Hi. I like this option combined with the option of "run until completion". I think for us with laptops with Dual Core, we need an option to select 1 task, and tell BOINC to run it ON THIS CORE until completion. This way, I still have the other core available to switch among the other projects I have.

I'd like this option specially to run climate prediction. I'd like to run these tasks non-stop on one the cores.

Maybe an option based not on task but on project? So I select one project and assign it one core, and let the other core for all the rest.

Regards,

Leo

I think this would be a good idea, as I have 2 GPU's and run 2 CUDA projects, SETI & GPUGrid. So to be able to allocate a project to 1 GPU if needed, or even permanantly. And as I too run ClimatePredictor, I would like the option to run on 1 CPU core 24/7.

But this would of course have to be linked to the Schedueler.

---

MAIN PC:
Q6600 @ 2.40GHz - Windows 7-64Bit 4GB RAM - BOINC 6.6.38 Win64 - 9800GX2 190.38
NETBOOK:
N280 @ 1.66GHz - Windows 7-32Bit 1GB RAM - BOINC 6.6.38 Win32

Work Unit Count:

Here's a simple idea. Okay, as everyone already knows, under the Projects tab, these are the normal fields/coloum headings: Project, Account, Team, Work done, Avg. work done, Resource share, status, now here my idea for a new heading...Work Unit/s Count

Now lets say you have 23 SETI WU's, and are currently running 4 WU's, then under this field you would see 4/23.

It would be nice to be able to look at projects, and see how many WU's I have and how many are running. If I had a project like SETI, where no WU's are running then 0/23 could be displayed instead.

As I am attached to 10 projects, and as sometimes I have an endless amount WU's displayed under Tasks, it would be nice just to be able to what I have, and what is running. I know I can sort them Status, but does not say how many WU's I have.

---

MAIN PC:
Q6600 @ 2.40GHz - Windows 7-64Bit 4GB RAM - BOINC 6.6.38 Win64 - 9800GX2 190.38
NETBOOK:
N280 @ 1.66GHz - Windows 7-32Bit 1GB RAM - BOINC 6.6.38 Win32

A BOINC Manager Task Window sidebar button or menu option, or both, to open up a file browser window into the data-directory. Too often the question to support when seeking error log info "Where can I find it?".

That's the job of a support person, to listen to (or in this case read) the same question again and again and again. You can always tell people to restart their BOINC, then have them read out the line that says Data directory:.

With that movement of the log files to a subdir of the data-dir, so we don't have to explain what their names are... make the extention .log instead of .txt too.

Explain that the logs are made in std*.txt files, that the .txt is so it can easily be opened in any text editor, while no OS will confuse it with anything else than a text file.

std means standard. (Although it could also be save to disk. ;-))
out means output.
err means error.
gui means graphical using interface. This is BOINC Manager.
dae means daemon. This is BOINC, the core client.

If people can't be bothered to learn a little about their program, you can always just point them to http://boinc.berkeley.edu/trac/wiki/ReportBugs, where it is explained what the various debug options do and how the various log files are used. :-)

---

Perhaps this has been asked before, but how about a proper right alignment of all the numbers in the Work Done and Avg. Work Done columns?

Can't see anything wrong in 6.6.12 interface.

Ok, that seems to be solved then. I'm still using 6.4.7 and it's still a minor issue there.

(b) "Projects" and "Tasks" listings: The capability of sorting the current project listings according to a column title of our choice, either ascending or descending. (For example, list projects in order of increasing time to completion, or by decreasing report deadline, or by progress or by project name).

They can already be sorted, since like a year ago...

---

Under Tasks, Boinc displays CPU Time...and because of CUDA projects would it be possible to display both GPU Time and total Elapsed Time?

I know that not everyone uses CUDA, so maybe these would only be displayed if Boinc detected a CUDA compatible GPU.

Nobody has yet figured out a good way to measure "GPU time".

---

1) where it can be seen that a workunit may not be returned within the specified deadline.

BOINC should already notice that and prioritize the workunit. If it doesn't, it's a bug; please give more details on a specific case where it doesn't work.

3) or suspending individual tasks of that 2nd project, which can be quite cumbersome when managing tasks that number 30 or more, as well very inefficient utilization.

You can select multiple tasks and click "suspend" just once.

