Posts by HAL9000

1) Message boards : The Lounge : The Seti is Down Cafe (Message 83852)
Posted 13 Dec 2017 by HAL9000
Post:
Seti is still down.
And Einstein has crashed under the strain. Going to be a fun day all round.

My backup projects are getting a bit more than normal amount of run time recently.
If my backup projects start to fall over under the extra load I may have to resort to my secondary backup projects.
2) Message boards : Projects : Which BOINC projects use the most memory? (Message 81628)
Posted 29 Sep 2017 by HAL9000
Post:
Now that the Lattice Project is closed, the highest demanding projects are typically those that run virtualbox applications. However, as SekeRob stated, look at WUProp's home page and click on Results. You can filter the projects in many ways and see what memory usage is being reported for various platforms.

https://wuprop.boinc-af.org/results.py

I had been wondering what happened to Lattice. One day it was down, later it looked like it was possibly being updated, and then it was gone.
3) Message boards : BOINC client : Starting BOINC clients and attaching projects, except from a single terminal (Message 80534)
Posted 29 Aug 2017 by HAL9000
Post:
Thanks BobCat, that's a link I didn't know about.

However, it doesn't look like --detach_console will complete the Initialization process on the client before returning control, so running a control command like
boinccmd --project_attach URL WeakKey

...immediately after starting the client will almost certainly cause the control command to be executed before the client is initialized, resulting in an error.

If --detach_console is to work, there will have to be a subsequent command to wait until the client is started.

All of that said, I think the --attach_project command (on the client itself) will work just fine.

Thanks for the help!

Well there is a command for that.
TIMEOUT [/T] timeout [/NOBREAK]

Description:
    This utility accepts a timeout parameter to wait for the specified
    time period (in seconds) or until any key is pressed. It also
    accepts a parameter to ignore the key press.

Parameter List:
    /T        timeout       Specifies the number of seconds to wait.
                            Valid range is -1 to 99999 seconds.

    /NOBREAK                Ignore key presses and wait specified time.

    /?                      Displays this help message.

NOTE: A timeout value of -1 means to wait indefinitely for a key press.

Examples:
    TIMEOUT /?
    TIMEOUT /T 10
    TIMEOUT /T 300 /NOBREAK
    TIMEOUT /T -1

Normally I'll give BOINC about 5 seconds before sending any commands to it.
I will also run a benchmark before attaching projects. So that I don't have to wait for it to run on its own.
4) Message boards : The Lounge : The Seti is Down Cafe (Message 80412)
Posted 22 Aug 2017 by HAL9000
Post:
Might we ask if you are involved in the 2017 WOW project?? I have my rigs working away on two separate teams for it. What are you doing?
He's looking for a 6 slice toaster to boost his crunching :-)

One day...
5) Message boards : Questions and problems : Boinc 7.6.33 on Win XP keeps exiting (Message 79149)
Posted 20 Jun 2017 by HAL9000
Post:
A couple of days ago my BOINC Manager 7.6.33 keeps exiting on start up.
I have installed it on a WinXP system (not as a service) for more than a year now crunching SETI, WCG, CPDN, Rosetta, Cosmology, Enigma and Milkyway without any problems.
I reinstalled the Manager a couple of time but the problem remains. I went to the event log but nothing appears!!
Any ideas on that?

If you have not already done so. Check the BOINC stdoutgui.txt. It may state the reason it is exiting in there.
6) Message boards : BOINC Manager : BOINC Manager also exiting when closing minimised window (Message 79148)
Posted 20 Jun 2017 by HAL9000
Post:
This seems to be an old problem
https://boinc.berkeley.edu/dev/forum_thread.php?id=8601

I don't believe the explanation that it's a windows function. Windows doesn't close programs by itself but sends a message. BOINC manager isn't handling it correctly.

It goes back even further than that. I use BOINC Manager 6.10.48 with the BOINC 7.6.33 client and it has the same behavior.
7) Message boards : News : NSF funds new model for BOINC (Message 78899)
Posted 11 Jun 2017 by HAL9000
Post:
Doesn't using a BOINC account manager eliminate the need to setup accounts on each project?
I have not used one. So I could be certainly be mistaken.
I was under the impression a user would create an account, log into the account manager, and then select projects to run.


I think to really catch on with mobile devices it would need to "play to the audience".
An app with a simple interface would be needed. Along the lines of scanning a QR code and then let them sign in with their social media account.
Then to provide the user some incentive to let the software run. Why will people buy a $3 pink wrist band to help fund breast cancer research, but not let an app run on their phone while it charges overnight?

