There are no such tasks. If you had a target01 or target02 before which ran say 6 hours, it will always be in that ballpark. Sometimes the time_stats or DCF get screwed up, and then you get ludicrous hours, high or low. If you job started with 100 hours and after 1 hour dropped to 93 hours (WALLCLOCK), then nothing to worry... just let it run and BOINC will slowly normalize. If the task is really getting 100% CPU time (which is different from wallclock aka elapsed), then something is wrong. Do a boot and observe if it speeds up.

What was the client version again?

OK, thanks for confirming that indeed your client (6.10.58) is/was just confused at the time of receipt of this job. Just let it run and it will finish in "Normal" CPU time.

target01/target02 are part of the WU name and indicate

A. The investigated target configuration.
B. An Anticipated regular run time. If you've seen 3 WU's of target02, then you'll know that any after will run the same time, same for the shorter target01 and the longer target03 and so on.

DCF - Duration Correction Factor. Sorry when I forget to expand on an acronym then it can be found in the Start Here FAQ's. Bad habit (US acquired), so I started collecting them in a list: http://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,11187. From there links go to pages that explain it further.