETools.Codecs.FLACCL
 				openCLCQ.EnqueueNDRangeKernel(
 					clFindRiceParameter,
 					groupSize, 1,
-					Math.Max(1, 8 * (2 << max_porder) / groupSize),
+					Math.Max(1, (2 << max_porder) / groupSize),
 					channels * frameCount);

 				//if (max_porder > 0) // need to run even if max_porder==0 just to calculate the final frame size
@@ -2531,18 +2534,18 @@ namespace CUETools.Codecs.FLACCL
 					groupSize,
 					channels * frameCount);

-				//openCLCQ.EnqueueReadBuffer(cudaBestRiceParams, false, 0, sizeof(int) * (1 << max_porder) * channels * frameCount, cudaBestRiceParams.HostPtr);
-				//openCLCQ.EnqueueReadBuffer(cudaResidual, false, 0, sizeof(int) * MAX_BLOCKSIZE * channels, cudaResidual.HostPtr);
-				openCLCQ.EnqueueMapBuffer(clBestRiceParams, false, MapFlags.READ, 0, sizeof(int) * (1 << max_porder) * channels * frameCount);
-				openCLCQ.EnqueueUnmapMemObject(clBestRiceParams, clBestRiceParams.HostPtr);
-				openCLCQ.EnqueueMapBuffer(clResidual, false, MapFlags.READ, 0, sizeof(int) * FLACCLWriter.MAX_BLOCKSIZE * channels);
-				openCLCQ.EnqueueUnmapMemObject(clResidual, clResidual.HostPtr);
+				openCLCQ.EnqueueReadBuffer(clBestRiceParams, false, 0, sizeof(int) * (1 << max_porder) * channels * frameCount, clBestRiceParams.HostPtr);
+				openCLCQ.EnqueueReadBuffer(clResidual, false, 0, sizeof(int) * FLACCLWriter.MAX_BLOCKSIZE * channels, clResidual.HostPtr);
+				//openCLCQ.EnqueueMapBuffer(clBestRiceParams, false, MapFlags.READ, 0, sizeof(int) * (1 << max_porder) * channels * frameCount);
+				//openCLCQ.EnqueueUnmapMemObject(clBestRiceParams, clBestRiceParams.HostPtr);
+				//openCLCQ.EnqueueMapBuffer(clResidual, false, MapFlags.READ, 0, sizeof(int) * FLACCLWriter.MAX_BLOCKSIZE * channels);
+				//openCLCQ.EnqueueUnmapMemObject(clResidual, clResidual.HostPtr);
 			}
-			//openCLCQ.EnqueueReadBuffer(cudaBestResidualTasks, false, 0, sizeof(FLACCLSubframeTask) * channels * frameCount, cudaBestResidualTasks.HostPtr);
-			openCLCQ.EnqueueMapBuffer(clBestResidualTasks, false, MapFlags.READ, 0, sizeof(FLACCLSubframeTask) * channels * frameCount);
-			openCLCQ.EnqueueUnmapMemObject(clBestResidualTasks, clBestResidualTasks.HostPtr);
+			openCLCQ.EnqueueReadBuffer(clBestResidualTasks, false, 0, sizeof(FLACCLSubframeTask) * channels * frameCount, clBestResidualTasks.HostPtr);
+			//openCLCQ.EnqueueMapBuffer(clBestResidualTasks, false, MapFlags.READ, 0, sizeof(FLACCLSubframeTask) * channels * frameCount);
+			//openCLCQ.EnqueueUnmapMemObject(clBestResidualTasks, clBestResidualTasks.HostPtr);

-			//openCLCQ.EnqueueMapBuffer(cudaSamplesBytes, false, MapFlags.WRITE, 0, samplesBufferLen / 2);
+			//openCLCQ.EnqueueMapBuffer(clSamplesBytes, false, MapFlags.WRITE, 0, samplesBufferLen / 2);
 		}
 	}
 }
--- a/CUETools.Codecs.FLACCL/flac.cl
+++ b/CUETools.Codecs.FLACCL/flac.cl
@@ -20,6 +20,12 @@
 #ifndef _FLACCL_KERNEL_H_
 #define _FLACCL_KERNEL_H_

