Posts by Richard Haselgrove

1) Message boards : Questions and problems : BOINC Manager not showing tasks (Message 103244)
Posted 10 hours ago by Richard Haselgrove
Post:
Ouch. That's why I also copy and paste, then amend. Not totally foolproof but it saves me relying on my memory.
Trouble is, copy'n'paste with incomplete edit can be even worse...

#3249 fixes #3247.
2) Message boards : Questions and problems : BOINC Manager keeps attaching to a different account (Message 103239)
Posted 11 hours ago by Richard Haselgrove
Post:
AFAIK not IP address as this may change.
I was thinking of the case where the server was trying to identify a machine from garbled information.

https://github.com/BOINC/boinc/blob/master/sched/handle_request.cpp#L78

// find the user's most recently-created host with given various characteristics
...
if (strlen(req_host.domain_name) && strlen(req_host.last_ip_addr) && strlen(req_host.os_name) && strlen(req_host.p_model)) {
...
I was forgetting that the computer name is the most fundamental of all... :-(
3) Message boards : Questions and problems : BOINC Manager keeps attaching to a different account (Message 103236)
Posted 12 hours ago by Richard Haselgrove
Post:
I'm finding it hard, from my knowledge of how BOINC is designed to work, to see how this solution could relate back to the original problem. I've had a look through my own registry, and in all the - multitude of - entries, I haven't seen a single one which relates to a user account key. Those are all stored in the files in the BOINC data folder.

The data folder - by design - does not get deleted when the BOINC programs are uninstalled. This is to allow for version upgrades: an upgrade requires the removal of the old version, and its replacement by the files for the new version. The plan is for the user's data to remain unchanged during this process, even to the extent of keeping part-processed work intact, and restarting it at the exact point the previous version left it at.

All of which is a long way of saying that this may be a temporary fix - don't be too disheartened if the problem returns in the future.

One thing that hasn't been discussed so far is the potential role of project servers in all this. The BOINC server code contains tools which try to re-unite a lost computer with its owner's account. From memory, it considers things like hardware configuration and IP address - there may be others. Might it be a possibility that one of the projects you're attached to is using a buggy version of that code? It would be a hard one to track down, but the first step would be to make a note of the computer's 'HostID' or 'Computer ID' at each of the projects it is attached to - it will be different in each case. Check it locally first ('Properties' button on the 'Projects' tab, BOINC Manager, Advanced view). Then check that the same number is used on your own account on each project's web site.

Then, if it changes again to the old account again in the future, check the list of HostIDs again. I would expect that at least one - hopefully just one - will also have changed. Go to the project web page, and see what account the new HostID is attached to. That would give us some clues to work with.
4) Message boards : Projects : bad news everyone, results of the BOINC Collatz project are invalid (Message 103233)
Posted 1 day ago by Richard Haselgrove
Post:
My concern is that, whatever the credit scale that the project - any project - has chosen to pay credits at, David converts them back into GFlops at the defined, Cobblestone, rate. And then claims that the resulting figure is a true measure of the contribution that volunteer computing collectively makes to science. In the case of the project which started this discussion, that claim is quite possibly fraudulent.

But It's David's reverse calculation from credit to flops which is fraudulent, not the project's way of allocating credits in the first place.

I wrote it up once at SETI@Home.
5) Message boards : Questions and problems : BOINC Manager not showing tasks (Message 103185)
Posted 2 days ago by Richard Haselgrove
Post:
The other, more permanent, way of fixing this is to add the owner to the boinc security group:

sudo usermod -a -G boinc ownername
And reboot
6) Message boards : BOINC client : Stuck on "BOINC screensaver loading" (Message 103138)
Posted 3 days ago by Richard Haselgrove
Post:
I think it's highly unlikely to be "the task" which causes this, but more likely to be "the type of task". Try to gather information about which type of task is running when this happens.
7) Message boards : The Lounge : Apple M1 & BOINC (Message 103126)
Posted 4 days ago by Richard Haselgrove
Post:
Bernd Machenschalk (administrator of Einstein @ Home) has announced that they have started beta testing the first (to my knowledge) native Apple silicon science application for BOINC:

We do have the first native app version for Apple Silicon (M1) available. It is a FGRP5 app. So with a new M1 Mac and the latest "universal" BOINC Client you should get native work for your machine. Note that this build is pretty hackish, so far I don't even know whether it runs on any other machine than it was built on. So far the app completed a handful of tasks, all pending validation. That's also something I awaiting eagerly, whether the results of this app really validate. Oh, and yes, this is a "Beta" App, so make sure in your preferences that you want to receive "Beta test work".
Feel free to test, and report feedback direct to Einstein.
8) Message boards : BOINC client : Move Client to another drive (Message 103102)
Posted 7 days ago by Richard Haselgrove
Post:
To save me typing it all over again, try message 98168
9) Message boards : Android : BOINC Android client 7.16.16 released (Message 103097)
Posted 7 days ago by Richard Haselgrove
Post:
Not quite ;-)
10) Message boards : The Lounge : Grumbles, Glory and Covid-19 (2) (Message 103082)
Posted 8 days ago by Richard Haselgrove
Post:
People who are invited to be vaccinated, make an appointment, go to get vaccinated at the selected location and time. THEN refuse to be vaccinated for no good reason at all.
Total oxygen wasters.
Probably the same people who go to a polling station to spoil their ballot papers. It's called 'abstaining in person'.

