Hey Guys, Wanted to understand if parallel query processing is possible in NebulaGraph? I.E. fetching & processing edges and nodes is possible in parallel or not?

Hi @Ankit Kumar, Fetching edges and nodes is performed on multiple storageds and thus can be considered as parallel. However, processing edges and nodes is currently processed by a single graphd, which can only achieve thread-level parallelism and cannot achieve parallelism similar to MPP.

Hi @Wenting, Correct me if I am wrong, what I understand is that thread level parallelism is enabled by default for fetching & processing of edges/nodes. If we have multiple CPUs on single instance then those CPUs should be utilised during query processing?

Hi @Ankit Kumar, yes it's correct. Please check this file https://github.com/vesoft-inc/nebula/blob/master/conf/nebula-graphd.conf.default#L127-L130