How to use graph layered parallel computing ？And whether there are corresponding statements in tigergraph that can be calculated？
TigerGraph applies a very natural model for parallel computation. In graphs, we start from one set of nodes of interest and then “traverse” across the edges that connect them to another set of nodes. There can be filtering and computation with the data on theses nodes and edges. These sets can be very large, so TigerGraph uses parallel processing of these sets to greatly accelerate graph traversal.
In a GSQL SELECT statement, these sets are defined by the FROM … WHERE clauses. The computational actions to be taken are defined in the ACCUM and POST-ACCUM clauses, which automatically employ parallel (more specifically, multi-agent) processing.