Having the infrastructure and an organization where groups/projects could go to have their data analyzed would probably let more types of research get done.
I have seen BOINC projects setup that try to do that, but most of them are not that great at it from the user end.
8) Message boards : Projects : SETI (Message 78073)
Posted 20 May 2017 by HAL9000
Post:
Would you be thinking of other kinds of communication that currently fall into the realm of scifi at the moment? Something like subspace communications on Star Trek?
Probably the closest thing we would have with our current level of knowledge would be quantum entanglement. I don't believe there is a way to detect the communication between entangled particles at this point in time.

As other communication methods are understood I imagine they would also be utilized in searches.
I figure it doesn't hurt to look with the technology we currently have. Since you can't find something if you never look.
9) Message boards : Questions and problems : Can I reduce I/O using memory? continuation (Message 77861)
Posted 11 May 2017 by HAL9000
Post:
IIRC threads are automatically locked after 60 days of inactivity.
Usually you can hit the red X and then request that you would like your thread unlocked.


I guess with rebooting everyday there is a higher chance something could go wrong saving the ramdisk data file. I usually don't reboot unless I need to. My HTPC has currently been going for 73 days w/o a reboot.
The version I'm using has an option to keep the previous disk image as a .bak. Maybe the one you are using has that feature as well?
10) Message boards : Questions and problems : energy saving , processor clock cycle frequency , environmentally harmful (Message 77571)
Posted 25 Apr 2017 by HAL9000
Post:
reasonable not programmers top scope to go in that direction . programmer and user interest differ a little in home use . i can only say , boinc has a chance to run on every desktop if it causes no costs . but programmings effort has to see a valuable benefit in calculations done , i understand . ok , thank you

Since BOINC is now more community driven development. Perhaps some kind of "green mode" could be devised?
It is unknown how desirable such a feature would be to users.
11) Message boards : Questions and problems : energy saving , processor clock cycle frequency , environmentally harmful (Message 77561)
Posted 25 Apr 2017 by HAL9000
Post:
Really modern computers don't have wasted idle time like older computers once did. Older computers, with Pentium 4 era CPUs, used only fractionally less power when not processing as they did at full load.
Since a modern desktop runs ~10w at idle and ~70w at full load there is less incentive to use a computers "idle time" for other purposes in environments outside of ones home.
12) Message boards : Questions and problems : Cloning a BOINC preinstalled VM (Message 77536)
Posted 22 Apr 2017 by HAL9000
Post:
Automatically adding projects can be done again in multiple ways. If you don't want to use a static list you could use an account manager or if you have a static list you can provide an initialization file.

HAL9000: with this process you should also run into the one host-CPID for all cloned hosts issue as the host-CPID is generated at first startup and is based on the first available MAC. If you clone the VM after this was set all clones will use the same host-CPID. The Client can generate a new host-CPID if it detects a clash but that will only work if the first available MAC on the cloned VM changed after cloning from the original VM.

FritzBox360: I'll send you a PM with my skype username.

For some reason I was thinking the CPID was generated when a host first contacted a project since I hadn't had an issue.
So I started up 10 VMs from the same image and watched the process.

All hosts initially had the same CPID.
Attached hosts the same project.
All hosts still had the same CPID after project initialization.
My normal process it to check the project/hosts_user.php for the new hosts I am attaching to a project.
If I don't see them I issue an update request to BOINC

Sent project update request to hosts.
Most hosts generated new CPIDs.
Repeated updated request, twice more, until all new hosts were listed on project/hosts_user.php

I now see why I wasn't finding an issue with my process.
The tasks on the hosts that generated new CPID were marked as "Cancelled by server". Which is an error I generally ignore & may not even see as I don't necessarily check the stats of tasks everyday.

I found a workaround for my process.
I found setting <host_cpid></host_cpid> before starting BOINC on each new VM host was sufficient to generate a new CPID.
Which lets me avoid each new VM running a full BOINC initialization and benchmark before getting started.
13) Message boards : Questions and problems : Cloning a BOINC preinstalled VM (Message 77511)
Posted 21 Apr 2017 by HAL9000
Post:
What you are trying to do sounds much like what I have done on my VM server.

