Weird host

If the results are valid, the project is happy to accept them.

As to why the crunch times seem fictitious, that's because they probably are. This doesn't affect the way this project works. If you drill down to the host details, it's a quite decent i5 running Linux. The recommended BOINC version for Linux is 7.2.42 and the most recent development version is 7.4.22. That host is running a version which identifies as 7.5.0. In other words, it's a version that may have been compiled outside the official BOINC system.

Since the source code is freely available, there's nothing to stop people modifying it for their own purposes and giving it whatever version number they want. BOINC doesn't do any science itself - it just provides the framework for the project apps to do their thing. Projects have systems in place to ensure that what is returned as a result is legitimate. I'm not at all suggesting there's anything suspicious about people who compile their own versions. There's lots of legitimate reasons.

The most likely reason for the very low elapsed time is that the task was stopped at some point close to completion and when it was restarted, the time counting started from scratch but the previous progress was able to be reloaded from the saved checkpoint. I've seen exactly this sort of behaviour with official BOINC versions, particularly after a crash of some sort. The 0.00 CPU time is probably an OS artifact. I used to see this a lot on earlier 32bit versions of the OS I run. I don't see it now with later 64bit versions.