A method of video processing is described. The method includes performing a conversion between a current video block of a video and a bitstream representation of the video, wherein, during the conversion, a prediction of the current video block is determined as a weighted sum of two or more predictions of the current video block, and a motion vector storing process for the current video block is determined according to a rule, wherein the current video block uses a partitioning prediction mode that is a first mode which is based on multiple first partitioning schemes or a second mode which is based on multiple second partitioning schemes, and wherein the rule specifies that a process used for determining which motion vector and how many motion vectors are stored is same for the first mode or the second mode.