---

As far as it hasn't been asked before:

1. An Update ALL (Projects) button
2. Split screens under the Projects tab (i.e. Projects and Tasks), so you can see the tasks for every project you click on in just one screen)
3. Proper shut down procedure (w/o asking to connect to another host every time (there's already a seperate menu option that does just like that))
4. Column with date and time of WU download
5. Sort WU's in order of date and time of download (now the sort is gone after a BOINC restart)

BTW, when I choose Accessible View in 6.4.7., the running WU entries are flickering every few seconds.

1. An Update ALL (Projects) button

Select all tasks, click Update.

---

1. An Update ALL (Projects) button

Select all tasks, click Update.

Great, thanks for the hint.

3. Proper shut down procedure (w/o asking to connect to another host every time (there's already a seperate menu option that does just like that))

5. Sort WU's in order of date and time of download (now the sort is gone after a BOINC restart)

These are available in BOINC 6.6

---

3. Proper shut down procedure (w/o asking to connect to another host every time (there's already a seperate menu option that does just like that))

5. Sort WU's in order of date and time of download (now the sort is gone after a BOINC restart)

These are available in BOINC 6.6

That's nice. Exactly how stable is 6.6 currently?

I wonder, Do any of the Crunchers here use BoincView?

So what do You think?

Notice the graphic? It says Boinc Manager - (Multihost) version: 7.6.66

And look at how many PCs are mentioned(up to 99).

I'd like to see BoincViews superior ability to monitor more than one PC at a time from one place(BV also does monitor multiple projects all at once by just clicking on the logged in icon for the PC one selects) Incorporated into the Boinc Manager. Instead of having separate windows as there is only so much desk space and only so much screen real estate.

It's only a picture of course, But It's what I mean, As in BoincView like(or Lite), This was a screen grab done with Paint Shop Pro 4.12 running under XP x64 that I altered abit. As this is what I meant earlier:


Projects: would show multiple PCs and what projects their attached to,
Tasks: would be expanded as shown above and Ready to report/Ready to start entries would be displayed for Multiple PCs on this tab too,
Transfers: would show the various hosts transferring their files to their respective projects,
Messages: would display the actions of attached PCs and their actions and/or what message they receive,
Statistics: would expand upon what It does already to multiple PCs,
Disk(usage): would be expanded to cover Multiple PCs.

I'd code It Myself If I knew C++(As I have the Borland C++ compiler here, If It's compatible with Boinc Managers present code base that is), So all I can do so far is suggest. As I did a Google search not too long ago(Approximately Recreated Here) and there is a demand for this ability as BoincView is no longer supported(the Author of BV Sebastian Masch[1.25/1.42/1.50b8] is either offline or dead, No one knows and no one knows where the source code for BV is as the Author never released It, Although BV is Free to share, I have tried in vain to find any trace of a contact through an existing website and email address, Both are gone and one is up for sale) and BV doesn't properly support cuda tasks(reporting functions to the screen on total time crunched is off as BV thinks the WU is a cpu task as BV doesn't know anything about CUDA/gpu tasks).

---

/SJ's PCSpecs

I would like to see some sort of 'speedometer' so I can see how fast BOINC is running.
The statistics graph doesn't show work done ... just credit.

I have no way to 'tune' things and measure changes in performance.

---

--
benDan

The statistics graph doesn't show work done ... just credit.

What measurement unit do you suggest using?

---

I will be happy if in
'Avancé'-> "Preferences"-> "Utilisation du processeur"
you add more fields to allows use of processor on multiple periods.

In France we have for elctricity, two period with not the same prices. This will allow to work on the wanted period that are chipper, not one only in the actual version of boinc manager on one period.

It will be too much if you add a total hour conter on each of these periods.

My second will, color the process in Task with the same color use in statistique graph, and when the wu is runing increase the color.

The statistics graph doesn't show work done ... just credit.

What measurement unit do you suggest using?

I don't really know.
I have been trying different tweaks but I have no way to measure changes in throughput.

"If you can't measure it you can't manage it."

Ideas anyone?

---

--
benDan

The statistics graph doesn't show work done ... just credit.

What measurement unit do you suggest using?

I don't really know.
I have been trying different tweaks but I have no way to measure changes in throughput.

"If you can't measure it you can't manage it."

Ideas anyone?

Some people consider the credits to be a measure of the througput. It works not too pretty bad. See how credits are calculated.

This can take hours or days to show up. I really want to see whats happening in realtime.

---

--
benDan

This can take hours or days to show up. I really want to see whats happening in realtime.

You can't get that in real time. For example, if you install an optimized application that creates invalid results, it would be lowering your throughput (to zero). But you won't know about that until the project compares your tasks against what other users processed, which may take hours or days.

---