+#ifdef DEBUG
+#pragma OPENCL EXTENSION cl_amd_printf : enable
+#endif
+
+#pragma OPENCL EXTENSION cl_khr_local_int32_base_atomics : enable
+
 //#pragma OPENCL EXTENSION cl_amd_fp64 : enable

 typedef enum
@@ -55,7 +61,7 @@ typedef struct
    int coefs[32]; // fixme: should be short?
 } FLACCLSubframeTask;

-__kernel void cudaStereoDecorr(
+__kernel void clStereoDecorr(
    __global int *samples,
    __global short2 *src,
    int offset
@@ -72,7 +78,7 @@ __kernel void cudaStereoDecorr(
    }
 }

-__kernel void cudaChannelDecorr2(
+__kernel void clChannelDecorr2(
    __global int *samples,
    __global short2 *src,
    int offset
@@ -87,7 +93,7 @@ __kernel void cudaChannelDecorr2(
    }
 }

-//__kernel void cudaChannelDecorr(
+//__kernel void clChannelDecorr(
 //    int *samples,
 //    short *src,
 //    int offset
@@ -102,7 +108,7 @@ __kernel void cudaChannelDecorr2(
 //#define __ffs(a) (33 - clz(~a & (a - 1)))

 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaFindWastedBits(
+void clFindWastedBits(
    __global FLACCLSubframeTask *tasks,
    __global int *samples,
    int tasksPerChannel
@@ -115,12 +121,13 @@ void cudaFindWastedBits(
    int tid = get_local_id(0);
    if (tid < sizeof(task) / sizeof(int))
 	((__local int*)&task)[tid] = ((__global int*)(&tasks[get_group_id(0) * tasksPerChannel].data))[tid];
+    
    barrier(CLK_LOCAL_MEM_FENCE);

    int w = 0, a = 0;
-    for (int pos = 0; pos < task.blocksize; pos += GROUP_SIZE)
+    for (int pos = tid; pos + tid < task.blocksize; pos += GROUP_SIZE)
    {
-	int smp = pos + tid < task.blocksize ? samples[task.samplesOffs + pos + tid] : 0;
+	int smp = samples[task.samplesOffs + pos];
 	w |= smp;
 	a |= smp ^ (smp >> 31);
    }
@@ -146,37 +153,39 @@ void cudaFindWastedBits(
 	tasks[get_group_id(0) * tasksPerChannel + tid].data.abits = a;
 }

+// get_num_groups(0) == number of tasks
+// get_num_groups(1) == number of windows
 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaComputeAutocor(
+void clComputeAutocor(
    __global float *output,
    __global const int *samples,
    __global const float *window,
    __global FLACCLSubframeTask *tasks,
-    const int windowCount, // windows (log2: 0,1)
    const int taskCount // tasks per block
 )
 {
    __local float data[GROUP_SIZE * 2];
-    __local float product[GROUP_SIZE];
+    __local float product[(MAX_ORDER / 4 + 1) * GROUP_SIZE];
    __local FLACCLSubframeData task;
    const int tid = get_local_id(0);
    // fetch task data
    if (tid < sizeof(task) / sizeof(int))
-	((__local int*)&task)[tid] = ((__global int*)(tasks + taskCount * (get_group_id(1) >> windowCount)))[tid];
+	((__local int*)&task)[tid] = ((__global int*)(tasks + taskCount * get_group_id(0)))[tid];
+
+    for (int ord4 = 0; ord4 < (MAX_ORDER / 4 + 1); ord4 ++)
+	product[ord4 * GROUP_SIZE + tid] = 0.0f;
    barrier(CLK_LOCAL_MEM_FENCE);
    
    int bs = task.blocksize;
-    int windowOffs = (get_group_id(1) & ((1 << windowCount)-1)) * bs;
+    int windowOffs = get_group_id(1) * bs;

    data[tid] = tid < bs ? samples[task.samplesOffs + tid] * window[windowOffs + tid] : 0.0f;

    int tid0 = tid % (GROUP_SIZE >> 2);
    int tid1 = tid / (GROUP_SIZE >> 2);
-    int lag0 = get_group_id(0) * 4;
    __local float4 * dptr = ((__local float4 *)&data[0]) + tid0;
-    __local float4 * dptr1 = ((__local float4 *)&data[lag0 + tid1]) + tid0;
+    __local float4 * dptr1 = ((__local float4 *)&data[tid1]) + tid0;
    
