Posts by tekwyzrd

1) Message boards : BOINC Manager : Recent modifications to Boinc Mgr look - not liking it [have included suggestions] (Message 38392)
Posted 13 Jun 2011 by Profile tekwyzrd
Post:
As for the malariacontrol message, look at the date. That was five years ago on a long-gone dual p3. This computer is a single processor and is watching for orbit@home work but running einstein@home. One task at a time. BOINC's disk usage is set to 10GB with 58.49GB of the 80GB /home drive free.

Who's being defensive now? As if it could possibly NOT be anything to do with what I said. The setting up of the disk usage preferences hasn't changed in all these years, so the date doesn't matter. It also doesn't directly matter what CPU you have.

What will matter is what all the preference settings are, for memory and disk use. As then it does coincide with your old "This left no space on the partition for malariacontrol to save data."

And if that doesn't help, then you best ask at the project what could cause the error, as it's their science application throwing it. They will know how much real and virtual memory their application takes up, and thus what's best for the preference settings on those, especially the percentages.

It's very possible that a simple reboot helps out there. I see three tasks threw a signal 8 error with a stack fault, the other threw a signal 11. It all indicates something wrong with (virtual) memory or stuck in (virtual) memory.


It's obvious you get some perverse pleasure from jumping all over people after not totally comprehending what they write. It's sad when a formerly reasonable individual (and I used to consider you a knowledgeable and reasonable person when I was a frequent poster) turns into someone who rabidly attacks any criticism. I stated my opinion. I explained the problem I experienced and the reason for my actions. I'm tired of trying to explain to someone with a closed mind. It's people like the one you've become that turned me away from the boinc related fora. Don't bother posting a reply. I won't return any time soon. Thank you for showing me that the comments by boinc users that are anything but gushing praise are not welcome.
2) Message boards : BOINC Manager : Recent modifications to Boinc Mgr look - not liking it [have included suggestions] (Message 38384)
Posted 13 Jun 2011 by Profile tekwyzrd
Post:
Why are you so defensive? I wasn't blaming boinc for the problem. I had recently seen in the messages that a new boinc version was available but I ignored it intending to build the new version at a more convenient time. When I saw there were two more tasks that didn't complete I thought it might be due to server side updates and a new app version and an incompatibility with v6.10.58.

Contrary to what's indicated all four failed under v6.10.58. With the last two errored tasks I shut down the client and exited boinc and the next day I tried a couple unsuccessful builds (I usually compile my own with my own cflags to compensate for the win vs linux benchmark difference) and installed the premade v6.12.31 and v6.12.32. I had to create a libnotify.so.1 link to libnotify.so.4.0.0 to get v6.12.31 to run and found the messages tab gone. While running v6.12.31 I let boinc report and request new tasks. Maybe it reports this version because it was the version running when uploaded and reported or maybe boinc tried to re-run the tasks - I don't know. Then I exited boinc and I installed v6.12.32 and found the messages tab absent there too.

As for one person being the reason for the column changes... it took months for the elimination of the picflags bug that caused the unusually low benchmarks on linux computers a few years ago. I found it, fixed it, notified boinc staffers and posted on the seti forum about it but it took about eight months before it was fixed. Back then I'd download new versions and build every week and when the benchmarks were still low I'd edit the appropriate files and build again to fix it.

I haven't been posting on these fora much but that doesn't mean I'm new to boinc.

As for the malariacontrol message, look at the date. That was five years ago on a long-gone dual p3. This computer is a single processor and is watching for orbit@home work but running einstein@home. One task at a time. BOINC's disk usage is set to 10GB with 58.49GB of the 80GB /home drive free.
3) Message boards : News : BOINC 6.12.26 released to public (Message 38380)
Posted 13 Jun 2011 by Profile tekwyzrd
Post:
If you are going to hide messages because they "scare" some users and make them quit, what's next? I'm sure some people quit because their computer becomes sluggish when they're playing a game with boinc projects running and others when they see how long some tasks are expected to take. If you hide each feature or bits of information that may "scare" some potential user you will reach a point where you will lose more users who don't like the changes than you keep by making them and the ones you lose will likely be the more experienced, dedicated, and productive. It's easy to lose user trust if you hide too much.
4) Message boards : BOINC Manager : Recent modifications to Boinc Mgr look - not liking it [have included suggestions] (Message 38379)
Posted 13 Jun 2011 by Profile tekwyzrd
Post:
OS: Sabayon Linux

I have to say I'm not too impressed by what I see in v6.12.32 for linux. I realize it's not an official release but the recommended v6.10.58 is almost a year old. I had four einstien@home tasks die and thought it might be due to changes to the project implemented for a newer boinc version. Now that I have v6.12.32 running I find a notices tab with nothing to tell me and the messages tab gone with it's contents hidden in the Advanced menu. Why do so many projects decide to dumb things down and hide features instead of educating users?

