Boincview and Boinc 6.6.14

[KarVi]

Has anybody else experienced problems with boincview and boinc 6.6.14.

Got 6.6.14 up and running yesterday, and at first things where fine.
But now I cant start boincview, it crashes when it contacts boinc on my server running 6.6.14. If I remove the server from Boincviews "computer.ini" file, it runs fine, but when I add the 6.6.14 server in boincviews "Computers" dialog, it crashes immediatly when contacting it.

I am a little sorry this has happened, since I liked the ability of Boincview, to monitor more computers from one program.

Are there any good alternative to Boincview out there?

[Richard Haselgrove]

I am running BoincView 1.4.2 on my P4 with BOINC v5.10.13, and using it to view and control six other computers, three of which are running BOINC v6.6.14

There are minor problems with displayed estimated completion times, but some of the earlier problems with BV and BOINC v6.6.12 have been eliminated - David Anderson has commented on changes he has made specifically to ensure continued compatibility with BoincView. It certainly doesn't crash, or even show error messages.

Are you sure that the problem isn't caused by some other factor? BV is notoriously memory-hungry, and it might be being crashed by the size of the new caches made possible by BOINC v6.6.14

There is also a new BOINC v6.6.15 (similar pre-Beta status to v6.6.14) which has still more BV compatibility changes. You could try that.

[KarVi]

I forgot to mention that I was running the latest available Beta of boincview. Something must have broken in that version, but reverting to 1.2.6 made it work perfectly again.

Will look into getting a never version running, but for now I'm OK.

[Richard Haselgrove]

BoincView website has vanished off the web (as has the author), but my original download of BV 1.4.2 (final non-Beta version) is attached.

Included licence file grants permission for re-distribution like this.

[attachment deleted by admin]

[KarVi]

Thanks.

Will try it out.

Yeah the website vanished around New years eve. But the developement was stopped long before that and the author never posted, or commented in the forum.

Sad that this very good tool isn't developed anymore, and that it does not seem anybody has the source code, so someone else could take over.

[maugou]

Hi everybody,

first, I'm not english native language, forgive some quirks in my expression,

I'm currently on my way to develop a BV replacement. While looking for knowledge on the boinc measurements in BV, I stepped into this forum and decided to talk about for the first time.
I recently decided to give a try on BV and since then I enjoyed it much. So, I wandered in the many forums and discovered the latest sayings of BV's author. I became unhappy with the BV's fate. Being currently in an unemployed state, I decided to begin this adventure, for one month now.

I'm currently able to mimic most viewing part of BV.

I'm looking for the method used by BV to determine 'estimated speed', 'efficiency', 'estimated credits'. I've downloaded the boinc sources, so I begin to have an overall understanding on the matter, but I'd appreciate some advices to speed up the development. Even a suggestion on who to ask, where to look, would be great.


I've tried to attach an image of the current state of my app. Let's see if it worked!

Regards,
Maurice


[attachment deleted by admin]

[Mike O]

I to am in the process of doing a re-creation of BV in VB.net
Maybe we can work together on this? What Language are you usings and what version?
Im using VB.net  9 of Visual Studio 2008
As far as the CPU Eff. What I found is that its a ratio between, cpu time and elapsed time in the form of : cpu time divided by elapsed time. This works but a better way is to keep 2 samples of the cpu time.. 2 samples of the elapsed time .. subtract the first elpased sampled from the 2nd (t1) and then subtract the first cpu time from the 2nd cpu times (t2)..  and then divide t2 by t1. This should give a better number then the other way.

Im trying to find a good way to deal with the network connections to the hosts.. Im seeing times fo 1/10 to 7/10ths of a second doing RPC_Request calls.
i have threads setup for the host table data creation but the comm is killing the CPU.. It seems to be loading it much MUCH more than I ever expected.
Heres A pic of what I have gotten done in the last month or so. (I work full time so I dont have the time to put in to this that i'd like to)

[Pappa]

I am not sure if I am quailified to answer your question. I can talk about behavior
With My seting BV is set to poll 10 machines every 20 seonds. The reason is it it is set to poll on intervals shorter than that it starts using considerable CPU Time.

I/you could setup a sniffer to look at the actrual transactions, to see what infromation is actaully being requested/returned. Then one would has to figure out the best filter to actually get the correct information.

Thast said the information you are reqeusting is very "broad" and takes time to prepare.

What I suspect is that somewhere in there he looked at the specific Pecent of CPU usage for the Seti process and running CPU time and Processes time until he stuck a balance ( this was compared to the total number of processes).

So as when the RPC happened he retireved  the CPU time/percentage for ALL Processes

From that, if he has the reported information about CPU(s) Speed then things can be averaged over wahtever contact/reporting time.


An Array can be created for each host to hold the information
Then the polling interval contols the growthe/decay of the calculated process.

Regards

[Darr]

Hi everybody,

first, I'm not english native language, forgive some quirks in my expression,

I'm currently on my way to develop a BV replacement. While looking for knowledge on the boinc measurements in BV, I stepped into this forum and decided to talk about for the first time.
{snip}

Please let us know when you have a version of BMW ready for beta (or even alpha) testing, Maurice.

Yours too, Mike.

And an excellent sniffer is WireShark - http://www.wireshark.org/download.html

They have a U3 version there for those menus on USB sticks, and also a portable version for other types of USB sticks, SD cards, et al... so you don't have to install it on every computer where you want to use it... just install it on a thumb drive and run it from that. Source is there too, since it's a sourceforge.net project.

If you guys decide to collaborate on a BV replacement, you might consider making it a SourceForge project as well.

Anyone got a link for the 1.5 beta of BOINCView?

[maugou]

Hi Darr,

Mike and I have just started to exchange about our projects. I'm trying hard to get a testable version for the end of may, but this will be at best a preprealpha. I will probably devote less time after that moment. For short, there will be many issues at start.

BMW is obviously a humorous attempt to name the application. Another proposal would be Boinc Workbench (BW), more neutral, shorter, maybe clearer.

I know well wireshark, I'm most of the time a networking consultant. It helped me start this project but is less used at my current project status.

On my side, I've created an account on SF and not yet created the project, but this is one of my goals. I may also get some support from my Boinc team (l'Alliance francophone); that would be a big improvement in the handling of the tests and support.

I'll let Mike express his own feelings about it.

Here is a link where I found BV 1.5.8beta, http://www.setigermany.de/upload/BoincView1.5.8beta.zip, it has some nice additions but is less stable (freezes, memory hog, impossible connection to some host) than the other versions I've tested.

Regards,
Maurice

[efmer (fred)]

I stumbled on this thread.
And I started a similar project on my own. For different reasons.. I could not read my own computers with the BOINC manager because I had too many WU probably....
http://www.efmer.eu/boinc/boinc_tasks/ The program is almost far enough for me to start using it on my own computers.
But I still have a lot of ideas I want to implement.