receive are dependent on your relative contribution to the BOINC network.
In 4.20 and later, the CPID chosen is the one from the project that has the earliest join date. In fact the CPID (which is used to identify researchers) is an md-5 hash of. in 4.19 and earlier, the CPID chosen is the numerically larger. What the new CPID changes to depends on the version of the BOINC client. As per my OP, 'I asked him if he had copied any files over to the other host and he said he hadn't.', also, 'He reinstalled BOINC, yet still got the same CPID.' (He had made sure to delete the data directory as well.) I'd also like to point out that the projects themselves are still reporting the same CPID for both hosts to this day.
(Only Crash Collection does not as far as I know).ĥ) The projects must export the stats XML.Ħ) The stats site must import the stats XML. This means that if you have 2 hosts and each is not talking to the same 'oldest', to include inactive, the 2 hosts can converge to different CPIDs.
However if both crunch LHC as well, then the CPIDs will eventually line up.ģ) There must be a communication of the CPID from a host that has the correct CPID to every project server.Ĥ) The project servers must have recent enough code. The CPID a host converges to is the 'oldest' generated at a project, to which a client is talking. Example: Machine A crunches and Machine B Crunches Einstein, these are two distinct pools, and the two CPIDs will not ever line up. To get the CPIDs for all of the projects to line up, all of the following must happen / be true.ġ) The email addresses must be identical including case as they are hashed with the internal CPID to generate the external CPID without a change of case.Ģ) There must be a path from project server to host to project server that covers all projects.
0 kommentar(er)