-    float prod = 0.0f;
    for (int pos = 0; pos < bs; pos += GROUP_SIZE)
    {
 	// fetch samples
@@ -184,32 +193,26 @@ void cudaComputeAutocor(
 	data[tid + GROUP_SIZE] = nextData;
 	barrier(CLK_LOCAL_MEM_FENCE);

-	prod += dot(*dptr, *dptr1);
+	for (int ord4 = 0; ord4 < (MAX_ORDER / 4 + 1); ord4 ++)
+	    product[ord4 * GROUP_SIZE + tid] += dot(dptr[0], dptr1[ord4]);

 	barrier(CLK_LOCAL_MEM_FENCE);

 	data[tid] = nextData;
    }
-    product[tid] = prod;
-    barrier(CLK_LOCAL_MEM_FENCE);
+    for (int ord4 = 0; ord4 < (MAX_ORDER / 4 + 1); ord4 ++)
 	for (int l = (GROUP_SIZE >> 3); l > 0; l >>= 1)
 	{
 	    if (tid0 < l)
-	    product[tid] = product[tid] + product[tid + l];
+		product[ord4 * GROUP_SIZE + tid] += product[ord4 * GROUP_SIZE + tid + l];
 	    barrier(CLK_LOCAL_MEM_FENCE);
 	}
-    if (tid < 4 && tid + lag0 <= MAX_ORDER)
-	output[get_group_id(1) * (MAX_ORDER + 1) + tid + lag0] = product[tid * (GROUP_SIZE >> 2)];
+    if (tid <= MAX_ORDER)
+	output[(get_group_id(0) * get_num_groups(1) + get_group_id(1)) * (MAX_ORDER + 1) + tid] = product[tid * (GROUP_SIZE >> 2)];
 }

-//#define DEBUGPRINT
-
-#ifdef DEBUGPRINT
-#pragma OPENCL EXTENSION cl_amd_printf : enable
-#endif
-
 __kernel __attribute__((reqd_work_group_size(32, 1, 1)))
-void cudaComputeLPC(
+void clComputeLPC(
    __global FLACCLSubframeTask *tasks,
    __global float *autoc,
    __global float *lpcs,
@@ -250,7 +253,7 @@ void cudaComputeLPC(
    shared.ldr[get_local_id(0)] = 0.0f;
    float error = shared.autoc[0];
    
-#ifdef DEBUGPRINT
+#ifdef DEBUGPRINT1
    int magic = shared.autoc[0] == 177286873088.0f;
    if (magic && get_local_id(0) <= MAX_ORDER)
        printf("autoc[%d] == %f\n", get_local_id(0), shared.autoc[get_local_id(0)]);
@@ -261,8 +264,8 @@ void cudaComputeLPC(
    {
 	// Schur recursion
 	float reff = -shared.gen1[0] / error;
-	error += shared.gen1[0] * reff; // Equivalent to error *= (1 - reff * reff);
-	//error *= (1 - reff * reff);
+	//error += shared.gen1[0] * reff; // Equivalent to error *= (1 - reff * reff);
+	error *= (1 - reff * reff);
 	float gen1;
 	if (get_local_id(0) < MAX_ORDER - 1 - order)
 	{
@@ -272,7 +275,7 @@ void cudaComputeLPC(
 	barrier(CLK_LOCAL_MEM_FENCE);
 	if (get_local_id(0) < MAX_ORDER - 1 - order)
 	    shared.gen1[get_local_id(0)] = gen1;
-#ifdef DEBUGPRINT
+#ifdef DEBUGPRINT1
 	if (magic && get_local_id(0) == 0)
 	    printf("order == %d, reff == %f, error = %f\n", order, reff, error);
 	if (magic && get_local_id(0) <= MAX_ORDER)
@@ -304,7 +307,7 @@ void cudaComputeLPC(
 }

 __kernel __attribute__((reqd_work_group_size(32, 1, 1)))
-void cudaQuantizeLPC(
+void clQuantizeLPC(
    __global FLACCLSubframeTask *tasks,
    __global float*lpcs,
    int taskCount, // tasks per block
@@ -449,8 +452,12 @@ void cudaQuantizeLPC(
    }
 }