EDIT: I replaced boincmgr v6.12.32 with the v6.10.58 I compiled a few months ago but I still don't understand changes like this.
5) Message boards : BOINC client : Will someone Please fix the Linux client benchmark! (Message 7225)
Posted 2 Jan 2007 by Profile tekwyzrd
Post:
Will someone Please fix the Linux client benchmark!


You could install the new 5.7.5 (alpha) Linux manager.
Benchmarking has been improved and is almost equal to a Win host.

Besides from that... it also looks much better.

;-)



Yeah, I noticed the improvement in BOINC Manager's appearance. The standard BOINC manager now looks just like the ones I've been building locally since Sept. 2005. That's the case on my computer running Sabayon 3.2. I haven't tested it on my other computer that runs SuSE 10.1 yet.
6) Message boards : BOINC client : How to manage scheduling on dual-core cpus ? (Message 7204)
Posted 1 Jan 2007 by Profile tekwyzrd
Post:
It would be nice to be able to dedicate processors on a per project basis. For months I argued the benefit of being able to run two different projects at once rather than running two of one then two of the other. The standard response I got was "there's no difference". Finally I gave up.

Though many disagree there is a difference. On one computer I observed that seti tasks took approximately 20% less time to run when running seti on one processor and einstein on the other.

Maybe some day they'll implement this feature.
I hope it happens, but I doubt it will.
7) Message boards : BOINC client : Will someone Please fix the Linux client benchmark! (Message 7179)
Posted 29 Dec 2006 by Profile tekwyzrd
Post:
I'm not sure what the discussions are but know that WCG and Dr. A. agreed, that the next formal release i.e. 5.8 would have a fixed Linux benchmark to bring the result on par with the Windows benchmark.

In quorum 3 situations like WCG one needs 3 Linux machines to have reasonable claims to arrive at a median, which is awarded to all in the quorum. "Outliers"....those overclaiming get normalised or penalised. A single machine having been 'optimized' in a quorum is thus chucked for median calculation purposes.

A quorum is never a mix of multiple OS results and they give slightly different output.

Oh, and the benchmark for Windows version is the one suspected of having been inflated in the past....Linux benchmark is the one actually close to reality!


I find it a bit disturbing that most of what users say goes ignored yet if WCG wants something it gets immediate attention. It's a bit disappointing that the benchmark disparity has been discussed many times and was always dismissed because "only 10% of the computers are running Linux". Now WCG wants it fixed and it's being addressed. Benchmark work units were suggested a long time ago and dismissed as a bad idea. Now WCG wants them and it's going to be implemented. I think it's obvious who's really in control of BOINC.
8) Message boards : BOINC Manager : Unable to open Web Browser (Message 7166)
Posted 29 Dec 2006 by Profile tekwyzrd
Post:
If you have a file named .profile in your user directory you could try adding the entry:

export BROWSER="/usr/bin/firefox"

9) Message boards : BOINC client : Will someone Please fix the Linux client benchmark! (Message 7165)
Posted 29 Dec 2006 by Profile tekwyzrd
Post:
This gets irritating if you run several different computers/OS's. If I set preferences to get enough files to keep the Linux machines running (twice the buffer it would require if the client would calculate the estimated time correctly), my windows machines get buried in tasks. I know I can "work around this" by declaring different locations and preferences for the Windows and Linux machines, but the question is-
Why should I have to??????? It seems that everyone knows that the Linux benchmark is about 1/2 of the real operating speed, is it that hard to fix?




I've been dealing with this since I switched to linux last August. The only solution I've found is to use an optimized version compiled specifically for your processor. It eliminates a good portion of the disparity between linux and windows and though some projects frown on it my computers claim credits very close to what is claimed by others running windows.
10) Message boards : Web interfaces : Request for additional features in User Detailed Information. (Message 4297)
Posted 12 May 2006 by Profile tekwyzrd
Post:
I've been working on a Firefox extension that displays a user's stats for BOINC and the individual projects in the status bar. While working on this and while testing the BOINC User Totals function I noticed a problem on my personal stats at http://boinc.netsoft-online.com/ . I'm not sure if the problem is with XtremLab or or boinc-netsoft-online. The problem is an error message on the ExtremLab and Total Credit JpGraphs. The error message reads Unknown color: grey5. Could someone look into this?


Ok, fixed the typo.


