CE application ce7-pub-beta 1.03 work units are overtaking my systems

j
jjch ID: 21776 Posts: 11
11 Jun 2014 05:52 AM

A few days ago CE pushed out a new application ce7-pub-beta 1.03 and it is overtaking several of my systems. The work units seem to keep coming in even though I have tried setting the "No new tasks option" It also appears that they are not following the application switching time which I have set out to a longer timeframe on most systems. These systems are all running BOINC version 7.2.42.

The only way I can allow other project work units to run is suspend the CE project and abort the CE work units. Even when I do that some CE work units still download but since the project is suspended they don't run. I don't mind the extra points for CE but the other projects are taking a hit and I prefer to keep an even workload. As far as I can tell ce7-pub-beta is not following BOINC rules.

Another oddity is that my settings do not have Beta Test applications selected. I really wouldn't mind running Beta Test units on one or two systems if I had requested them and set them to run that way. I just don't like them being pushed out and not having the appropriate control over them. Please look into this and make the needed adjustments so CE will play along fairly with the others.

Thanks,

JJCH

Matt ID: 44 Posts: 222
11 Jun 2014 03:26 PM

Hi JJCH.  Sorry to hear you’re having troubles.  A couple comments/questions — 

By “overtaking your system” — I assume you don’t mean that general performance of your device(s) was affected, but rather that CE took a greater share of your designated volunteered computing resources, i.e. at the expense of other BOINC projects you may have added in addition to CE?  But can you confirm?

If the above is true, the issue has likely now been resolved (*also please confirm or give further detail if not).  But having said that: Running BOINC projects other than those managed by CE is not our general use case and is not fully supported.  That is: CE is not a BOINC project like SETI@home or Einstein@home; it mostly plays nice when run alongside them, but there will from time to time be (*non-critical) wrinkles in such scenarios, and the resource allocation issue you encountered was one of them.

As for the beta test — the CE7 app, like other apps, was initially released only to beta users (i.e. not to you).  after testing, the app was released to the general population.  But: it was not renamed at that time (i.e. “beta” was not dropped from the filename), so this quite understandably led to confusion.  apologies for that.  and, of course, this confusion is compounded since you had trouble — furthering your reasonable conclusion that the app was a beta app.  anyhow, we’ll take additional steps to manage app filenames moving forward.

j
jjch ID: 21776 Posts: 11
12 Jun 2014 05:17 AM

Matt,

Yes, the CE work units were taking over the resource share and not allowing other projects to run. It does appear that the issue has been resolved now. I am seeing CE work units running along with the others.

 

I have been able to run CE along with other BOINC projects fairly well. I have the resource share set so that all the projects get a fair amount of time and spread out the work evenly.

The one other thing that throws off the balance is when the work units go into high priority and run full power for awhile. This is most noticible on the GPU units as then they ignore the temperature settings I have them set at.

When you move the beta applications to production it would be great if you can rename them so users don't think they are getting beta test units. Maybe change Beta to Prod or just leave it out.

I may consider running Beta test units in the future but I haven't figured out a way to set them up so they only run on one or two machines. Since I would prefer to keep most of them running the real work.

Thank you,

jjch

 

M
Mardo ID: 643026 Posts: 2
24 Dec 2014 04:16 AM

Hi Matt,

Let me clarify this with you. Ideally, if a user wishes to participate in the prize draw s/he must let CE account manager allocate the tasks for their PC to crunch and not select any projects outside CE?

If that user selects projects outside CE to work alongside their CE assigned tasks, this can cause system errors/instability issues? Points not calculated accurately? Unable to participate in the prize draw?

Have you heard of Folding@Home? They have their stand alone program separate from BOINC. What do you think of it? Would this be an ideal setup for CE?

Thank you for your time.

 

Mardo

Matt ID: 44 Posts: 222
24 Dec 2014 08:12 PM

Hi Mardo.

Yes, the idea is that participants should let CE manage their projects.  99%+ of CE users have never heard of BOINC and expect (and require) the app to be ‎fully automatic.

Having said that, it is possible to add BOINC projects outside the set managed by CE (ie, from the desktop software "add project" UI).  In this instance, {1} you will get fewer points, because only a fraction of your compute time will be allocated to CE; and also {2} in some cases there will be issues and we're less able to support these cases. Which is to say: participating in BOINC projects should work and doesn't disqualify you, but it does limit the support we can provide because issues will be somewhat more frequent and debugging irregular setups gets very labor intensive.

(I'm not sure what you mean about F@H?  Maybe others here have direct experience, but I'd advise against running F@H and BOINC or CE at the same time.)

M
Mardo ID: 643026 Posts: 2
27 Dec 2014 01:03 PM

Hi Matt,

An explanation of Folding@Home from

http://folding.stanford.edu/home/faq/faq-simulation/

Like both examples, it’s important for us to know about the motion of a folding protein, although we also want to know the intermediate steps along the way. Our simulation methods construct models of both of these properties. One way that makes Folding@home different from some other distributed computing projects (Rosetta@home for example) is that we want to see how the car parks, not just the end state of seeing it parked. While that’s an important result, it doesn’t shed any light on how or why a protein sometimes misfolds. By attempting to study all of the possible paths that the bouncing ball can take down the mountain, we learn a lot about the question “How did we get here?” It also gives us the opportunity to introduce changes – such as drugs – into the process that modify the probability of misfolded results.

Matt ID: 44 Posts: 222
28 Dec 2014 05:08 PM

Thanks; yes, I've heard about Folding@home.  They do very good work (*though I think the above does not do justice to the very excellent work which is likewise being done by Rosetta).  --- I just wasn't sure what you were asking vis-a-vis F@H and CE.