N-body moved forward some version steps today.
We moved from 1.18 to 1.32.
Only binaries 1.30 and 1.32 were pushed to the servers.

1.30 and later versions address the segmentation fault reported by some users.
1.32 addressed an error in the mathematical return value of the likelihood calculation.

There may be some tweaks in this release as we watch the results come back.
To test the release the following runs were started:

de_nbody_07_11_no_dark
ps_nbody_07_11_no_dark

de_nbody_07_11_dark
ps_nbody_07_11_dark

Please post any errors or issues you may encounter to this thread.


Thank you,


Jeff

Bad task here and here.

Also, there was a Win 7 (B 7.0.28) host assigned to the default plan_class on the former.

This one doesn't look good. :-(

<edit> Although after look further into it, this host doesn't have any successes showing for any version of nBody. So most likely the problem here is specific to this host, possibly the version and/or configuration of Linux its running.

Thank you. The first one is peculiar. I will look into it.

I just started running these n-body tasks, so I'm not familiar with them, in general. I know that it's a small sample size, but I'm running an error rate of 15%-20% overall (as of now, 8 errors in 43 tasks). All stderr output is similar to this:

<core_client_version>7.0.64</core_client_version>
<![CDATA[
<message>
The system cannot find the drive specified.
(0xf) - exit code 15 (0xf)
</message>
<stderr_txt>
<search_application> milkyway_nbody 1.32 Windows x86_64 double OpenMP, Crlibm </search_application>
Using OpenMP 4 max threads on a system with 4 processors
<search_likelihood>1.#QNAN0000000000</search_likelihood>
Failed to calculate likelihood
21:45:13 (3196): called boinc_finish

A variety of hosts are involved, a mix of Win7-64, Linux, Intel, and AMD. All hosts with errors have returned valid results. It doesn't look like it, but are there any unusual RAM requirements? Any problems running n-body alongside GPU tasks that require a dedicated CPU or make significant use of one?

MarkR

---

The error you are seeing will be resolved in the next release. It is a consequence of correcting the likelihood calculation. For now, there is nothing you can do about it.

Jake

The error you are seeing will be resolved in the next release...For now, there is nothing you can do about it.

Thanks for the feedback, Jake. Is there any way of knowing when that's going to be? Just curious...

---

I am just about to announce the release!

I am just about to announce the release!

Excellent...Thanks for the update, Jake.

---