Posts by Ageless

log in
21) Questions and Answers : Windows : No WU since a long time. Why? (Message 1813218)
Posted 28 Aug 2016 by Profile Ageless
Open BOINC Manager->(View->Advanced view->)Projects tab->select Seti@Home->Click Reset->Acknowledge. This will remove all old entries about files and applications from the client_state.xml file and the ..projects\\ directory.

It'll reset the state of Seti to as if you've just added it. It won't lose you any old credit, or anything like that. Just the old muck that you cannot use anymore to crunch with.
22) Questions and Answers : Windows : No WU since a long time. Why? (Message 1813214)
Posted 28 Aug 2016 by Profile Ageless
The simplest solution is to remove your app_info.xml

... and then reset the project. Just removing the app_info.xml file from the C:/Programdata/BOINC/projects/ directory won't cut it, as all the misinformation about the old applications is then still in the client_state.xml file.

Resetting the project will fix that.
23) Questions and Answers : Web site : New notices message. (Message 1812298)
Posted 24 Aug 2016 by Profile Ageless
That will stop the notification bubbles, it will not stop the Notifications tab from showing a number, so as to tell you there are new notices. That still happens after so many moons, but usually when a project has one new news item, BOINC Manager refreshes all older news as new news. Shrug.
24) Questions and Answers : Web site : Screensaver Missing (Message 1812210)
Posted 24 Aug 2016 by Profile Ageless
I wonder where you're seeing that my computer has completed only one (01) work unit?

When you click on anyone's name, you'll get to their external account.
Click on the "View" link for Computers.
Under "Computer ID" click on the "Tasks" link.
This opens onto a page showing all of the tasks that the computer in question has downloaded (In progress) and returned (other messages).

In your case that comes out on
When you look in the "Application" column, you can see what device was used to run the tasks. And although now all show SETI@home v8 v8.12 (opencl_intel_gpu_sah) windows_intelx86, I'm sure that when BilBg checked earlier this week, that one of them showed it had run on the CPU.

The database deletes tasks 24 hours after they were returned, as when they're all kept the database will grow too enormous to keep. The result data of the ran workunit was stored in the scientific database, that's the only one that holds results from all tasks run over the years. It's several terabytes big.
25) Questions and Answers : Web site : New notices message. (Message 1812205)
Posted 24 Aug 2016 by Profile Ageless
Open BOINC Manager->Options->Other options...->General tab->Notice reminder interval->set to Never->OK.
26) Questions and Answers : Unix/Linux : Raspberry Pi Build Memory Leak? (Message 1811864)
Posted 23 Aug 2016 by Profile Ageless
I'm starting to suspect a memory leak in the boinc manager.

Because it cannot be trouble with the memory on the device itself?
BOINC Manager is what we call the GUI, it doesn't do any of the calculations, it's just there to easily control the client. The BOINC client itself doesn't do any of the calculations either, it's merely storing the work, scheduling what runs next and doing the communications.

The actual work and memory load is done by the science applications, in your case solely Seti v8. These don't tax the memory that much, what I saw from when I ran it earlier this year, perhaps 50-75MB per task. So if you run on all 4 cores at the same time, it may be that you use 200-300MB total. But if somewhere in the first or second group of RAM there's an error, that can cause freezes.

Best thing to do is run a memory checker over it, most Linux versions have Memtest86 in their start-up menu. And else download it.

Something else you can try to run on it, to see if it is your system or BOINC, is Prime95 (GIMPS). Install and run for during a time line such as you ran Seti, see if that hangs the device. If it does, it's definitely hardware.
27) Questions and Answers : Web site : Screensaver Missing (Message 1811501)
Posted 22 Aug 2016 by Profile Ageless
You are running work on your Intel GPU. GPU applications do not have the graphics application, they never had those here at Seti. When in the past the development team ported over the science application to use Nvidia's CUDA, they decided to only port over the sc ience app, not the graphics app.

This decision was made with in mind that the graphics processing unit (GPU) was doing the calculations and that adding the graphics app/screen saver might make the error rate go up.

