用OpenCL实现HEVC中ME模块的测试数据分析

 

 

使用opencl来实现编码算法中运动搜索模块！

下面测试数据时在GTX570上的测试结果：

LCU为32x32, 100帧720P, CPU上纯C算法使用搜索时间是67s, GPU上是0.915s

LCU为16x16, CPU 是76.8s,   GPU上是1.6s

LCU为8x8, CPU 是82.5s,   GPU上是4.2s

 

 

同样的程序， CPU改为SSE实现， GPU做一个小的改动， 使用缩减算法！ 结果如下：

 

 

 

从上面数据可以看出， sse 比C语言快5倍左右， 新的GPU kenel快了20%左右， 其中LCU为8x8的快了好几倍！

 

综合看来

OPENCL实现 比C语言实现接近100倍的级别， 比SSE快了接近20倍左右！

另外提一句， 如果OPENCL不适用__local 内存的话， 会慢一半！

下面贴出部分代码供参考：

 

#define SearchRange 16
#define Edge_SIZE_T 48
//32x32 version of kernel
__kernel void opencl_me_32x32(const __global short* p_ref, __global short* p_cur, __global int* outputBuf, __local int* local_refBuf, __local int* local_curBuf, __local int* mv_cost)
{
	int searchrange = SearchRange;
	int edeg = Edge_SIZE_T;
	int width = get_global_size(0);
	int height = get_global_size(1);
	int block_w = get_local_size(0);
	int block_h = get_local_size(1);
	int local_x = get_local_id(0);
	int local_y = get_local_id(1);
	int lcu_x = get_group_id(0);
	int lcu_y = get_group_id(1);
	int stride = width + 2 * edeg;
	int lcu_adr_offset = edeg * stride + edeg;
	int local_refBuf_stride = block_w + 2 * searchrange;
			
	//LCU blcok adr
	lcu_adr_offset += lcu_y * stride * block_h + lcu_x * block_w;
	int ref_lcu_adr_offset = lcu_adr_offset - searchrange - searchrange * stride;
	//thread adr
	int global_thread_adr_offset = local_y * stride + local_x;
	int thread_adr_offset  = local_y * local_refBuf_stride + local_x;

	local_curBuf[local_y * block_w + local_x]										= p_cur[lcu_adr_offset + global_thread_adr_offset];

	local_refBuf[thread_adr_offset]												= p_ref[ref_lcu_adr_offset + global_thread_adr_offset ];
	local_refBuf[thread_adr_offset + block_w]									= p_ref[ref_lcu_adr_offset + global_thread_adr_offset + block_w];
	local_refBuf[thread_adr_offset + local_refBuf_stride * block_h]				= p_ref[ref_lcu_adr_offset + global_thread_adr_offset + stride * block_h];
	local_refBuf[thread_adr_offset + local_refBuf_stride * block_h + block_w]	= p_ref[ref_lcu_adr_offset + global_thread_adr_offset + stride * block_h + block_w];

 barrier(CLK_LOCAL_MEM_FENCE);
  {
   int i;
  int uiSum = 0;
  for( int i = 0; i < block_h; i++ )
  {

计算sad

}

{

  比较最小SAD 保存bestcost

}

 if((local_y ==0) && (local_x == 0))
 {
  int best_sad  = mv_cost[local_y*2*SearchRange*3 + local_x*3 + 2];
  int best_mvx  = mv_cost[local_y*2*SearchRange*3 + local_x*3 + 0];
  int best_mvy  = mv_cost[local_y*2*SearchRange*3 + local_x*3 + 1];
  outputBuf[(lcu_y * get_num_groups(0) + lcu_x)*3 + 0] = best_mvx;
  outputBuf[(lcu_y * get_num_groups(0) + lcu_x)*3 + 1] = best_mvy;
  outputBuf[(lcu_y * get_num_groups(0) + lcu_x)*3 + 2] = best_sad; 
  //printf("\nxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx = %d, y = %d, sad = %d",best_mvx, best_mvy, best_sad);
 }