Another project with low scores - CAS@HOme

  • Thread starter Deleted member 88227
  • Start date
D

Deleted member 88227

Guest
I see phoenicis is getting points on this. Care to share your success?

Running a script? Just staying attached to the project?
 

Gilthanis

[H]ard|DCer of the Year - 2014
Joined
Jan 29, 2006
Messages
8,574
Looks like only about 3 users have points today so far according to Free-DC

https://stats.free-dc.org/proj/cas

I'm betting he is running the same scripts for pulling work as SZTAKI. CAS is one of the projects I have recommended people also run from their VM's dedicated to just pulling work from the hard to get projects.
 
D

Deleted member 88227

Guest
Yeah, that's why I started this thread.

I modified a batch script. Seems to be working (no errors) but there is no work to pull right now so only time will tell if I can get anything from it.

for /l %%x in (1, 1, 10000) do (echo %%x
"c:\Program Files\Boinc\boinccmd.exe" --project http://casathome.ihep.ac.cn/ update
ping 127.0.0.1 -n 180 > nul)
 
Last edited by a moderator:

Gilthanis

[H]ard|DCer of the Year - 2014
Joined
Jan 29, 2006
Messages
8,574
Well..when CAS has work, last I had it reported to me was that there was a limit of 40 work units per host in progress at a time. So, maybe changing the core count would be wise to try and load up as well...
 

phoenicis

[H]ard|DCer of the Year 2018
Joined
Oct 3, 2010
Messages
590
I see phoenicis is getting points on this. Care to share your success?

Running a script? Just staying attached to the project?

I'm running a script with a 180s delay between update requests. It used to be mostly just me and OCUK's PDW targeting CAS but recently XS have been piling on en masse.

A few dozen work units appear daily sometime between 00.00 and 06:00 UTC. Can't be sure exactly when as that's sleep time for me.

Every few months a large batch is released which typically lasts for a day or so. I snagged about 200K credits last time this happened so best to stay attached 24/7.

Be great if others ran CAS. The more the merrier.
 
D

Deleted member 88227

Guest
Yeah, I'm gonna stay on top of these low tasks projects. XS is coming for our throats.
 

phoenicis

[H]ard|DCer of the Year 2018
Joined
Oct 3, 2010
Messages
590
Yeah, I'm gonna stay on top of these low tasks projects. XS is coming for our throats.

Xansons is another low task project I could use a hand with. It releases a few hours work every two weekends. This has recently commenced at around 19:00 to 21:00 UTC on Saturdays. It's both cpu and gpu but cpus are best used elsewhere.

After a slow start, XS have pushed themselves to 2nd place. With more people on this and enough of a lead, we could screw with their position later in the year.
 
D

Deleted member 88227

Guest
Ill make a new thread addressing that project when I get back home.
 
D

Deleted member 88227

Guest
Can confirm I am getting work using the script on this project.
 
D

Deleted member 88227

Guest
motqalden you seem to be very good at spinning up multiple VMs to get a lot of WUs from projects that don't have a lot of work. This project should be right up your alley. Granted, you don't need to go to the extent you did with Sztaki since we are in 1st by a good bit already, but the project does need to hit 2M overall points this year which it has not done yet. So there are an extra 25 points just waiting for us to claim.
 

motqalden

[H]ard|DCOTM x5
Joined
Jun 22, 2009
Messages
2,982
motqalden you seem to be very good at spinning up multiple VMs to get a lot of WUs from projects that don't have a lot of work. This project should be right up your alley. Granted, you don't need to go to the extent you did with Sztaki since we are in 1st by a good bit already, but the project does need to hit 2M overall points this year which it has not done yet. So there are an extra 25 points just waiting for us to claim.

Alright I will have a look. I did have to power down a few VMs yesterday. I started citizen's science grid and man that project eats ram
 

phoenicis

[H]ard|DCer of the Year 2018
Joined
Oct 3, 2010
Messages
590
Nice one Skillz. You start a thread and the next thing you know work appears.

I just checked my rigs and there's CAS work galore. Not sure when it started or when it will end but I've already picked up 50K credits.

Of course XS is all over it. Get the work while it's there guys!
 
D

Deleted member 88227

Guest
Nah, being the CRO and having a new found Luckz title just proves I've got dem Skillz. ;)
 

MN Scout

Supreme [H]ardness
Joined
Jan 26, 2003
Messages
4,723
It only sent me 10, and the event log is saying it is having trouble connecting to the server.
 
D

Deleted member 88227

Guest
You need to have at least 20 threads/cores to get 40 tasks. All my quad core rigs were getting 8 tasks, but once I bumped the CPU count to 48 it got 40 tasks. Then I discovered my 20 thread intel rig had 40 tasks without having to modify anything so I am pretty sure the limit is 20 threads to get 40 tasks.

Also I keep getting connection errors as well, but it eventually connects.
 

MN Scout

Supreme [H]ardness
Joined
Jan 26, 2003
Messages
4,723
Also found out it needs 953.67MB of Ram. Will need to bump up the virtual machines. Edit the tasks are only taking 2MB to 30MB right now. It is complaining for, I don't know why.

Used these lines in cc_config to get my 40 tasks.

