I seem to recall that there were threads about problems with hibernating way, way back in the early days here on the WCG forums. My own solution was to simply run my machines 24x7. With laptops, I simply set them to do nothing in response to the lid being closed and to keep crunching when running on batteries. That would let them keep crunching during the commute to/from work. From a practical standpoint, shutting down a machine after all of the running wu's have checkpointed is the key. Typically, projects checkpoint frequently enough that there is not a timing issue. With CEP, the checkpoints are much less frequent, as you've seen. With a CEP wu running, you may have to wait an hour or two for it to checkpoint before shutting down. If you are running more than one CEP wu at the same time, the timing can get complicated quickly. Your approach is a good solution for that. Logically, it should not cause a problem with the validity of the completed wu's. As long at the memory information that gets written to disk on hibernating is reloaded on restart, you won't lose any progress. If your approach is going to cause validity problems, you probably would have already seen that in your results status. One thing I would recommend though is that any changes such as OS or BOINC updates be avoided at start up and instead be done when you can manually stop BOINC. I also recommend defragging on at least a weekly basis as fragmentation will be much worse than typical due to all the wu's that get downloaded when crunching on a machine.

I run at home on Win7 64bit and have so far not had any problems with suspending crunching before hibernating my machine on CEP units. I do only run one CEP at a time.
Admittedly I don't hibernate every day but have done so on quite a few occasions.

It's a while since I hibernated and restarted a machine that was running BOINC, but I have done it successfully in the past, mostly on a machine running Windows XP-32.
However, I did get a BSOD on startup a couple of times, so after that I always suspended BOINC first. If you do this, make sure that LAIM is set before suspending and don't forget to resume BOINC after you start up again!

The biggest problem I've had with this is that Windows will only hibernate on about 1 machine in 5, and there are no diagnostic messages that I can find that explain the failure to hibernate. MS say that all devices and drivers on/in the machine must support hibernation. I suspect that this may be a frequent cause, but the lack of diagnostics is extremely annoying. All part of the policy "No error messages appeared so there can't possibly be an error and we don't have to admit it or fix it".

And hibernating doesn't help when you've got CEP2 WUs running and you want to reboot the O/S, eg after installing or updating software.

I stopped doing WU from Clean Energy project because all time I go to hibernating or switch of my pc I need to do all work again even I reached already 95%. I'm using Win7 and normaly have 9 WU active in parallel. I don't want to watch boinc manager if a clean energy project is not jet finished.
Sorry when there is a correct thread somewhere I didn't see.

I have no problem in hibernating although it has been a while since I did that. Generally if I foresee myself having to completely shut down the computer, I uncheck CEP a day ahead of time, then I won't have CEP running when I do a shutdown. And check CEP when I am ready to resume.