ExpandPixel2X2For8BitGrad_Kernel_2<<<BlockNum, ThreadNum>>>(Width, Height, 2, dev_ImageGrad, dev_ImageSrc, dev_ImageDes);
cudaFree(dev_ImageSrc);
cudaFree(dev_ImageGrad);
byte * ImagePolatation;
cudaMalloc(&ImagePolatation, LineByteOut * Height * 2);
Filter3X3_Kernal<<<BlockNum, ThreadNum>>>(dev_ImageDes, Width * 2, Height * 2, ImagePolatation);
我释放cudaFree(dev_ImageSrc);cudaFree(dev_ImageGrad);之后程序运行时不正确的,如何把cudaFree(dev_ImageSrc);cudaFree(dev_ImageGrad);放在Filter3X3_Kernal<<<BlockNum, ThreadNum>>>(dev_ImageDes, Width * 2, Height * 2, ImagePolatation);后面运行才正确,我想问一下
(1)这个是不是属于核函数之间的同步问题?
(2)若想使用完指针之后立即释放,应该如何操作?