Welcome to MilkyWay@home

Posts by rbpeake

1) Message boards : Number crunching : How many CPUs? (Message 69250)
Posted 12 Nov 2019 by rbpeake
Post:
Thanks!
2) Message boards : Number crunching : How many CPUs? (Message 69246)
Posted 11 Nov 2019 by rbpeake
Post:
Is there an app_config.xml configuration to limit the number of cores the multi core CPU application uses? I tried adjusting the Use Only x% of the CPU's, but that does not work.

Thanks!


Yes this line does that:
<avg_ncpus>4</avg_ncpus>
In this case it's limiting the program to only use 4 cpu cores per workunit.


Would you happen to know the entire app_config.xml statement/expression? I am not very familiar with these files.

Thanks!
3) Message boards : Number crunching : How many CPUs? (Message 69237)
Posted 8 Nov 2019 by rbpeake
Post:
Is there an app_config.xml configuration to limit the number of cores the multi core CPU application uses? I tried adjusting the Use Only x% of the CPU's, but that does not work.

Thanks!
4) Message boards : Number crunching : Long crunch time on new N-Body simulations? (Message 69232)
Posted 6 Nov 2019 by rbpeake
Post:
Is there a difference in runtimes between the work units? A recent 4-core unit took longer than a 3-core unit!
5) Message boards : MilkyWay@home Science : Absolute magnitude distribution (Message 963)
Posted 8 Dec 2007 by rbpeake
Post:
Thanks for the update and letting us know where we are headed.

Before I forget, will there be a break or downtime between the switch in WUs? If there is can someone post when it may be so that we do not have our machines running on empty?

Also, how will the work units be designated, so we can see which are the longer ones?

Thanks, it is nice to see the research advancing! :)
6) Message boards : Number crunching : Started Particle Swarm Searches (Message 910)
Posted 3 Dec 2007 by rbpeake
Post:
Both genetic search and particle swarm are monte carlo algorithms (basically global search methods using randomization: http://en.wikipedia.org/wiki/Monte_Carlo_method ).

Wikipedia isn't bad for information on either of these:
http://en.wikipedia.org/wiki/Particle_swarm_optimization
http://en.wikipedia.org/wiki/Genetic_algorithm

Theres a lot of interesting stuff to look into :)


Thanks very much for the very informative update! :)

For me BOINC has been a great way to learn about all these fascinating topics! A number of other BOINC projects utilize Monte Carlo optimizations as well as I am sure you are aware, and it is fascinating to see how these methods are used and improved on by the various projects each according to their needs. Thanks again, I am very much looking forward to helping you guys! :)
7) Message boards : Number crunching : Computational Errors (Message 400)
Posted 11 Nov 2007 by rbpeake
Post:
So the errors are just g5 WU's since I just got one g5 right now, probably won't crunch until after all the g4's I have.


Ni

I know from other projects that there is a way for the project administrator to delete them from the system so they do not "recycle" again and again, but I do not believe they have looked into that here yet.
8) Message boards : MilkyWay@home Science : Graduate Student Paper (Message 213)
Posted 8 Nov 2007 by rbpeake
Post:
Just curious if the paper the grad student will write based on results from this project, is for a dissertation, for publishing, or for some other purpose?

Thanks!
9) Message boards : Number crunching : New version (Message 146)
Posted 9 Oct 2007 by rbpeake
Post:
..It will download automatically, and the disk usages will be about 8 MB.

Edit: I've also reduced the memory usage by about half. So, it should run with only about 17,000 k of memory now.

Very nice! :)

Any idea from the Astronomy folks on when they will have work units? Thanks!
10) Message boards : Cafe MilkyWay : Any Requests? (Message 130)
Posted 6 Oct 2007 by rbpeake
Post:
..Once we have the windows app working and the genetic search in place, I'll try to convince my professor to let me work on the graphics.

Sounds like a good plan! :) Any word yet on when the Windows app might be fixed?

Thanks! :)
11) Message boards : Number crunching : News RSS broken (Message 108)
Posted 4 Oct 2007 by rbpeake
Post:
Works for me know! using opera's built in feed reader

Thanks!
Paul

Works fine with Firefox's built-in feed reader, too! :)
12) Message boards : Number crunching : Getting work? (Message 87)
Posted 4 Oct 2007 by rbpeake
Post:
I haven't gotten too deep into debugging this, but hopefully I can have it fixed for you soon!

You certainly are a very busy guy tonight, and a very productive one! :D
13) Message boards : Number crunching : Unable to attach a machine (Message 83)
Posted 4 Oct 2007 by rbpeake
Post:
..Apparently the windows version thinks there is a memory leak, and its outputting all the modules it's dumping to stderr.txt... I'm working on a fix for that too ;D

This result for which I ended the process would certainly seem to confirm that! ;)
14) Message boards : Number crunching : Unable to attach a machine (Message 82)
Posted 4 Oct 2007 by rbpeake
Post:
Apparently the windows version thinks there is a memory leak, and its outputting all the modules it's dumping to stderr.txt... I'm working on a fix for that too ;D

Already over 45MB... how big a file can I expect in the end?

It might be endless! ;)
15) Message boards : Number crunching : Unable to attach a machine (Message 78)
Posted 4 Oct 2007 by rbpeake
Post:
HAHA yeah, it took me a while, but it should be fixed now =D

Yup, I've got one result and am crunching it as I type this. :-)

Now everyone attached to just http://milkyway.cs.rpi.edu should detach and reattach to http://milkyway.cs.rpi.edu/milkyway

Keep an eye out on the runtime if you are running Windows!

I seem to go over 100% complete with the 1.05 Windows application, but this time I will let it run a bit longer to see if it does end somewhere! ;)
16) Message boards : Number crunching : How Long Should They Run? (Message 46)
Posted 3 Oct 2007 by rbpeake
Post:
I got this work unit, and see that you processed it in under two minutes.

I am running Windows XP and the progress bar went to 100% after a few minutes but the work unit kept on processing! It is now up to 18 minutes! It is like a runaway train I guess! ;)

I will abort it after 20 minutes or so, so you can take a look at it. Maybe it is an issue with the Windows app.
17) Message boards : Number crunching : Unable to attach a machine (Message 43)
Posted 3 Oct 2007 by rbpeake
Post:
I am attached to my Win 5.10.20 client but if I try to get some of these rare work units the following message occurs:
03.10.2007 20:57:11|Milkyway@home|Deferring communication for 1 days 0 hr 0 min 0 sec
03.10.2007 20:57:11|Milkyway@home|Reason: 6 consecutive failures fetching scheduler list


Hmmm.... ;-)

I get the same message! The scheduler does not want to give us work! ;)
18) Message boards : Number crunching : time (Message 9)
Posted 2 Oct 2007 by rbpeake
Post:
10/2/2007 2:46:22 PM|Milkyway@home|Resuming task test_wu_28_12 using astronomy version 104
the work unit is stuck
task manager shows know cpu usage
I stop and started the wu three or four times
what to do?

Since the 1.04 application apparently has a bug in it, imho it is probably best if you just abort the unit and wait until a fixed appl. is available.
19) Message boards : Number crunching : Add "Server Status" (Message 7)
Posted 2 Oct 2007 by rbpeake
Post:
At some point you may wish to add a "Server Status" link on the main page, so that we can see if work units are available and if things are running smoothly.

Thanks!




©2020 Astroinformatics Group