Why this message from project?

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4699
Credit: 17542421714
RAC: 6374041
Topic 195704

I decided to run an anonymous platform so I could crunch 2 tasks at a time on my two Fermi class GPU's. Why does the project now throw this message?

(| Einstein@Home | To get more Einstein@Home work, finish current work, stop BOINC, remove app_info.xml file, and restart.)

I seem to able to get all the work I desire. Both CPU's and GPU's are getting fed well.

 

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5842
Credit: 109381836165
RAC: 35971149

Why this message from project?

Quote:
... Why does the project now throw this message?


My assumptions about this are along the following lines:-

AP was never intended as a means of running the standard apps on known platforms. It was always intended as a mechanism for running user compiled apps on otherwise unsupported platforms. It has been found that AP can be made to work if you want to run a standard app under non-standard conditions (ie two tasks simultaneously) on a known platform. The server-side portion of BOINC notices that you have a supported platform and so (in its humble opinion) thinks that you shouldn't be using AP - hence the message, together with a 4 hour backoff presumably to give you time to correct your 'error'. Provided you have a sufficient cache of work on hand to survive these 4 hour backoffs, you should never have a problem. Once the backoff counts down, you will always be able to get more work - and another backoff as well :-).

For various purposes, I've been using AP for years with various versions of both client and server BOINC software. The behaviour described above has been much the same for as long as I can remember and is not specific to any particular BOINC version.

Cheers,
Gary.

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4699
Credit: 17542421714
RAC: 6374041

RE: RE: ... Why does the

Quote:
Quote:
... Why does the project now throw this message?

My assumptions about this are along the following lines:-

- hence the message, together with a 4 hour backoff presumably to give you time to correct your 'error'. Provided you have a sufficient cache of work on hand to survive these 4 hour backoffs, you should never have a problem. Once the backoff counts down, you will always be able to get more work - and another backoff as well :-).

Thanks for the reply and explanation. I did in fact notice that after every communication with the project, that I now do get a 4 hour backoff. I was worrying about that also. I seem to be in good shape with a four day cache setting and also running two other projects. I have plenty of work usually to keep all the cores and GPU's occupied.

Cheers, Keith

 

DanNeely
DanNeely
Joined: 4 Sep 05
Posts: 1364
Credit: 3562358667
RAC: 1580

Is the 4 hour backoff new? I

Is the 4 hour backoff new? I don't recall ever seeing it when running Akos's custom apps back in the day.

Gary Roberts
Gary Roberts
Moderator
Joined: 9 Feb 05
Posts: 5842
Credit: 109381836165
RAC: 35971149

The backoffs didn't happen

The backoffs didn't happen way back then but they have been around for at least the last couple of years. I don't really remember when they started, but it seems like ages ago. They don't happen every single time. My impression is that around 10% of the time, an AP work request will NOT trigger the 4 hour backoff. I did discuss this previously and I did give an opinion about why there was no backoff with some requests but there were other comments which put doubt on the correctness of that explanation.

Cheers,
Gary.

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2139
Credit: 2752648717
RAC: 1486471

RE: RE: ... Why does the

Quote:
Quote:
... Why does the project now throw this message?

My assumptions about this are along the following lines:-

AP was never intended as a means of running the standard apps on known platforms. It was always intended as a mechanism for running user compiled apps on otherwise unsupported platforms. It has been found that AP can be made to work if you want to run a standard app under non-standard conditions (ie two tasks simultaneously) on a known platform. The server-side portion of BOINC notices that you have a supported platform and so (in its humble opinion) thinks that you shouldn't be using AP - hence the message, together with a 4 hour backoff presumably to give you time to correct your 'error'. Provided you have a sufficient cache of work on hand to survive these 4 hour backoffs, you should never have a problem. Once the backoff counts down, you will always be able to get more work - and another backoff as well :-).

For various purposes, I've been using AP for years with various versions of both client and server BOINC software. The behaviour described above has been much the same for as long as I can remember and is not specific to any particular BOINC version.


As Gary says, Anonymous Platform is a standard feature of BOINC and has been for many versions and many years. However, the 4-hour work fetch backoff when using AP is specific to the Einstein project.