Thank you.
11) Message boards : Web interfaces : Request for additional features in User Detailed Information. (Message 4278)
Posted 9 May 2006 by Profile tekwyzrd
Post:
I've been working on a Firefox extension that displays a user's stats for BOINC and the individual projects in the status bar. While working on this and while testing the BOINC User Totals function I noticed a problem on my personal stats at http://boinc.netsoft-online.com/ . I'm not sure if the problem is with XtremLab or or boinc-netsoft-online. The problem is an error message on the ExtremLab and Total Credit JpGraphs. The error message reads Unknown color: grey5. Could someone look into this?
12) Message boards : Web interfaces : Request for additional features in User Detailed Information. (Message 4266)
Posted 9 May 2006 by Profile tekwyzrd
Post:
OK, I've found a couple working sourceforge servers.
They seem pretty unstable right now.

BOINC Stats v0.2.2.1 is available here and here.
It has been uploaded to mozilla and is awaiting approval.
Thanks to Steve Cressman for the suggestion regarding the xml tree display problem.
13) Message boards : Web interfaces : Request for additional features in User Detailed Information. (Message 4265)
Posted 9 May 2006 by Profile tekwyzrd
Post:
Post deleted.
14) Message boards : Web interfaces : Request for additional features in User Detailed Information. (Message 4257)
Posted 8 May 2006 by Profile tekwyzrd
Post:
I've released a Firefox extension known as BOINC Stats. It is awaiting approval at mozilla but v0.2.1 is available via sourceforge as are the release notes.

The BOINC User Totals are derived from the xml data from http://boinc.netsoft-online.com/ and left-clicking on the data displays the associated data in a browser tab. Unfortunately there's no style information associated with the xml file. Would it be possible to have a style implemented for this xml data?
15) Message boards : BOINC client : setting cpu usage per project (Message 3678)
Posted 28 Mar 2006 by Profile tekwyzrd
Post:
Hello !

It would be nice if it would possible to limit the usage of the available cpus per project. For example I have a host with 2 cpus (only virtual, but never mind), and per default BOIC runs two workunits at once.

That's not a problem with most projects, but f.e. the new seasonal climate models need a lot of memory and I want to prevent BOINC to run 2 of such WUs at the same time, because this is stressing my virtual memory and the other applications slows down.

Ok, I could use the preferences to limit the general cpu usage to one, but that wastes cpu cycles if other projects (like predicotr@home) with little memory requirements are running.

I hope, I explaned all well.
Regards, Frank


This has been suggested more than once but like benchmark parity between platforms and work requests proportional to resource allocation it's not a feature to expect any time soon.
16) Message boards : BOINC Manager : BOINC in Linux (Message 3541)
Posted 18 Mar 2006 by Profile tekwyzrd
Post:
I can solve that for you. In linux a series of environment variables are stored that relate to certain default attributes like HOSTNAME or LOCALE. One of these environment variables is called BROWSER and often it is not set by default.

You need to be root to do this.

edit the file /etc/profile with a convenient app like vi or nano and add this to the end,

BROWSER="<path/to/your/browser"

save and you're done. The simplest way for this to be read by BOINC is simply to reboot but you could try maybe export $BROWSER from the command line and just restart BOINC.

By way of example, I use mozilla and the mozilla binary's full path is /opt/mozilla/bin/mozilla so my extra line is

BROWSER="/opt/mozilla/bin/mozilla"

You can find the exact path using the whereis command,

whereis mozilla

You will also find the acroread package useful. In ArchLinux this is as simple as,

pacman -Sy acroread

but I'm sure SuSe/RH/FC/Debian have something similar
enjoy.


Sorry to disappoint you but this doesn't work on SuSE 10.0 with KDE.

It would make things so much easier if boinc referred to a config or preferences file in the boinc directory for this info. Just open the file with a text editor and type or paste the path to the user's preferred browser. Or, better yet, an Advanced\\Select Browser option with a file system browser window to select the web browser and set the necessary value in the config or preferences file.
17) Message boards : BOINC client : BOINC 5.3.25 - Official vs locally compiled (Message 3397)
Posted 8 Mar 2006 by Profile tekwyzrd
Post:
CORRECTION: The Results from the Berkeley version are for 5.3.24.

It would be nice if they made a corresponding source package available when they release a new version.
18) Message boards : BOINC client : BOINC 5.3.25 - Official vs locally compiled (Message 3393)
Posted 8 Mar 2006 by Profile tekwyzrd
Post:
Well, I've finally figured out why I've been having problems since December compiling from the dev cvs source. The following is a comparison of benchmarks from the official version of BOINC 5.3.25 and versions I compiled from the dev cvs sources downloaded 2006-03-07 22:53 UTC. The work was done on an extremely up-to-date copy of SuSE 10.0 using gcc 4.0.2. The computer is a rock steady dual P3 700 overclocked to 933.

