determining, for a conversion between a current video block of a video picture and a bitstream of the video picture, a prediction mode being applied to the current video block;

determining a block vector of the current video block based on the prediction mode;

maintaining a virtual buffer comprising reference samples derived from blocks of sample values of the video picture;

deriving, in the prediction mode, prediction samples of the current video block based on the block vector and the virtual buffer; and

performing the conversion based on the prediction samples,

wherein a width of the virtual buffer is one or multiple times of a width of a virtual unit, and the width of the virtual unit is min (ctbSizeY, 64), where ctbSizeY denotes a size of a luma CTB comprising the current video block, and wherein a height of the virtual buffer is one or multiple times of a height of the virtual unit, and the height of the virtual unit is min (ctbSizeY, 64),

wherein locations of one or more unavailable samples in the virtual buffer for deriving the prediction samples are determined based on the width of the virtual unit and the height of the virtual unit, and

wherein a maximum number of available reference samples in the virtual buffer is less than a size of the virtual buffer.