We have finished testing a new version of Nbody.
I will be packaging the release to push it in 24 hours.

Due to the holidays I am posting this ahead of time so everyone can be aware that a new version of the nbody software will be available. It will be version 1.04.

We will also be starting a run with the new nbody version.

I keep updates in this thread as we go live with the new version.

Thank you,
Jeff Thompson

in case of errors (hopefully there are none), should we be posting them here or should we start another thread?

---

Team Belgium

I was going to post everything to this thread. To keep it in one place.


Jeff Thompson

I have updated the binaries on the server and am getting ready to push a run of nbody code. The version reported by the nbody software should display as 1.04.

Thank you,


Jeff Thompson

Hello,

I can proudly confim that it works, reports and validates.

Congrats

Thank you for the confirmation.

I am noticing a fair amount of linux users posting with version 0.94 of the nbody software reporting errors. Please update your version to 1.04 if this is happening to you.


Thank you,

Jeff Thompson

I'm getting work units and they are computing, reporting, and validating successfully.

Jake

works good here.
just wondering what its computing there. anymore info there?

Ameyah

Right now, we are trying to fit several orbital parameters of a fictitious stellar stream that is loosely modeled after the Orphan Stream. Each work unit is computing a different set of parameters, and the idea is to see if we can obtain the expected values (the ones that we generated the stream with). After doing this test, we can move on to give you guys workunits that will be intended to fit real streams, ones where we don't know the parameters since they aren't generated according to our will (unfortunately).

Thanks to the users for their continued support and being patient with us as we push this application forward.

Jake

Pleased to say that the Windows version (Win7/64) is running fine on host 465695.

It's not trusted enough for instant validation yet, but the signs are good. I'll check how it's doing it in the morning.

I seem to be having issues since the new version on nbody 1.04. 2 items, one is it no longer uses all available cpu threads, in the past it used all 8 threads and now is using between 2 and 7. and when I checked log it seems that all of the new work I've completed states "computation error".

Phil
ID 481742

---

no issues here, all crunches fine and validates. One question, though. How come the granted credit is so low for nbody? I have one nbody WU at 550 seconds crunch time and it only granted a few points.

---

Team Belgium

CPU Runtime variation is huge: 32 to 4700 seconds. I wonder why?
Wasn't nbody supposed to be multi-thread? Currently, the app runs single thread (app per core)

I missed this topic before (and accidentally started my own in Number Crunching), but to me, the 104s are multithreaded, which seems to work well. However runtimes vary between 10 seconds and several hours, I pre-emptively aborted a 62(!)hour WU, since my computer may well shutdown before that time, and I'm not willing to test checkpointing on such a large job.

Although the WU says it's OpenCL, it does nothing with my GPU, has the -np 6 --device 0 options, and is only using CPU. Credits are thus low to very low, but that might be correct.

I have 4 nbodies running one at 14hrs and 60hrs to go, one at 14hrs and 1000+ hrs to go and 2 at 100+ hrs to go.should I abort these.I also notice that that they are version 104, but in my account they are listed as 100.

Please look in this thread in numbers as well, these WUs should probably not be sent to GPUs, as they just obviously waste their power.

---

Grüße vom Sänger

The CPU usage variations or workunits are coming from the parameter sweeps, some sets of parameters will be computationally easy while others more complex.

We are looking at the range of parameters we are going to use on the next runs that should address some of these issues in the variation in CPU usage of workunits.

We are looking into the other issues and may have some tweaks released in the near future. I don't have specifics right now but I am taking the units with error and looking at the different issues with the group.

For the threads and GPU usage I can not answer on right now.


I tested the threads because of observations on a previous run where only one thread was generated and could not duplicate the issue but I will look at host specifics on that.

Updates will follow in this thread and I will get a reply in the other threads shortly.


Jeff Thompson

The 1.04 work units are very bad on a linux machine.

boinc     3439  0.2  0.3  64572 15496 ?        Ss   18:57   0:02 ./boinc --allow_remote_gui_rpc --daemon
boinc     7164 35.3  0.0  14940  1820 ?        RNl  18:57   7:05  \_ ../../projects/volunteer.cs.und.edu_subset_sum/SubsetSum_0.11_x86_64-pc-linux-gnu 52 27 190600788427830 2203961430
boinc     7165 34.3  0.0  14940  1828 ?        RNl  18:57   6:54  \_ ../../projects/volunteer.cs.und.edu_subset_sum/SubsetSum_0.11_x86_64-pc-linux-gnu 52 27 190605196350690 2203961430
boinc     7166 35.1  0.0  14940  1820 ?        RNl  18:57   7:03  \_ ../../projects/volunteer.cs.und.edu_subset_sum/SubsetSum_0.11_x86_64-pc-linux-gnu 52 27 189606801822900 2203961430
boinc     7167 33.9  0.0  14940  1824 ?        RNl  18:57   6:48  \_ ../../projects/volunteer.cs.und.edu_subset_sum/SubsetSum_0.11_x86_64-pc-linux-gnu 52 27 189646473128640 2203961430
boinc     7645  326  1.1  77160 46400 ?        RNl  19:05  40:07  \_ ../../projects/milkyway.cs.rpi.edu_milkyway/milkyway_nbody_1.04_x86_64-pc-linux-gnu_mt__opencl_amd_ati -f nbody_parameters.lua -h histogram.t

Yes, it is taking all of 3 1/4 cores from a 4 core machine leaving precious little for other work. Is this intentional behavior?

just had to abort a WU whose estimated time was 70 hours to complete... it crunched for 10 hours before I noticed it.

---

Team Belgium