<start_delay>60</start_delay>
<ncpus>100</ncpus>

Will change ncpus down after I hopefully get a lot of tasks.... Eek all 40 tasks are running at once! The 15 Ubuntu VMs are also running. I can't believe this poor computer is actually responsive right now.
 
Last edited:
D

Deleted member 88227

Guest
You only need to do <ncpus>20</ncpus> to get them.
 
Last edited by a moderator:

MN Scout

Supreme [H]ardness
Joined
Jan 26, 2003
Messages
4,723
Weird the Cas@Home tasks aren't starting in my Ubuntu VM. It just says "Ready to Start". I wasn't able to fix this.

Really messed up this VM instance. Won't even boot now. Maybe it has to do with me allowing it to use 2CPUs in the Virtual Box instance, and then changing it back to 1.

I think I'm going to be OK with just my main Windows box on this project now.
 
D

Deleted member 88227

Guest
Not sure what could have happened, but why did you change the number of CPUs in the VM?
 

Gilthanis

[H]ard|DCer of the Year - 2014
Joined
Jan 29, 2006
Messages
8,574
To correct again above, I believe they limit it to 2 work units in progress per thread up to a total of 40 per host in progress. So, you would only need to set your NCPU's to 20 to get the 2 per thread limit. Please correct me if I'm wrong.
 
D

Deleted member 88227

Guest
Okay guys. I set my quad core CPUs to NCPUS=20 and then switched back and forth every so often on percent of CPU so it would keep 40 tasks, but only run 4. Then I tried using an app_config to limit tasks running to 4. This allows you to keep the CPU percent at 100%, number of CPUs at 20 and the project will download the 40 limit tasks. This is my app_config.xml file for it.

<app_config>
<app>
<name>treeThreader</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>


Change max_concurrent to actual number of processes you want to run which should be equal to or less than the number of cores/threads you have.
 

Gilthanis

[H]ard|DCer of the Year - 2014
Joined
Jan 29, 2006
Messages
8,574
According to the server status, over a thousand users joined today....


CAS.PNG
 

Gilthanis

[H]ard|DCer of the Year - 2014
Joined
Jan 29, 2006
Messages
8,574
possibility considering only 63 computers registered in the last 24 hours
 
D

Deleted member 88227

Guest
Yeah and the number of recent credit also which includes all members not just newly registered or even anyone who got credit in the past month probably
 

pututu

[H]ard DC'er of the Year 2021
Joined
Dec 27, 2015
Messages
2,574
Yeah, I think it is the spammers. Lots of them. Some are fun to read. See this new CAS spam just joined today: http://casathome.ihep.ac.cn/team_display.php?teamid=9122, particularly the "Description" section.

Not all Rastafari are homophobic, but TIME Magazine referred to Jamaica as the most homophobic place on Earth” because of the dozens of murders and a huge selection of attacks on gays on the island each year. Furthermore, successful multi-generation hybridisation with ruderalis has spawned the 4th generation of Super Autoflowering cannabis.
She is typically a small herb (around 80 cm), though some growers have reported higher levels when growing outside Don't allow her small size strategy you into pondering you can hide her away everywhere- Afghan Kush Ryder has a robust skunky odor that gives her away every time.
:meh::wtf:

Scratching my head.
 

motqalden

[H]ard|DCOTM x5
Joined
Jun 22, 2009
Messages
2,982
alright i am onboard now. not all in yet, but onboard :D currently 40 some ish in progress
 

Gilthanis

[H]ard|DCer of the Year - 2014
Joined
Jan 29, 2006
Messages
8,574
Kind of a larger batch really...normally goes quicker than that....
 

atp1916

[H]ard|DCoTM x1
Joined
Jun 18, 2004
Messages
4,850
So i took the script for motqalden and dumped in CAS / WEP. It now pins Sztaki / WEP / CAS. No work from CAS/WEP after a day of pinging away.

Code:
for /l %%x in (1, 1, 100000) do (echo %%x
"c:\Program Files\Boinc\boinccmd.exe" --project http://casathome.ihep.ac.cn/ update
ping 127.0.0.1 -n 140 > nul

echo %%x
"c:\Program Files\Boinc\boinccmd.exe" --project http://szdg.lpds.sztaki.hu/szdg/ update
ping 127.0.0.1 -n 140 > nul

echo %%x
"c:\Program Files\Boinc\boinccmd.exe" --project http://bearnol.is-a-geek.com/wanless2/ update
ping 127.0.0.1 -n 140 > nul
)

Also: McCHillin the QDBW is going out tomorrow!
 
D

Deleted member 88227

Guest
WEP doesn't need a script. It has too much work.
 

MN Scout

Supreme [H]ardness
Joined
Jan 26, 2003
Messages
4,723
I tried to use the Docker instance https://hub.docker.com/r/boinc/client/ . I messed up in some way because the Docker instance would never have any used cpu seconds when I ran boinccmd --get_tasks. I'm guessing part of the problem was running a swarm of 50 BOINC dockers inside of a AlpineLinux virtual machine that was not made with enough RAM. Tonight I'm going to try to run the swarm inside of Windows.

I did get it to attach to the project under my name. Just no cpu seconds even though it was Executing, and no CPU % viewed in Windows task manager.

I'm also not sure if there is any way to persist the docker data, it looks like a docker command can be called to write the data to a TAR/Zip file before closing.
 
Last edited:
Top