MaggieQi
Results
13
comments of
MaggieQi
You can load different indexes into different servers. By configuring an aggregator (see Get Start page), clients can send requests to the aggregator and get the aggregated results from different...
@Yukikaze-CZR and @tinkerlin, you can look at the “Dump” and “Load” implementations in Wrappers/src/CLRCoreInterface.cpp.
We are exploring GPU speedup for the index build. For index serving, it is fast enough (several milliseconds).