Posts by Carl Christensen |
| log in |
|
1)
Questions and Answers :
Unix/Linux :
Phidgets.com sensors
(Message 4673)
Posted 2 hours ago by Carl Christensen a cheap phidget is fine. they are a little tricky under Linux ie you have to run BOINC/QCN as "root" due to the way they use the Linux USB driver. |
|
2)
Questions and Answers :
Windows :
Since month only errors ?!
(Message 4672)
Posted 2 hours ago by Carl Christensen it looks like you're getting too many timing errors. it could be a Windows patch was applied that for some reason is messing up QCN. also if you run other BOINC projects that keep your cmoputer at 100% utilization, that can cause a lot of resets on QCN -- even though we are "non-CPU intensive" high CPU usage does seem to slow down the USB bus so that the sensors get reset a lot. |
|
3)
Message boards :
News :
NEES/QCN Open House for Future Teachers
(Message 4630)
Posted 26 days ago by Carl Christensen NEES@UCSB personnel hosted an open house for future science teachers in the Givertz Graduate School of Education at UC Santa Barbara in April 2013. Sandra Seale and Francesco Civilini, NEES project specialists, gave a training session in “Make Your Own Earthquake” using the QCN sensor. The teachers downloaded the QCN software and practiced making their own quakes and printing their graphs. Find out more information about NEES and Make-Your-Own-Earthquake here. |
|
4)
Questions and Answers :
Unix/Linux :
MK802
(Message 4623)
Posted 32 days ago by Carl Christensen if you're getting that error and boinccmd is in the present directory, then your machine is probably a different architecture --- try running "uname -a" and post the output (ie you may be an ARM v6 not a v5) u |
|
5)
Questions and Answers :
Unix/Linux :
MK802
(Message 4620)
Posted 33 days ago by Carl Christensen it looks like you don't have a /home/qcn directory, which is usually what you'd have if you made a user account called "qcn", and logged in (or typically ran "sudo su - qcn") as that account. so your first error message is that "/home/qcn" directory isn't found, then your boinc-armv5.zip isn't there to be unzipped etc. |
|
6)
Questions and Answers :
Unix/Linux :
QCN and BOINC for Raspberry Pi (Debian "Wheezy" Linux)
(Message 4602)
Posted 40 days ago by Carl Christensen thanks, I'll look into this, I'm not sure it can be set by a boinc program though? but it may provide a workaround |
|
7)
Questions and Answers :
Windows :
Ensuring my participation - what to look for
(Message 4601)
Posted 40 days ago by Carl Christensen http://qcn.stanford.edu/sensor/show_host_detail.php?hostid=26966 you do seem to have a higher than usual (>100) number of resets, I guess you're running other BOINC projects and keeping this computer at 100% CPU usage? That can affect the sensor monitoring - even though we're a "low CPU" project - if the CPU is pegged at 100% it seems to slow down the USB data transfer accuracy. |
|
8)
Questions and Answers :
Windows :
Trouble updating project
(Message 4593)
Posted 43 days ago by Carl Christensen is it my QCN program or the BOINC client? I think perhaps BOINC has been trying to "take over" the completion estimates since many projects have a hard time estimating the gigaflops for a workunit. And with QCN there's no real CPU component ie our workunits & credits are based on "clock time" -- i.e. a workunit, which will give you 50 cobblestones of credit, is 24 hours of "real time" sensor monitoring. I know in the older days the BOINC client would just accede to my "completion done" calls so you would basically get a 24-hour "countdown" -- but nowadays this does not happen? |
|
9)
Questions and Answers :
Windows :
No close tag in scheduler reply...
(Message 4586)
Posted 47 days ago by Carl Christensen hmm that's odd, I just checked my sched_reply.xml and it looked fine - and it seems people are getting workunits & triggering etc. Is BOINCTasks a pretty old version, maybe there are new XML fields from BOINC (I updated the server fairly recently ie the new year) and BOINCTasks isn't handling them? |
|
10)
Questions and Answers :
Unix/Linux :
QCN and BOINC for Raspberry Pi (Debian "Wheezy" Linux)
(Message 4581)
Posted 48 days ago by Carl Christensen I meant a newer RPi - the sensors are just USB and limited by the USB implementation of the hardware and operating system. Unless someone makes an accelerometer that can bypass the USB ie by mounting on the RPi expansion pins directly, any USB sensor will run into this reset problem with current RPi's. |
|
11)
Message boards :
News :
BBC (UK) TV - Looking for videos of participants - what do you think of QCN?