All science applications running on the GPU since that time are a continuation of that early app and are without graphics application thus without the project screen saver. If you want to see it, run Seti work on your CPU.
28) Questions and Answers : Unix/Linux : stock astropulse and app_info.xml (Message 1811326)
Posted 22 Aug 2016 by Profile Ageless
What you could start with is to post your present (if available) app_info.xml file, and tell if you want to run the Astropulse on the CPU or GPU. Perhaps also useful for a helper who doesn't have Linux is to tell what the application name is/are.

You can find that in the data directory, depending on what installation you used - Berkeley installer uses the home directory, OS maintained package uses /var/lib/boinc/ or /var/lib/boinc-client/ - in the /projects/ directory.

If in doubt about what data directory your BOINC uses, BOINC says that at the start of the start-up messages, found under BOINC Manager->Tools->Event Log, or CTRL+SHIFT+E.
29) Questions and Answers : Unix/Linux : Trouble downloading work to Raspberry Pi 3 (Message 1811310)
Posted 22 Aug 2016 by Profile Ageless
You get "Communication deferred" messages when BOINC doesn't recognize something in your setup, like the operating system used, or the CPU. The project has got
Seti v8 applications for ARM CPUs, but it needs to be specific.

What you best can do now is post your startup messages, all the way from Starting BOINC client version 7.4.23 for *** till you've put a message from the project in. The messages will tell what's wrong. They can be reached through the Event Log (Tools->Event Log or CTRL+SHIFT+E).
30) Questions and Answers : Windows : How to disable new tasks to repurpose computer (Message 1809366)
Posted 15 Aug 2016 by Profile Ageless
Open BOINC Manager -> (View -> Advanced view ->) Projects tab -> Select (hi-light) Seti@Home -> click "No new tasks". The button will change to "Allow new tasks" as this button shows the option you have, not the state it is in.

Run down your cache, report everything.
Once you're done with the changes, click "Allow new tasks" to get new work. If at least you didn't reformat and started anew.
31) Questions and Answers : Windows : Boinc upgrade sequence. (Message 1808854)
Posted 13 Aug 2016 by Profile Ageless
Thanks, I'll add that to the rewritten FAQ.
32) Questions and Answers : Windows : Boinc upgrade sequence. (Message 1808798)
Posted 12 Aug 2016 by Profile Ageless
I've set up the BOINC FAQs for just these things. One of them in there is Windows Installer Error 1706: Setup cannot find the required files / 1714: older version cannot be removed / BOINC.msi cannot be found / The feature you are trying to use is on a network resource that is unavailable / Error 1325. is not a valid short name with solutions on how to fix the problem.
33) Questions and Answers : Windows : Is it possible to run SETI@home without virtualBox? (Message 1807986)
Posted 8 Aug 2016 by Profile Ageless
VirtualBox installs its ow set of network drivers which replace the existing ones.

That is not true, Rob.
What VirtualBox will do is install a virtual network for its own use. This works on the side of the normal network, uses virtual hardware and is just 100Mbit fast. To be able to install the virtual hardware, the present network needs to be taken down, then after initialization of the virtual network it'll be restarted without any changes.

VirtualBox will not install drivers for the physical network card.
For more rechnical information, see
34) Questions and Answers : GPU applications : GPU show a very large work file (Message 1807898)
Posted 8 Aug 2016 by Profile Ageless
the missing ATI GPU sounds like it might be the problem.

Definitely not, because you don't have any work in cache assigned to an ATI GPU. There's only a reference in client_state.xml to an application that is used on an ATI GPU, but according to your task list there's no work assigned to it.

Your other host has an ATI GPU. Have you moved this GPU to that machine from the one you have the long turn around problem with? If you did so, and you installed the Nvidia GPU, did you uninstall the ATI drivers, and run a cleaner program to get rid of all remnants of the previous GPU?

Or does this host have two GPUs and did you not (re)install ATI drivers for the second GPU?
In any case, it's got nothing to do with your present problem of tasks showing a long time to completion. So if you want to solve that problem, I suggest you post a separate thread about it, so not to confuse this thread too much.