Official Berkeley version

Benchmark results:
Number of CPUs: 2
466 floating point MIPS (Whetstone) per CPU
811 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

Locally compiled (un-optimized)

Number of CPUs: 2
456 floating point MIPS (Whetstone) per CPU
871 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

Locally compiled (optimized)

export CFLAGS="-march=pentium3 -O3 -malign-double -fomit-frame-pointer -funroll-loops -fforce-addr -ffast-math -ftracer"
export CXXFLAGS=$CFLAGS

Benchmark results:
Number of CPUs: 2
638 floating point MIPS (Whetstone) per CPU
1098 integer MIPS (Dhrystone) per CPU

work errored out immediately

Locally compiled (optimized)

export CFLAGS="-march=pentium3 -O3 -fomit-frame-pointer -funroll-loops -fforce-addr -ffast-math -ftracer"
export CXXFLAGS=$CFLAGS

Benchmark results:
Number of CPUs: 2
295 floating point MIPS (Whetstone) per CPU
1096 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

removal of -malign-double eliminated the work-killing error
this wasn't a problem until mid December, and work errored out on all optimized versions I compiled. I'm glad this has finally been resolved.
the repeatable low floating point score was a major surprise

Locally compiled (optimized)
(line containing @PICFLAGS@ removed from /boinc/Makefile.incl prior to ./autosetup)

export CFLAGS="-march=pentium3 -O3 -fomit-frame-pointer -funroll-loops -fforce-addr -ffast-math -ftracer"
export CXXFLAGS=$CFLAGS

Benchmark results:
Number of CPUs: 2
982 floating point MIPS (Whetstone) per CPU
1416 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

So far the last version listed seems to be doing well.
The only oddity I've noticed is the absebce of the BOINC logo in the about box but that's a minor cosmetic issue concerning the boinc manager.
19) Message boards : BOINC client : Problems running locally compiled BOINC (Message 2901)
Posted 31 Jan 2006 by Profile tekwyzrd
Post:
Looking back on it I should have added that the last BOINC dev cvs source I was able to compile and run was v5.3.3.
20) Message boards : BOINC client : Problems running locally compiled BOINC (Message 2897)
Posted 31 Jan 2006 by Profile tekwyzrd
Post:
Once again I tried to compile a new version of BOINC.
This time v5.3.16
If I install the standard version it runs without errors but benchmarks are:

Benchmark results:
Number of CPUs: 2
455 floating point MIPS (Whetstone) per CPU
813 integer MIPS (Dhrystone) per CPU
Finished CPU benchmarks

These results are WAY too low for P3 processors @933MHz and if I run it I'll be requesting approximately 9 credits for a short runtime (current estimates are 3:18:22) einstein result. Boincmgr looks like something straight out of windows 3.1

So I drop in the boincmgr I compiled. It works fine and looks great. Then I drop in the optimized boinc app (after suspending and exiting boinc and the science apps). After starting boinc and switching back to "Run based on preferences" every item in the task tab errors out.

This has happened with EVERY version after v5.2.15 I've tried to compile. In addition it seems to ignore the "Network access disabled" setting and in the messages tab indicates it is attempting to get more work after the existing work has errored out.

I'm using Suse 10.0 and gcc 4.0.2
From the build process.

during _autosetup

Bootstrapping configure script and makefiles:
Checking version of 'make' >= 3.79... succeeded. (3.80)
Checking version of 'm4' >= 1.4... succeeded. (1.4)
Checking version of 'pkg-config' >= 0.15... succeeded. (0.20)
Checking version of 'autoconf' >= 2.58... succeeded. (2.59)
Checking version of 'automake' >= 1.8... succeeded. (1.9)
aclocal -I m4 && autoheader && automake && autoconf
/usr/share/aclocal/progsreiserfs.m4:13: warning: underquoted definition of AC_CHECK_LIBREISERFS
run info '(automake)Extending aclocal'
or see http://sources.redhat.com/automake/automake.html#Extending-aclocal
/usr/share/aclocal/pilot-link.m4:1: warning: underquoted definition of AC_PILOT_LINK_HOOK
Done, now run ./configure

I'm doing everything the same as with v5.2.15

export CFLAGS="-march=pentium3 -O3 -malign-double -fomit-frame-pointer -funroll-loops -fforce-addr -ffast-math -ftracer"

export CXXFLAGS=$CFLAGS

./configure --disable-server

make clean

make

no errors during configure

during make

then mv -f ".deps/extract.Tpo" ".deps/extract.Po"; else rm -f ".deps/extract.Tpo"; exit 1; fi
./unzip/extract.c:221: warning: parameter names (without types) in function declaration

lse rm -f ".deps/boinc_gui-BOINCGUIApp.Tpo"; exit 1; fi
BOINCGUIApp.cpp:170:9: warning: extra tokens at end of #endif directive

Would someone please take a bit of time and help figure out what the problem is?

Thank you.


Next 20

Copyright © 2024 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.