Zhejiang Tmall Technology Co., Ltd.,

Xiaojian Luo

Jingbo Xu

Wenfei Fan

Wenyuan Yu

A graph data storage method for non-uniform memory access architecture (NUMA) processing system is provided. The processing system includes at least one computing device, each computing device corresponding to multiple memories, and each memory corresponding to multiple processors. The method includes: performing three-level partitioning on graph data to obtain multiple third-level partitions based on a communication mode among computing device(s), memories, and processors; and separately storing graph data of the multiple third-level partitions in NUMA nodes corresponding to the processors. A graph data storage system and an electronic device are further provided.

Graph data storage method, system and electronic device

A method for scheduling a computational operation based on graph data, which includes: dividing graph data into multiple partitions; scheduling the partitions to be allocated to multiple processors; successively assigning the partitions to corresponding threads of the processors for computation according to a respective degree of criticality of each partition; determining whether an idle processor exists, a processor being the idle processor when the processor includes one or more idle threads; communicating the idle processor with other processor(s) to find partition(s) to be computed and corresponding to the other processor(s); selecting and determining a partition to be computed; and migrating and binding the determined partition to be computed to thread(s) of the idle processor for processing. The present disclosure further relates to a system, a computer readable medium, and a device for scheduling a computational operation based on graph data.

Growing community of inventors

Beijing, China

Xiaojian Luo