Mac SMP Client Eating WUs

[chileman]

Hey guys, just seeing if I'm the only one seeing this issue. I'm running the Mac SMP client on my Macbook (2.0ghz, Core Duo). I run it through Increase 2.2. Of course I move my laptop around and I have it pause the folding when it goes to battery power. Issue is that even thought it pauses when I open it up and plug it back in, it usually deletes the WU and gets another, waisting thousands of points and time. Here's a snippet from the log.

[23:32:55] Completed 370000 out of 500000 steps (74 percent)
[01:26:38] CoreStatus = 1 (1)
[01:26:38] Client-core communications error: ERROR 0x1
[01:26:38] Deleting current work unit & continuing...
[01:31:07] - Warning: Could not delete all work unit files (8): Core returned invalid code
[01:31:07] Trying to send all finished work units
[01:31:07] + No unsent completed units remaining.
[01:31:07] - Preparing to get new work unit...
[01:31:07] + Attempting to get work packet

 

[chileman]

Oh, and after that in the log I found this

[15:41:43] Project: 2605 (Run 8, Clone 101, Gen 17)
[15:41:43]
[15:41:43] Assembly optimizations on if available.
[15:41:43] Entering M.D.
[15:41:49] mdrunner cpfilename:
[15:41:49] Rejecting checkpoint
[15:41:51] OPC
[15:41:51] Writing local files
[15:41:51]
[15:41:51] Writing local files
[15:41:52] Extra SSE boost OK.
[15:41:53] Writing local files
[15:41:53] Completed 0 out of 500000 steps (0 percent)

 

[gigabyte1024]

This is a beta client.
Maybe instead of telling the client to pause, you should just stop the client before you unplug and start it when you get the juice flowing again. Or just let it run on battery too, the Core Duo is pretty effeciant.

 

[chileman]

I know, bitten by the beta. I was just seeing if I was the only one seeing this or I was doing something wrong.

 

[Macaholic]

Be sure to check out the known bugs list. As mentioned, it may likely be an issue with the pause on battery that you are doing. Keep in mind that it can take up to 10 to 15 seconds to stop the client when you simply Control-C in terminal. Pausing and quickly starting back up again may be causing havoc. Check your OS X crash logs to see if anything is amiss. As for the "rejecting checkpoint" statement in your FAHlog, that is normal fare when you start up the client again (one of those pesky bugs).

 

[BillR]

I know, bitten by the beta. I was just seeing if I was the only one seeing this or I was doing something wrong.

I’ve been looking for the post at the community forum, but I recall VJ mentioning that one or more of the beta clients had to have the checkpoint part of the program disabled.

You might want to do a search as well because at the moment I can’t find that post again.

If you fall into that category you are kind of forced into running two separate standard clients since you shut your machine down.

If you have been folding successfully with that client and the problem just started I have found that almost none of the SMP clients clean up after themselves very well, if at all. My cure for this is at least once a week after a WU completes I shut down FAH and delete all the contents of the work directory then restart the client. After you do that expect a few error notices but then the client will pick up and run find again.

Hope that helps a little anyway.

 

[chileman]

I’ve been looking for the post at the community forum, but I recall VJ mentioning that one or more of the beta clients had to have the checkpoint part of the program disabled.

You might want to do a search as well because at the moment I can’t find that post again.

If you fall into that category you are kind of forced into running two separate standard clients since you shut your machine down.

If you have been folding successfully with that client and the problem just started I have found that almost none of the SMP clients clean up after themselves very well, if at all. My cure for this is at least once a week after a WU completes I shut down FAH and delete all the contents of the work directory then restart the client. After you do that expect a few error notices but then the client will pick up and run find again.

Hope that helps a little anyway.

Well it kinda puts my back against the wall. They don't have an Intel single core client, just a Intel SMP. And trust me, it's not worth running the PPC version in Rosetta (43days/wu). So I think I need to wait a little longer after pausing/stopping the client and I think I might be ok. It doesn't always do it so it's probably me. Maybe I'll put together a Mac SMP install guide and maybe one to do with Increase. Thanks for the help guys! Hope I don't lose any more WUs, I hate losing 930PPD.