My process:
1. Create OS environment. OS was configured use the hostname of BNC-xxxxxx. Where xxxxxx is the last 6 of the MAC address from the VM.
2. Add BOINC
3. Configure BOINC client. Run benchmarks, set permissions, and configure for remote access.
4. Create start script for BOINC that attaches BOINC client to desired projects after BOINC starts.
NOTE: BOINC should not be attached to any projects until after the image is saved.
5. Save OS image.
6. Launch multiple instances of VM image.
7. Crunch lots of projects data.

sample BOINC start script (for Windows VMs)
boinc --detach
timeout 30
boinccmd --project_attach http://project1 auth
boinccmd --project_attach http://project2 auth
boinccmd --project_attach http://project3 auth
14) Message boards : Questions and problems : Windows 10 getting BOINC certificate error (Message 77410)
Posted 18 Apr 2017 by HAL9000
Post:
The ca-bundle.crt file in 7.6.33 expires on 2018/08/13.
On my Windows 10 notebook (also Version 10.0.14393) I'm not getting and messages about expired certs.
I do not have BOINC installed as a service,but I don't believe that should matter... unless it does.
15) Message boards : Questions and problems : Win10 Creator's Edition update breaks nVidia driver, manual reinstall fixes it (Message 77393)
Posted 16 Apr 2017 by HAL9000
Post:
I updated my Win10x64 to the Creator's edition and BOINC was throwing the GPU unavailable error message.

A manual install of the latest driver fixed the problem. 381.65

My first guess is that the video driver was updated.
The video drivers provided via Windows update are not the ones you want to use for processing tasks.
Normally they are stripped of the required components for the GPGPU applications to function. Such as CUDA or OpenCL.
16) Message boards : Questions and problems : energy saving , processor clock cycle frequency , environmentally harmful (Message 77392)
Posted 16 Apr 2017 by HAL9000
Post:
Running BOINC on a CPUs lowest power state is easy to do. At least in Windows.
Define a power policy with the CPU maximum set to the desired value, such as 25%, and then only have BOINC run applications when that policy is active.
I actually use this method on one of my notebooks.
17) Message boards : The Lounge : The Seti is Down Cafe (Message 77332)
Posted 12 Apr 2017 by HAL9000
Post:
In the BOINC code it still is server_status.php: https://github.com/BOINC/boinc/blob/master/html/user/server_status.php
It even says in the documentation:
BOINC-based projects offer the following XML export at <code>URL/server_status.php</code>
But Seti being Seti, they always do things differently.

@Gary, the HTML page was officially deprecated in favor of the PHP page. It's still being updated (and cached), but it's not the official page.
The HTML page was deprecated because the new bootstrap code couldn't handle the page being HTML. Remember that when you were on the HTML page that you didn't see that you were logged in (on the right), and didn't have a link to your account from Project? That's why it is now no longer used, and instead we have the PHP page.

**Edit: Aarrgghhh, extra white lines.

It makes things even more odd since SETI@home Beta is using server_status.php
18) Message boards : The Lounge : The Seti is Down Cafe (Message 77327)
Posted 11 Apr 2017 by HAL9000
Post:
Yup, they've changed the page name again. It used to be server_status.php, and is now show_server_status.php
So if you have a bookmark, like good old me, then it points to the old page, which in turn goes to the old HTML link when the project is down. And because the old HTML link is no longer used, it's the colourless B&W page.

I thought that changed with the new layout?
The bookmark I have points to show_server_status.php. The oldest backup of my bookmarks I currently have is from Jan which also has show_server_status.php.
19) Message boards : BOINC client : My Wish List (Message 77288)
Posted 11 Apr 2017 by HAL9000
Post:
BOINC needs to be made NUMA aware especially now since AMD's Ryzen processors appear to act as if they.... This work will also benefit Intel Xeon processors that are set up to use cluster on die mode....


Seems that Numa is an old request:
https://github.com/BOINC/boinc/issues/1357
http://boinc.berkeley.edu/dev/forum_thread.php?id=10124#60953

I don't understand if Numa is now supported or not

On my dual Xeon systems BOINC seems to run fine across both NUMA nodes.
I did try a few tests comparing NUMA enabled and disabled. I didn't see anything difference between the two configurations when processing project tasks.
20) Message boards : Questions and problems : Controling blend of mutlicore and single core VM's (Message 77287)
Posted 11 Apr 2017 by HAL9000
Post:
I finally got it to work

The project_max_concurrent is jobs, so it will always run that many tasks regardless of the number of cores per task.

I forced the number of CPU's as per your suggestion.

Now it respects the number of cores correctly. I can also limit the number of cores in use with the Use at most N% or the <ncpus>.

I think the concept of a job is quite confusing, not sure why it would be useful, max cores per project would be better?

I would guess because only a few projects use multi-threaded apps that there are probably some unexpected issues when running such apps.
Milkyway is the only project I can think of right now that has multi-threaded apps. Which I need to get around to throwing one of my dual E5-2670 machines at someday.


Next 20

Copyright © 2019 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.