+#ifndef PARTORDER
+#define PARTORDER 4
+#endif
+
 __kernel /*__attribute__(( vec_type_hint (int4)))*/ __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaEstimateResidual(
+void clEstimateResidual(
    __global int*output,
    __global int*samples,
    __global FLACCLSubframeTask *tasks
@@ -459,7 +466,7 @@ void cudaEstimateResidual(
    __local int data[GROUP_SIZE * 2];
    __local FLACCLSubframeTask task;
    __local int residual[GROUP_SIZE];
-    __local int len[GROUP_SIZE / 16];
+    __local int len[GROUP_SIZE >> PARTORDER];

    const int tid = get_local_id(0);
    if (tid < sizeof(task)/sizeof(int))
@@ -471,7 +478,7 @@ void cudaEstimateResidual(

    if (tid < 32 && tid >= ro)
 	task.coefs[tid] = 0;
-    if (tid < GROUP_SIZE / 16)
+    if (tid < (GROUP_SIZE >> PARTORDER))
 	len[tid] = 0;
    data[tid] = 0;

@@ -514,35 +521,49 @@ void cudaEstimateResidual(
 #endif
 	    ;
    	
-	int t = data[tid + GROUP_SIZE] - ((sum.x + sum.y + sum.z + sum.w) >> task.data.shift);
+	int t = nextData - ((sum.x + sum.y + sum.z + sum.w) >> task.data.shift);
 	// ensure we're within frame bounds
 	t = select(0, t, offs >= ro && offs < bs);
 	// overflow protection
 	t = clamp(t, -0x7fffff, 0x7fffff);
 	// convert to unsigned
 	residual[tid] = (t << 1) ^ (t >> 31);
-	barrier(CLK_GLOBAL_MEM_FENCE);
+	barrier(CLK_LOCAL_MEM_FENCE);
+	data[tid] = nextData;

 	// calculate rice partition bit length for every 16 samples
-	if (tid < GROUP_SIZE / 16)
+	if (tid < (GROUP_SIZE >> PARTORDER))
 	{
-	    __local int4 * chunk = ((__local int4 *)residual) + (tid << 2);
-	    int4 sum = chunk[0] + chunk[1] + chunk[2] + chunk[3];
-	    int res = sum.x + sum.y + sum.z + sum.w;
-	    int k = clamp(27 - clz(res), 0, 14); // 27 - clz(res) == clz(16) - clz(res) == log2(res / 16)
-#ifdef EXTRAMODE
-	    sum = (chunk[0] >> k) + (chunk[1] >> k) + (chunk[2] >> k) + (chunk[3] >> k);
-	    len[tid] += (k << 4) + sum.x + sum.y + sum.z + sum.w;
+	    //__local int4 * chunk = (__local int4 *)&residual[tid << PARTORDER];
+	    __local int4 * chunk = ((__local int4 *)residual) + (tid << (PARTORDER - 2));
+#if PARTORDER == 3
+	    int4 sum = chunk[0] + chunk[1];
+#elif PARTORDER == 4
+	    int4 sum = chunk[0] + chunk[1] + chunk[2] + chunk[3]; // [0 .. (1 << (PARTORDER - 2)) - 1]
+#elif PARTORDER == 5
+	    int4 sum = chunk[0] + chunk[1] + chunk[2] + chunk[3] + chunk[4] + chunk[5] + chunk[6] + chunk[7];
 #else
-	    len[tid] += (k << 4) + (res >> k);
+#error Invalid PARTORDER
+#endif
+	    int res = sum.x + sum.y + sum.z + sum.w;
+	    int k = clamp(clz(1 << PARTORDER) - clz(res), 0, 14); // 27 - clz(res) == clz(16) - clz(res) == log2(res / 16)
+#ifdef EXTRAMODE
+#if PARTORDER == 3
+	    sum = (chunk[0] >> k) + (chunk[1] >> k);
+#elif PARTORDER == 4
+	    sum = (chunk[0] >> k) + (chunk[1] >> k) + (chunk[2] >> k) + (chunk[3] >> k);
+#else
+#error Invalid PARTORDER
+#endif
+	    len[tid] += (k << PARTORDER) + sum.x + sum.y + sum.z + sum.w;
+#else
+	    len[tid] += (k << PARTORDER) + (res >> k);
 #endif
 	}
-
-	data[tid] = nextData;
    }

    barrier(CLK_LOCAL_MEM_FENCE);
-    for (int l = GROUP_SIZE / 32; l > 0; l >>= 1)
+    for (int l = GROUP_SIZE >> (PARTORDER + 1); l > 0; l >>= 1)
    {
 	if (tid < l)
 	    len[tid] += len[tid + l];
@@ -553,7 +574,7 @@ void cudaEstimateResidual(
 }

