I got one. Ran it for just over two minutes for it to checkpoint at 1% or so.
Restarted it with LAIM off and it went back to 0.204%.

got 6...

workunits with 298 in them are the start of the flexible work units.

295, 296, 297 are rigid work units with more than one job per workunits.

298 will only have 1 job per workunit. They are more complicated and may run longer.

Thanks,
-Uplinger

I got 4 on two machines: 3 flex and one rigid. The flex ones are running well and do seem to be checkpointing a little more often, and closer to the "Write to disk at most every" time which I have set to 180sec. The rigid one ran for nearly 28 minutes before it checkpointed (at 20% progress), but shortly thereafter I suspended and resumed it with LAIM off and it seemed to correctly restart from that checkpoint. (I'm monitoring checkpoints with BoincTasks 1.66 which may not be fully reliable in that regard.)

However (if the times are accurate) I think 28 minutes is way too long, and now at 37min progress appears stuck at 20%.

Edit: At 38 minutes progress appears to have retreated to 10%. This is similar to the strange numbers that appear on long-running tasks that don't checkpoint, so maybe not too surprising but certainly unwelcome (and very off-putting for newbies). No further checkpoints have occurred either.

Edit 2: Checked stderr.txt in the slot directory and, yes, it does look like it correctly restarted. Also, looking at the times of the .ckp files, it does appear that the checkpoints are at extended intervals. It just checkpointed for a second time at 55 minutes.

That rigid unit checkpointed a third time at 69 minutes. This time I noticed that the number of checkpoint files is growing with each checkpoint, both in the slot directory and in the contained vina_checkpoint directory. Is that supposed to happen?