I did that once. A particularly pointless election, where nobody knew any of the candidates. I wrote a long version of "none of the above", complaining about the waste of money.

Spoke with the person in charge of the election, later (in the pub) after he was off duty. "You spent a long time writing in that voting booth", he said. And a few moments later: "A lot of people did that".
11) Message boards : The Lounge : Grumbles, Glory and Covid-19 (2) (Message 103063)
Posted 8 days ago by Richard Haselgrove
Post:
And that has something to do with the message board timing out?
Well, if it involved loading new software, and then reinstating a code database onto a replacement server that's trying to handle too much already, it might. Just a suggestion.
12) Message boards : The Lounge : Grumbles, Glory and Covid-19 (2) (Message 103061)
Posted 8 days ago by Richard Haselgrove
Post:
Who is DoS'ing the SETI messageboard?
It's timing almost every time.
No idea, I'm seeing the same.
Raistmer emailed Eric K yesterday, to ask him to re-instate the SVN repository holding all the SETI application code...

https://einsteinathome.org/content/igpu-beta-app-validation-issue#comment-183576
13) Message boards : Promotion : Is it ok if I put the BOINC logo on my car? (Message 103060)
Posted 8 days ago by Richard Haselgrove
Post:
I have stickers for organisations that I'm a member of on my car - they get me free parking at some venues. Other people have bumper stickers - either jokes, or showing support for environmental or political causes. I doubt something like that would cause any problems, and might spark some interesting conversations.

But I'd steer clear of a giant-sized paint job that might imply sponsorship or endorsement by the wider BOINC community. Might not look good if you found yourself in an embarrassing position while driving it!
14) Message boards : Android : BOINC Android client 7.16.16 released (Message 103052)
Posted 8 days ago by Richard Haselgrove
Post:
You could look at https://github.com/BOINC/boinc/milestone/24?closed=1, but I don't think it'll be very helpful.
15) Message boards : Android : BOINC Android client 7.16.16 released (Message 103032)
Posted 10 days ago by Richard Haselgrove
Post:
Saw that. He probably wouldn't have bothered if I'd reported it myself...

One down, still 11 to go.
16) Message boards : Android : BOINC Android client 7.16.16 released (Message 103025)
Posted 10 days ago by Richard Haselgrove
Post:
Very wise. Time to hand the shift over to the upper-left USA. See if they can sort out some fixes by the morning!
17) Message boards : Android : BOINC Android client 7.16.16 released (Message 103022)
Posted 10 days ago by Richard Haselgrove
Post:
Another one for you. When downloading the new version from the BOINC website (as advised in the announcement), the BOINC download page advises you to go to the Google play store instead. Left hand, right hand?
18) Message boards : Questions and problems : GPU tasks skipped after scheduler overcommits CPU cores (Message 103005)
Posted 11 days ago by Richard Haselgrove
Post:
It's probably easiest on Einstein if you work out a reasonable target for what you can work off before you start hitting deadlines (given your various resource constraints), and then abort the excess sooner rather than later. No shame in that, if it's going to happen anyway, and better for the project if it happens sooner, so the resends can go out quickly.

If the website says you have more in progress than you can see locally, brace yourself for resent lost tasks - Einstein has that setting switched on permanently.
19) Message boards : Questions and problems : GPU tasks skipped after scheduler overcommits CPU cores (Message 103002)
Posted 11 days ago by Richard Haselgrove
Post:
Well, Einstein's 384 limit is implemented as a single global value. So it's a pretty blunt weapon, more of a long stop than something you should expect to ride on day by day.

Ease off any app_config.xml files that contain a <max_concurrent> value, and drop your cache (OK, it's low enough already) down to well under a day. Your first task is to burn off enough of the overfetch to get out of EDF, so that CPU tasks become normal priority again. By then, you'll have returned a few tasks and have some slack to play with.

I'm not sure how Einstein implemented the back stop: if it's saying "daily limit", you'll probably have to wait until midnight (German time) to start filling the cache with GPU tasks. "Max tasks in progress" might have been better here, but I suspect their rathed old server doesn't have that option.
20) Message boards : Questions and problems : x86 or x64 - what is faster? (Message 102998)
Posted 11 days ago by Richard Haselgrove
Post:
Also the better addressing of large memory spaces can help when pulling data from memory rather than disk
Thus encouraging bloatware, instead of mean and clean programming - and slowing things down again.


Next 20

Copyright © 2021 University of California. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.