 __kernel __attribute__((reqd_work_group_size(32, 1, 1)))
-void cudaChooseBestMethod(
+void clChooseBestMethod(
    __global FLACCLSubframeTask *tasks,
    __global int *residual,
    int taskCount
@@ -621,7 +642,7 @@ void cudaChooseBestMethod(
 }

 __kernel __attribute__((reqd_work_group_size(64, 1, 1)))
-void cudaCopyBestMethod(
+void clCopyBestMethod(
    __global FLACCLSubframeTask *tasks_out,
    __global FLACCLSubframeTask *tasks,
    int count
@@ -636,7 +657,7 @@ void cudaCopyBestMethod(
 }

 __kernel __attribute__((reqd_work_group_size(64, 1, 1)))
-void cudaCopyBestMethodStereo(
+void clCopyBestMethodStereo(
    __global FLACCLSubframeTask *tasks_out,
    __global FLACCLSubframeTask *tasks,
    int count
@@ -690,7 +711,7 @@ void cudaCopyBestMethodStereo(

 // get_group_id(0) == task index
 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaEncodeResidual(
+void clEncodeResidual(
    __global int *output,
    __global int *samples,
    __global FLACCLSubframeTask *tasks
@@ -756,10 +777,10 @@ void cudaEncodeResidual(
    }
 }

-// get_group_id(0) == partition index
+// get_group_id(0) == partition index / (GROUP_SIZE / 16)
 // get_group_id(1) == task index
 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaCalcPartition(
+void clCalcPartition(
    __global int *partition_lengths,
    __global int *residual,
    __global FLACCLSubframeTask *tasks,
@@ -767,51 +788,53 @@ void cudaCalcPartition(
    int psize // == task.blocksize >> max_porder?
    )
 {
-    __local int data[GROUP_SIZE];
-    __local int length[GROUP_SIZE / 16][16];
+    __local int pl[(GROUP_SIZE / 16)][15];
    __local FLACCLSubframeData task;

    const int tid = get_local_id(0);
    if (tid < sizeof(task) / sizeof(int))
 	((__local int*)&task)[tid] = ((__global int*)(&tasks[get_group_id(1)]))[tid];
-    barrier(CLK_LOCAL_MEM_FENCE);
-
-    int k = tid % 16;
-    int x = tid / 16;
-
-    int sum = 0;
-    for (int pos0 = 0; pos0 < psize; pos0 += GROUP_SIZE)
+    if (tid < (GROUP_SIZE / 16))
    {
-	int offs = get_group_id(0) * psize + pos0 + tid;
-	// fetch residual
-	int s = (offs >= task.residualOrder && pos0 + tid < psize) ? residual[task.residualOffs + offs] : 0;
-	// convert to unsigned
-	data[tid] = min(0x7fffff, (s << 1) ^ (s >> 31));
-	barrier(CLK_LOCAL_MEM_FENCE);
-
-	// calc number of unary bits for each residual sample with each rice paramater
-	for (int pos = 0; pos < psize && pos < GROUP_SIZE; pos += GROUP_SIZE / 16)
-	    sum += data[pos + x] >> k;
-	barrier(CLK_LOCAL_MEM_FENCE);
+	for (int k = 0; k <= 14; k++)
+	    pl[tid][k] = 0;
    }
-    
-    length[x][k] = min(0x7fffff, sum);
    barrier(CLK_LOCAL_MEM_FENCE);