(Message 4575)
Posted 57 days ago by Carl Christensen Dear all, We are working on a piece for British television for a science programme called the Dara O’Briain Science Club that will be broadcast in the UK to several million and many more around the world. We are working with the project leader Elizabeth Cochran to make a short insert for a prime time science show explaining the QCN. Elizabeth has kindly given me permission to contact the community via this forum. We are trying to demonstrate the success of the network so far – show how this is really just at the beginning but already people from all over the world are taking part. We are looking for participants to film themselves to tell us why they got involved with the network and what it means to them. It could be you, your partner, your children – anybody taking part! It could be filmed on an iPhone, webcam or similar – the quality isn’t important. We’d hope you’d be able to upload the video to a private YouTube channel on use drop box – a free upload service: https://www.dropbox.com/home Ideally what we are looking for is: • Your first name – and why you joined the quake Cather network – (about 30 seconds) • What are your hopes for the network? - (20 seconds roughly) • Wave at camera with name on card or similar (licence to be creative here!) - (about 20 seconds). Please contact me directly on duncan.singh@bbc.co.uk if you have any questions and to submit short videos. We aim to make a montage with the footage for use in the programme – to demonstrate the scale and success of the network. Thank you for your help! Duncan |
|
12)
Questions and Answers :
Windows :
App crashes at Win 7 x64 / CPU: Xeon E3-1230V2
(Message 4573)
Posted 60 days ago by Carl Christensen this computer seems to be running OK: http://qcn.stanford.edu/sensor/show_host_detail.php?hostid=11796 the other one seems to have stopped a few weeks ago and no new tasks, so I'm assuming you are interested in host 11796. a common problem seems to be if running other boinc apps 100%, even though qcn is a "low cpu app" it seems to take away from the USB transfers and cause a lot of resets (i.e. 100-1000). If you just have a few resets ie < 30 then it's working OK. |
|
13)
Message boards :
News :
New Versions of QCN for All Platforms (v 7.29)
(Message 4568)
Posted 62 days ago by Carl Christensen 7.31 is a very minor change -- we are changing the reporting of the -1/1/2/4 second data we send in the followup trigger/trickle after an event (I used to take the maximum of the whole range, now it's the max of each sub-range ie between 2-4 sec so we can see if the signal is ramping up or down). |
|
14)
Questions and Answers :
Windows :
Seismic data to webpage ?
(Message 4566)
Posted 63 days ago by Carl Christensen to save participant's disk space, we only write out triggers, which are in the boinc/projects/qcn.stanford.edu_sensor/trigger subdiretory in zip files. You can probably unzip each one somewhere else and then process the .sac files to generate a chart locally if you want to plot the waveforms. |
|
15)
Message boards :
News :
QCN Featured in TEDx Talk
(Message 4561)
Posted 76 days ago by Carl Christensen Project collaborator Debi Kilb gave a TEDx talk on ‘Using Videogames to Cultivate Future Scientists’ which includes a videogame created around the QCN project. Check out the video of her talk here. |
|
16)
Questions and Answers :
Unix/Linux :
QCN and BOINC for Raspberry Pi (Debian "Wheezy" Linux)
(Message 4560)
Posted 77 days ago by Carl Christensen I wouldn't spend too much time with the current state of the RPi & QCN - I think with the massive reset problems with the USB bus we'd have to wait for a newer unit. The disk space used should be small something like 100MB or so. BOINC is conservative about disk space & memory usage so if any of the boinc "rules" are violated (i.e. probably 25% of disk space) it won't run an app. |
|
17)
Questions and Answers :
Unix/Linux :
QCN and BOINC for Raspberry Pi (Debian "Wheezy" Linux)
(Message 4555)
Posted 81 days ago by Carl Christensen I guess it could get that high CPU usage since the RaspPi CPU isn't so strong, and my RPi Boinc runs QCN at the highest priority. |
|
18)
Questions and Answers :
Macintosh :
lots of errors
(Message 4544)
Posted 88 days ago by Carl Christensen yeah I upgraded my Mac & Xcode so I think you need 10.6 to run QCN now. |
|
19)
Questions and Answers :
Unix/Linux :
QCN and BOINC for Raspberry Pi (Debian "Wheezy" Linux)
(Message 4537)
Posted 91 days ago by Carl Christensen you'll want to run "./boinccmd" executable, not "./boinc" e.g.: ./boinccmd --project_attach http://qcn.stanford.edu/sensor ACCOUNT_KEY you get your ACCOUNT_KEY from your page here (must be logged in): http://qcn.stanford.edu/sensor/weak_auth.php |
|
20)
Questions and Answers :
Windows :
App crashes at Win 7 x64 / CPU: Xeon E3-1230V2
(Message 4533)
Posted 94 days ago by Carl Christensen the problem is actually you are having too many resets, ie too many "time slips" by the sensor (usually over a half-second between points, when it should be .02 seconds). This can happen if you're running a lot of other BOINC apps so your CPU is at 100% -- even though QCN with a USB sensor will run at .1% or .01% CPU, it seems that it slows down the USB bus enough so that it can't sustain our 50Hz monitoring rate (i.e. a reading every .02 seconds). this is the error message from your stderr.txt: qcn_util::getBOINCInitData requested at 1361014284.732985 Timing error encountered t0check=1361015607.931722 t0active=1361015608.299227 diff=0.000000 timeadj=0 sample_size=0, dt=0.020000, lErr=1, lErrCum=1001 resetting... Last values were 0.115510 -0.344143 -10.085820 1361015608.299227 |