Gamma ray searches not showing cruncher details anymore?

Jasper
Jasper
Joined: 14 Feb 12
Posts: 63
Credit: 4032891
RAC: 0
Topic 198663

This morning, I noticed something new (to me). When clicking on a work unit to see its details, this would usually bring up a list of crunchers working on that particular unit, me and one or several others depending on deadlines. Not so today (my TZ): the list is not there anymore, ’Tasks in progress’ shows ’suppressed pending completion’. This is only true for FGRPB1, not GW; I don´t know for others, as these are the only two I get for not having a suitable GPU in my iMac.

Apart from the words stating accurately what I see, nothing, what does this mean really? Should I worry?

Christian Beer
Christian Beer
Joined: 9 Feb 05
Posts: 595
Credit: 188466662
RAC: 241178

Gamma ray searches not showing cruncher details anymore?

Interesting. I didn't know that would happen. We activated a feature called adaptive replication where we don't send out a second task if we think the host can be trusted to deliver a good result. This is only activated for the FGRPB1 search. Don't worry. I'll investigate and see what that really means next week.

What I think this means is that you don't see if there is a wingman until you deliver the result and then see if you are the only one crunching this job or not.

Jasper
Jasper
Joined: 14 Feb 12
Posts: 63
Credit: 4032891
RAC: 0

Thank you for the mighty

Thank you for the mighty quick reaction, Christian!

I´m happy enough already, knowing that there is no reason to switch off that part of the crunching. It does not seem working allright quite yet though, as I picked that feature from my completed but pending validation tasks list, not from the ones in progress. I guess that you will look at that also next week, cannot imagine there to be a hurry with this.

floyd
floyd
Joined: 12 Sep 11
Posts: 133
Credit: 186610495
RAC: 0

Jasper, you're confusing your

Jasper, you're confusing your task and the corresponding work unit. You are looking at the work unit details page and there it says "suppressed pending completion." Apparently you're not supposed to know who's working on a work unit until it is completed. I can't imagine the reason for this but I've seen it before, at Malaria Control if I remember correctly.

Jasper
Jasper
Joined: 14 Feb 12
Posts: 63
Credit: 4032891
RAC: 0

I don´t think I´m all that

I don´t think I´m all that much confused - although that´s arguably subject to debate - but just discussing what I see as of this morning, as opposed to until yesterday. ðŸ˜

I am indeed looking at a WU details as seen here when clicked from my validation pending taks, or also as seen here when clicking from my tasks in progress. Both are for FGRPB1. Until yesterday however, this showed exactly like seen here when clicked for a GW work unit.

It´s not really important of course, especially not when there´s no wingman at all, but in that case it should not keep showing pending methinks. It´s for sure not something requiring urgent action in any case.

Richard Haselgrove
Richard Haselgrove
Joined: 10 Dec 05
Posts: 2143
Credit: 2958199595
RAC: 713174

Presumably, if your task is

Presumably, if your task is pending, you can infer that this particular task has been selected for re-verification by a wingmate.

From what I've seen in the past at MilkyWay, where adaptive replication was also used (from my observation - the staff there didn't seem to be very clear about how they have BOINC set up), the scheduler decides whether to send a check task only when the initial sole replicant is returned: and at MilkyWay, you could see that process happening in open view. I don't know whether the difference in visibility is a server version difference or a configuration difference.

Edit - it's configurable:

Quote:

Do not hide incomplete results when using adaptive replication.


http://boinc.berkeley.edu/trac/wiki/ProjectOptions#website

floyd
floyd
Joined: 12 Sep 11
Posts: 133
Credit: 186610495
RAC: 0

What I tried to say is that

What I tried to say is that "suppressed pending completion" on the work unit page means completion of the work unit, not completion of your task. There is no "your task" or "you" on this page and obviously the work unit is not yet complete, probably because you're waiting for some unknown wingman. I'm sure all will be revealed once all is done.

Just for the record, I don't like it this way.

rbpeake
rbpeake
Joined: 18 Jan 05
Posts: 266
Credit: 1132497797
RAC: 753717

RE: ...We activated a

Quote:
...We activated a feature called adaptive replication where we don't send out a second task if we think the host can be trusted to deliver a good result. This is only activated for the FGRPB1 search.


This seems such a great way to increase efficiency for the project as a whole! I hope it works out.

Christian Beer
Christian Beer
Joined: 9 Feb 05
Posts: 595
Credit: 188466662
RAC: 241178

I thought a bit more about

I thought a bit more about this suppression of pending jobs (workunits). There are two possible problems when you know if there is a wingman or not.

The first is if you know there will be no wingman you can manipulate the result on your computer and potentially harm the project by sending in falsified scientific results.

The other one is when you have tasks that have a wingman and tasks that don't you might be inclined to abort or starve the ones with a wingmen because you are in a race with another team and want Credit as quick as possible.

For now we are going to leave it as it is. I'll bring this to the attention of the FGRPB1 scientists and they can discuss if this might have an impact on their scientific results.

Jasper: you obviously have wingmen on those tasks where you already uploaded and reported the tasks. Once those tasks are in and validated you will see who it was.

rbpeake
rbpeake
Joined: 18 Jan 05
Posts: 266
Credit: 1132497797
RAC: 753717

RE: I thought a bit more

Quote:

I thought a bit more about this suppression of pending jobs (workunits). There are two possible problems when you know if there is a wingman or not.

The first is if you know there will be no wingman you can manipulate the result on your computer and potentially harm the project by sending in falsified scientific results.

The other one is when you have tasks that have a wingman and tasks that don't you might be inclined to abort or starve the ones with a wingmen because you are in a race with another team and want Credit as quick as possible.

For now we are going to leave it as it is. I'll bring this to the attention of the FGRPB1 scientists and they can discuss if this might have an impact on their scientific results.

Jasper: you obviously have wingmen on those tasks where you already uploaded and reported the tasks. Once those tasks are in and validated you will see who it was.


Makes sense. Some people will always try to beat the system.

Jasper
Jasper
Joined: 14 Feb 12
Posts: 63
Credit: 4032891
RAC: 0

RE: Jasper: you obviously

Quote:
Jasper: you obviously have wingmen on those tasks where you already uploaded and reported the tasks. Once those tasks are in and validated you will see who it was.

Yes, that (now) makes perfect sense to me: I completed one about an hour ago and it is showing validated right away, here.

@Richard: thank you for your input, this is exactly how I´m used to see it once in a while at MilkyWay, when for whatever reason I am getting no new WU from Einstein. However, if I recall correctly, the WUs there are always showing their details, whether I´m alone or getting company. I can easily be wrong on that though, it has been a long time ago when that happened.

Comment viewing options

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