-    if (x == 0)
+    int start = get_group_id(0) * psize * (GROUP_SIZE / 16);
+    int end = min(start + psize * (GROUP_SIZE / 16), task.blocksize);
+    for (int offs = start + tid; offs < end; offs += GROUP_SIZE)
+    {
+	// fetch residual
+	int s = (offs >= task.residualOrder && offs < end) ? residual[task.residualOffs + offs] : 0;
+	// convert to unsigned
+	s = clamp(s, -0x7fffff, 0x7fffff);
+	s = (s << 1) ^ (s >> 31);
+	// calc number of unary bits for each residual sample with each rice paramater
+	int part = (offs - start) / psize;
+	for (int k = 0; k <= 14; k++)
+	    atom_add(&pl[part][k], s >> k);
+	    //pl[part][k] += s >> k;
+    }   
+    barrier(CLK_LOCAL_MEM_FENCE);
+
+    int part = get_group_id(0) * (GROUP_SIZE / 16) + tid;
+    if (tid < (GROUP_SIZE / 16) && part < (1 << max_porder))
+    {
+	for (int k = 0; k <= 14; k++)
 	{
-	for (int i = 1; i < GROUP_SIZE / 16; i++)
-	    length[0][k] += length[i][k];
 	    // output length
 	    const int pos = (15 << (max_porder + 1)) * get_group_id(1) + (k << (max_porder + 1));
-	if (k <= 14)
-	    partition_lengths[pos + get_group_id(0)] = min(0x7fffff,length[0][k]) + (psize - task.residualOrder * (get_group_id(0) == 0)) * (k + 1);
+	    partition_lengths[pos + part] = min(0x7fffff, pl[tid][k]) + select(psize, psize - task.residualOrder, part == 0) * (k + 1);
+	 //   if (get_group_id(1) == 0)
+		//printf("pl[%d][%d] == %d\n", k, part, min(0x7fffff, pl[k][tid]) + (psize - task.residualOrder * (part == 0)) * (k + 1));
+	}
    }
 }

 // get_group_id(0) == task index
 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaCalcPartition16(
+void clCalcPartition16(
    __global int *partition_lengths,
    __global int *residual,
    __global int *samples,
@@ -887,6 +910,9 @@ void cudaCalcPartition16(
 	s = clamp(s, -0x7fffff, 0x7fffff);
 	// convert to unsigned
 	res[tid] = (s << 1) ^ (s >> 31);
+
+	// for (int k = 0; k < 15; k++) atom_add(&pl[x][k], s >> k);
+
 	barrier(CLK_LOCAL_MEM_FENCE);
 	data[tid] = nextData;

@@ -906,22 +932,22 @@ void cudaCalcPartition16(
 // get_group_id(0) == k
 // get_group_id(1) == task index
 __kernel __attribute__((reqd_work_group_size(128, 1, 1)))
-void cudaSumPartition(
+void clSumPartition(
    __global int* partition_lengths,
    int max_porder
    )
 {
-    __local int data[512]; // max_porder <= 8, data length <= 1 << 9.
+    __local int data[256]; // max_porder <= 8, data length <= 1 << 9.
    const int pos = (15 << (max_porder + 1)) * get_group_id(1) + (get_group_id(0) << (max_porder + 1));

    // fetch partition lengths
-    data[get_local_id(0)] = get_local_id(0) < (1 << max_porder) ? partition_lengths[pos + get_local_id(0)] : 0;
-    data[get_local_size(0) + get_local_id(0)] = get_local_size(0) + get_local_id(0) < (1 << max_porder) ? partition_lengths[pos + get_local_size(0) + get_local_id(0)] : 0;
+    int2 pl = get_local_id(0) * 2 < (1 << max_porder) ? *(__global int2*)&partition_lengths[pos + get_local_id(0) * 2] : 0;
+    data[get_local_id(0)] = pl.x + pl.y;
    barrier(CLK_LOCAL_MEM_FENCE);

    int in_pos = (get_local_id(0) << 1);
-    int out_pos = (1 << max_porder) + get_local_id(0);
-    for (int bs = 1 << (max_porder - 1); bs > 0; bs >>= 1)
+    int out_pos = (1 << (max_porder - 1)) + get_local_id(0);
+    for (int bs = 1 << (max_porder - 2); bs > 0; bs >>= 1)
    {
 	if (get_local_id(0) < bs) data[out_pos] = data[in_pos] + data[in_pos + 1];
 	in_pos += bs << 1;
@@ -929,131 +955,96 @@ void cudaSumPartition(
 	barrier(CLK_LOCAL_MEM_FENCE);
    }
    if (get_local_id(0) < (1 << max_porder))
-	partition_lengths[pos + (1 << max_porder) + get_local_id(0)] = data[(1 << max_porder) + get_local_id(0)];
+	partition_lengths[pos + (1 << max_porder) + get_local_id(0)] = data[get_local_id(0)];
    if (get_local_size(0) + get_local_id(0) < (1 << max_porder))
-	partition_lengths[pos + (1 << max_porder) + get_local_size(0) + get_local_id(0)] = data[(1 << max_porder) + get_local_size(0) + get_local_id(0)];
+	partition_lengths[pos + (1 << max_porder) + get_local_size(0) + get_local_id(0)] = data[get_local_size(0) + get_local_id(0)];
 }