Can you go to and scroll to the bottom of that page, then select with the mouse (mouse over the text while holding the left button down) everything with percentages. When everything is selected, press CTRL+C to copy to clipboard. In an answer here, click the window to select it, press CTRL+V to paste it. You may want to check whether there's spaces needed, but else just post it.
35) Questions and Answers : Windows : Back after 10 years - New BOINC - reboot problem (Message 1807678)
Posted 7 Aug 2016 by Profile Ageless
99% = all but one core on any machine with 190 cores or less.

But my examples were to show that this "statement" is wrong

Yes, you're right. Teaches me not to walk away in the middle of an answer and come back later to then finish it with my mind set to the thing I was called away on.
36) Questions and Answers : Windows : Back after 10 years - New BOINC - reboot problem (Message 1807277)
Posted 5 Aug 2016 by Profile Ageless
This is because BOINC can't work out what to do with fractional cores when allocating tasks so it rounds to whole cores.

BOINC makes use of integers in the case of determining the amount of CPU cores. This translates into 100% = everything, 99% = all but one core on any machine with 190 cores or less. It'll get interesting at situations above 190 cores. But that's not the case here.
37) Questions and Answers : Getting started : Allowed WU by user (Message 1806762)
Posted 3 Aug 2016 by Profile Ageless
"a high-end GPU" = 1 GPU

BOINC isn't able to recognize a high end GPU from a cheap P.o.S. GPU. While it will read the amount of peak flops the GPU can do, it doesn't do anything with that information, neither does the project server.

The project also is not able to recognize a high end GPU from a cheap P.o.S. GPU.

So for that, it doesn't matter what GPU you have in the system.
Of course a high end GPU returns the work faster than a cheap P.o.S. GPU, but then it will also refill its cache faster.

There's really no need for a bit of a high end GPU to have thousands of tasks cached, for what if the inevitable happens and that machine dies? Or just another SNAFU happens?

Besides, I'd expect that for a bit of a competition you'd want to return your tasks ASAP when done, not have them sit in a cache and wait their turn to be done, then only reported when the long minimum (store for N days) is met. (Late edit: Unless the competition is to have as many tasks in cache as humanly possible...?)
38) Questions and Answers : GPU applications : GPU show a very large work file (Message 1806666)
Posted 2 Aug 2016 by Profile Ageless
Host 775768 shows an unusual large turnaround time (Average turnaround time 6.18 days) for a system with a GPU. That can only be reached when you have large restrictions in network availability, or you only sporadically run BOINC.

When you @densvin look at the details of this host, what are the percentages showing for "Percentage of time BOINC is allowed to run", "Percentage of time GPU is allowed to run" etc.? (Wording may be different)

If the time BOINC is allowed to run and do work is low, that will translate itself in large estimated run times that you have on tasks.
39) Questions and Answers : Unix/Linux : from Windows to Linux: can I move the "" folder? (Message 1806664)
Posted 2 Aug 2016 by Profile Ageless
You cannot use the Windows data directory or any sub-directories in it for Linux. I assume you want to try to run work you downloaded for Windows on Linux, but that isn't possible.

Tasks downloaded on Windows are registered against the hostID BOINC made for Windows. That hostID must also upload and report the finished results.

When you switch to Linux, a new hostID will be made by the BOINC you have on it. A hostID is composed of a brand and model CPU and an operating system. You cannot have a hostID with just a brand and model CPU and then switch between different operating systems using the same hostID. That would be possible if you switch from e.g. Windows XP to Windows 10 (as long as both are the same bitness), but not if you were to switch from any Windows to Linux.
40) Questions and Answers : Unix/Linux : boinccmd --project nomorework command returns error 02-Aug-2016 01:48:47: GUI RPC error: No such project Operation failed: Error -1 (Message 1806629)
Posted 2 Aug 2016 by Profile Ageless
Operation failed: Error -1

This means you have no permission to do this operation, possibly because you haven't set a password in gui_rpc_auth.cfg and made BOINC read it.

So try this:
boinccmd --passwd password --project nomorework

The URL doesn't need the http:// or the trailing /

If you do not supply the password in gui_rpc_auth.cfg, this file should be in the current directory, where you do all the commands from. Else you should supply the password.

By default BOINC adds a random 32 character hexadecimal password to this file. You best add your own.
Do not finish with an Enter, there should be no end of line or Return characters after the password.

Previous 20 · Next 20

Copyright © 2016 University of California