Hey All,

I just put up some new nbody runs. They will be the first to run off the new version (v170). I waited some time to put these up so the queue would clear out of the old workunits.

Please let me know if anyone encounters any issues with these runs.

As always, thank you for you continuing support,
Sidd

I never get any nbody WUs. I have both checked in my milkyway preferences. Any idea why not???

Hey Sid,

Noob here... I'm noticing (I think) that the nbody WUs seem to grab precedence from Wus already in progress, but maybe I am mis-reading my task window. I notice this when restarting computers after resting them at night.

Unrelated: I never see any transfer activity in the Transfers window. But Wus show Ready to Report, eventually disappear, and my project total keeps incrementing up. So I presume all is well. Just curious.

tnx!

chuck

Turned on Nbodys again. They still cease making progress after a while, but keep the processors. Turned off Nbodys and aborted those in my que.

Getting checkpoint errors on Nbody and Milkway tasks.
Nothing works

I am getting errors with work units on Milkyway. WU's don't seem to be going down in the ETA, I had one say was going to take 731 days, up from the 4 hours when it downloaded.

Seems there's some issue somewhere.

Sidd,

Starting to see nbody WUs stall after running for a few hours. Remaining time stops decrementing but can be induced to decrement by suspending and restarting. Both computers seen with this problem are dual core WIN7 SP1 with BOINC 7.12.1. One is Intel. Other is AMD. One has GPU other does not.

As a result I am not making a few nbody deadlines. Single core WUs seem to run normally.

Hi Sidd,

Will there be a Mac version again for the new nbody run?

I am experiencing problems with N-Body simulation 1.7. I am seeing almost no CPU activity although 8 CPUs are assigned. Elapsed time is accruing but remaining time is increasing as well. No progress. A 3-1/2 hour work unit has 19 hours elapsed and now shows 4-1/4 hours remaining. I suspended all N-Body 1.7 work units and Home 1.46 work units were then picked up; all CPUs now running at around 75%. Please advise.

Hey,

There seems to be an issue with openmp and the newest versions of mac, in that it is seemingly not supported anymore. We have someone working on that and will hopefully have a multi-threaded mac application soon. We have a single threaded application, but there was an error in putting it out; we are also looking into that.

All in all, we are currently working on getting our mac users an Nbody binary, but it may take some time unfortunately.

Thank you for your support,
Sidd

Hey,

That seems a bit odd. I would go ahead and cancel it. It could be that the workunit got stuck somehow. I will look into why that is.

Thanks for letting us know,
Sidd

More info on the N-Body issue.

Some are allocating 8 CPUs and only using 15%-20% of 4 CPUs.

Examples:
de_nbody_7_23_2018_v170_20k__data_1_1533467104_35769_0 de_nbody_7-23-2018_v170_20k__data_2_1533467104_35738_0

For the two I mentioned earlier, between 11:00 and 13:36 their progress stayed at 71.234% and 61.237% and so I canceled them.

I am still seeing issues with the N-Body. The work units get picked up and then they stall out while retaining the CPUs so other work units don't get processed.

The following example shows a 1.25 - 1.50 hour work unit that ran for almost 12 hours and attained only 7.52% completion. Estimated remaining time went from 1.5 hours to over 8 days.

8/25/2018
7:08:02 PM | Milkyway@Home | Starting task de_nbody_7_23_2018_v170_20k__data_2_1533467104_174546_0

2018-08-26
08:23 | Running (8 CPUs) - but 5 other single-CPU jobs running, plus 1 partial with GPU. All N-Body task estimates between 01:15 and 01:30
7.52% | 11:56:08 | 6d 02:49:41

12:50 |
7.52% | 15:56:28 | 8d 04:05:40