 // Finds optimal rice parameter for several partitions at a time.
-// get_group_id(0) == chunk index (chunk size is GROUP_SIZE / 8, so total task size is 8 * (2 << max_porder))
+// get_group_id(0) == chunk index (chunk size is GROUP_SIZE, total task size is (2 << max_porder))
 // get_group_id(1) == task index
 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaFindRiceParameter(
+void clFindRiceParameter(
    __global int* rice_parameters,
    __global int* partition_lengths,
    int max_porder
    )
 {
-    __local struct {
-	volatile int length[GROUP_SIZE];
-	volatile int index[GROUP_SIZE];
-    } shared;
    const int tid = get_local_id(0);
-    const int ws = GROUP_SIZE / 8;
-    const int parts = min(ws, 2 << max_porder);
-    const int p = tid % ws;
-    const int k = tid / ws; // 0..7
-    const int pos = (15 << (max_porder + 1)) * get_group_id(1) + (k << (max_porder + 1));
+    const int parts = min(GROUP_SIZE, 2 << max_porder);
+    const int pos = (15 << (max_porder + 1)) * get_group_id(1) + get_group_id(0) * GROUP_SIZE + tid;

-    // read length for 32 partitions
-    int l1 = (p < parts) ? partition_lengths[pos + get_group_id(0) * ws + p] : 0xffffff;
-    int l2 = (k + 8 <= 14 && p < parts) ? partition_lengths[pos + (8 << (max_porder + 1)) + get_group_id(0) * ws + p] : 0xffffff;
-    // find best rice parameter
-    shared.index[tid] = k + ((l2 < l1) << 3);
-    shared.length[tid] = l1 = min(l1, l2);
-    barrier(CLK_LOCAL_MEM_FENCE);
-//#pragma unroll 3
-    for (int lsh = GROUP_SIZE / 2; lsh >= ws; lsh >>= 1)
-    {
-	if (tid < lsh)
-	{
-	    l2 = shared.length[tid + lsh];
-	    shared.index[tid] = shared.index[tid + (l2 < l1) * lsh];
-	    shared.length[tid] = l1 = min(l1, l2);
-	}    
-	barrier(CLK_LOCAL_MEM_FENCE);
-    }
    if (tid < parts)
    {
+	int best_l = partition_lengths[pos];
+	int best_k = 0;
+	for (int k = 1; k <= 14; k++)
+	{
+	    int l = partition_lengths[pos + (k << (max_porder + 1))];
+	    best_k = select(best_k, k, l < best_l);
+	    best_l = min(best_l, l);
+	}
+
 	// output rice parameter
-	rice_parameters[(get_group_id(1) << (max_porder + 2)) + get_group_id(0) * parts + tid] = shared.index[tid];
+	rice_parameters[(get_group_id(1) << (max_porder + 2)) + get_group_id(0) * GROUP_SIZE + tid] = best_k;
 	// output length
-	rice_parameters[(get_group_id(1) << (max_porder + 2)) + (1 << (max_porder + 1)) + get_group_id(0) * parts + tid] = shared.length[tid];
+	rice_parameters[(get_group_id(1) << (max_porder + 2)) + (1 << (max_porder + 1)) + get_group_id(0) * GROUP_SIZE + tid] = best_l;
    }
 }

 // get_group_id(0) == task index
 __kernel __attribute__((reqd_work_group_size(GROUP_SIZE, 1, 1)))
-void cudaFindPartitionOrder(
+void clFindPartitionOrder(
    __global int* best_rice_parameters,
    __global FLACCLSubframeTask *tasks,
    __global int* rice_parameters,
    int max_porder
    )
 {
-    __local struct {
-	int length[32];
-	int index[32];
-    } shared;
-    __local int partlen[GROUP_SIZE];
+    __local int partlen[9];
    __local FLACCLSubframeData task;

    const int pos = (get_group_id(0) << (max_porder + 2)) + (2 << max_porder);
    if (get_local_id(0) < sizeof(task) / sizeof(int))
 	((__local int*)&task)[get_local_id(0)] = ((__global int*)(&tasks[get_group_id(0)]))[get_local_id(0)];
+    if (get_local_id(0) < 9)
+	partlen[get_local_id(0)] = 0;
+    barrier(CLK_LOCAL_MEM_FENCE);
+
    // fetch partition lengths
+    for (int offs = 0; offs < (2 << max_porder); offs += GROUP_SIZE)
+    {
+	if (offs + get_local_id(0) < (2 << max_porder) - 1)
+	{
+	    int len = rice_parameters[pos + offs + get_local_id(0)];
+	    int porder = 31 - clz((2 << max_porder) - 1 - offs - get_local_id(0));
+	    atom_add(&partlen[porder], len);
+	}
+    }
    barrier(CLK_LOCAL_MEM_FENCE);