My understanding is that, whilst Einstein is open-source, and user-compiled versions for unsupported platforms are tolerated, they aren't exactly encouraged. The project has mentioned concerns in the past that data from hosts running unverified software might be called into question during the scientific peer-review process prior to the publication of any results in academic journals.

That objection doesn't apply, of course, to people using the official project-supplied applications under unconventional operating conditions. But the whole point of anonymous platform, of course, that it is anonymous - the server doesn't know whether the software that is being run is official or homebrew. So the gentle 'discouragement' is applied to the AP mechanism as a whole, rather than to specific applications.

The other big use of AP at Einstein has been during the beta-testing of new software. I think the 4-hour backoff first appeared (or at least I first noticed it) at the end of an intensive period of software testing (was this when Akos's optimisations were being incorporated into the official codebase, I wonder?). I think at least part of the intention was to 'nudge' software testers back onto the straight-and-narrow of project-maintained, and hence automatically updateable, application versions.

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4699
Credit: 17542421714
RAC: 6374041

RE: ... Why does the

Quote:
... Why does the project now throw this message?

Quote:

That objection doesn't apply, of course, to people using the official project-supplied applications under unconventional operating conditions. But the whole point of anonymous platform, of course, that it is anonymous - the server doesn't know whether the software that is being run is official or homebrew. So the gentle 'discouragement' is applied to the AP mechanism as a whole, rather than to specific applications.

As I do in fact only run the official project applications for Einstein and MilkyWay at least (Seti - different story), when are we going to get a major rewrite of the BOINC platform so as to enable things like setting GPU's to do multiple tasks concurrently in the platform itself, instead of having to resort to anonymous platforms? It would be nice to have the BOINC platform automatically keep us updated with the latest apps like normal users instead of fumbling through edits of our custom app_info's. What is the latest news or rumors?

 

Bernd Machenschalk
Bernd Machenschalk
Moderator
Administrator
Joined: 15 Oct 04
Posts: 4265
Credit: 244921706
RAC: 16859

This message is a legacy from

This message is a legacy from the times when there was only one application on Einstein@home. The idea is to tell you there is no more work for the application that is configured in the app_info.xml, usually when a new GW run has started.

I'm currently preparing an update of our scheduler, I'll see if I can remove or otherwise fix it.

BM

BM

Keith Myers
Keith Myers
Joined: 11 Feb 11
Posts: 4699
Credit: 17542421714
RAC: 6374041

RE: This message is a

Quote:

This message is a legacy from the times when there was only one application on Einstein@home. The idea is to tell you there is no more work for the application that is configured in the app_info.xml, usually when a new GW run has started.

I'm currently preparing an update of our scheduler, I'll see if I can remove or otherwise fix it.

BM

That would be lovely, Bernd. I have run into lack of work due the 4 hour backoff imposed while the Seti project went dark yesterday for power testing at the lab. Anything you manage to do will be greatly appreciated.

Cheers, Keith

 

Darmok
Darmok
Joined: 18 Jun 09
Posts: 6
Credit: 147327812
RAC: 0

As I am a recent addition to

As I am a recent addition to the E@H crunchers community, though I have searched for an explanation for the message "Got server request to delete file XXXXXX" in the forums or FAQ, I did not find any. Can somebody tell me what it means, what I am supposed to do (or not), and if I do, why doesn't Boinc Manager delete the files automatically in "Tasks", if that is the intent. This message is daily and pertains to at least 10-20 tasks in waiting or processing.
Thanks

Mike Hewson
Mike Hewson
Moderator
Joined: 1 Dec 05
Posts: 6534
Credit: 284700169
RAC: 113726

RE: As I am a recent

Quote:
As I am a recent addition to the E@H crunchers community, though I have searched for an explanation for the message "Got server request to delete file XXXXXX" in the forums or FAQ, I did not find any. Can somebody tell me what it means, what I am supposed to do (or not), and if I do, why doesn't Boinc Manager delete the files automatically in "Tasks", if that is the intent. This message is daily and pertains to at least 10-20 tasks in waiting or processing.
Thanks


The task log for your computer lists very many user aborts. Are you aware of any reason for that ?

Cheers, Mike.

I have made this letter longer than usual because I lack the time to make it shorter ...

... and my other CPU is a Ryzen 5950X :-) Blaise Pascal

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.