-    for (int porder = max_porder; porder >= 0; porder--)
+    int best_length = partlen[0] + 4;
+    int best_porder = 0;
+    for (int porder = 1; porder <= max_porder; porder++)
    {
-	int len = 0;
-	for (int offs = 0; offs < (1 << porder); offs += GROUP_SIZE)
-	    len += offs + get_local_id(0) < (1 << porder) ? rice_parameters[pos + (2 << max_porder) - (2 << porder) + offs + get_local_id(0)] : 0;
-	partlen[get_local_id(0)] = len;
-	barrier(CLK_LOCAL_MEM_FENCE);
-	for (int l = min(GROUP_SIZE, 1 << porder) / 2; l > 0; l >>= 1)
-	{
-	    if (get_local_id(0) < l)
-		partlen[get_local_id(0)] += partlen[get_local_id(0) + l];
-	    barrier(CLK_LOCAL_MEM_FENCE);
-	}
-	if (get_local_id(0) == 0)
-	    shared.length[porder] = partlen[0] + (4 << porder);
-	barrier(CLK_LOCAL_MEM_FENCE);
+	int length = (4 << porder) + partlen[porder];
+	best_porder = select(best_porder, porder, length < best_length);
+	best_length = min(best_length, length);
    }

-    if (get_local_id(0) < 32 && get_local_id(0) > max_porder)
-	shared.length[get_local_id(0)] = 0xfffffff;
-    if (get_local_id(0) < 32)
-	shared.index[get_local_id(0)] = get_local_id(0);
-    barrier(CLK_LOCAL_MEM_FENCE);
-    //atom_min(shared.index[get_local_id(0)],);
-    int l1 = get_local_id(0) <= max_porder ? shared.length[get_local_id(0)] : 0xfffffff;
-    for (int l = 8; l > 0; l >>= 1)
-    {
-	if (get_local_id(0) < l)
-	{
-	    int l2 = shared.length[get_local_id(0) + l];
-	    shared.index[get_local_id(0)] = shared.index[get_local_id(0) + select(0, l, l2 < l1)];
-	    shared.length[get_local_id(0)] = l1 = min(l1, l2);
-	}
-	barrier(CLK_LOCAL_MEM_FENCE);
-    }
-    if (get_local_id(0) == 0)
-	tasks[get_group_id(0)].data.porder = shared.index[0];
    if (get_local_id(0) == 0)
    {
+	tasks[get_group_id(0)].data.porder = best_porder;
 	int obits = task.obits - task.wbits;
 	tasks[get_group_id(0)].data.size =
-	    task.type == Fixed ? task.residualOrder * obits + 6 + l1 :
-	    task.type == LPC ? task.residualOrder * obits + 6 + l1 + 4 + 5 + task.residualOrder * task.cbits :
+	    task.type == Fixed ? task.residualOrder * obits + 6 + best_length :
+	    task.type == LPC ? task.residualOrder * obits + 6 + best_length + 4 + 5 + task.residualOrder * task.cbits :
 	    task.type == Constant ? obits : obits * task.blocksize;
    }
    barrier(CLK_LOCAL_MEM_FENCE);
-    int porder = shared.index[0];
-    for (int offs = 0; offs < (1 << porder); offs += GROUP_SIZE)
-	if (offs + get_local_id(0) < (1 << porder))
-	    best_rice_parameters[(get_group_id(0) << max_porder) + offs + get_local_id(0)] = rice_parameters[pos - (2 << porder) + offs + get_local_id(0)];
+    for (int offs = 0; offs < (1 << best_porder); offs += GROUP_SIZE)
+	if (offs + get_local_id(0) < (1 << best_porder))
+	    best_rice_parameters[(get_group_id(0) << max_porder) + offs + get_local_id(0)] = rice_parameters[pos - (2 << best_porder) + offs + get_local_id(0)];
    // FIXME: should be bytes?
 }
 #endif