Trying BOINC, Confused

Linden

[H]ard|Gawd
Joined
Sep 8, 2005
Messages
1,199
I restarted one of my 4Ps today and am now trying out BOINC - Tears's optimized version. I'm using the BOINC manager to manage the client and BOINCstats to manage projects. I have zero experience with BOINC, so please bear with me if my questions might seem silly. (Oh, yes, for projects that are associated with team HardOCP, I will assign sign on.)

I have the BOINC Manager configured to get project coordination from BOINCstats/BAM!. Should I continue this or instead have the BOINC Manage go directly to the projects?

I have signed up for five projects at BAM!, only one, Rosetta@Home, shows when I synchronize the BOINC manager. I set all projects for 100% resource share. My Projects shows all projects attached to the correct host computer.

There appears to be no crunching going on at all. System Monitor shows 0% CPU activity for boincmgr. No other processes show any CPU activity either.

In the BOINC manager, it shoes Rosetta as active, but for over an hour, the status is "Downloading work from the server."

Is there a tutorial that I might have missed? It seems like with five projects selected, I should be able to see more than one in BOINC manager. Do I just need to be patient. I don't think it's supposed to be this clunky and inactive.
 
Also, in BAM!, when I try to create an account with World Community Grid, all get for a return is "No response from server." A Google search showed that to be rather common, but it seems there are many WCG contributors here. Is there a workaround?

Update: I shutdown the computer and restarted it. When I opened BOINC Manager, it showed Rosetta again plus one of the other projects I had selected at BAM! Both projects' status is "Downloading work from server." How long does it take! Rosetta's been in that status over an hour. On the same computer, networking and Internet seem to working just fine.

Is this relevant? http://boinc.berkeley.edu/dev/forum_thread.php?id=7632
 
Last edited:
Since you are just testing the BOINC waters, I would avoid managers and just use the BOINC client for now. Try out a few projects, but only run a couple at a time to narrow the range of issues until you are more familiar.
 
OK... I will see if I can help. This certainly sounds like something that can be a turn off for a first timer and no is not the norm. So, consider yourself lucky (meaning bad luck...lol).

No, it should not take a long time for Rosetta to download. Something is certainly going on. In BOINC Manager, go to the Advanced view. We are wanting to know what the Event Log says. You can find it under the Advanced option. I recommend doing this from a fresh start up of BOINC Manager. Copy/Paste like the first 30 or so lines. This will give us an idea what is happening and some of the settings you have set up.

For WCG, just make an account from WCG's website http://www.worldcommunitygrid.org/reg/viewRegister.do?recruiterId=338542 and then go to here http://www.worldcommunitygrid.org/team/viewTeamInfo.do?teamId=BP5XNJBR9N1 to join the team. Then just tell Boincstats to find your account. That will be easier.
 
Thanks for your assistance.

I would avoid managers and just use the BOINC client for now. Try out a few projects, but only run a couple at a time to narrow the range of issues until you are more familiar.
1) I am only running a couple, as for some reason, only two of the five projects I requested at BAM1 synchronized/appeared in BOINC Manager; 2) at this time, I can't revert to direct program selection - I've got to use BAM1, because the Linux client won't let me change. (Maybe it will, but I don't know how to do it.)

In BOINC Manager, go to the Advanced view. We are wanting to know what the Event Log says. You can find it under the Advanced option. I recommend doing this from a fresh start up of BOINC Manager. Copy/Paste like the first 30 or so lines. This will give us an idea what is happening and some of the settings you have set up.

For WCG, just make an account from WCG's website http://www.worldcommunitygrid.org/re...uiterId=338542 and then go to here http://www.worldcommunitygrid.org/te...Id=BP5XNJBR9N1 to join the team. Then just tell Boincstats to find your account. That will be easier.

1) In the Linux BOINC Manager, there is no advanced view. I really appreciate your willingness to help me, but I don't know where the logs are in the Linux files. 2) I already tried to tell BAM! to find my WCG account, which I did open at the WCG site. I'll try again.
 
In the Linux BOINC Manager, there is no advanced view

I don't have my linux box running right now, but from memory you need to hover your cursor at the top of the linux BOINC manager for those options to appear. In windows they are in the same place, but always visible.
 
There are two projects, the only ones showing in BOINC Manger, which no longer show their status as downloading. Both projects show their status as, "Suspended - computer in use." Huh? Well of course it's in use, I've got BOINC running! I checked CPU temps, and they are idle temps. System Monitor shows no CPU activity at all. Well, at least we've made progress, that the projects are no longer indicating downloading status.
 
Linden, take a look at your Preferences. However, first we need to get you into Advanced View. Look at my post that I linked to above. You will find pictures of the process to get there. Once you are in advanced view you can change your preferences. The reason you are seeing it suspended isn't because BOINC is active but because your mouse and keyboard are. BOINC by default is set to pause when the system is "in use" by the user.... You can change that behavior under the preferences.
 
My guess is that the reason you saw Rosetta not downloading earlier is because network activity was suspended due to the computer being in use by you moving the mouse or using the keyboard. By default, not only will processing be suspending but so will network usage... I think we are figuring out the problem already....
 
In the Linux BOINC Manager, there is no advanced view.

I remember thinking the same thing and it irritated me to no end for a while. This is due to, as KMac said, the menu hiding. The screen shots below are from Ubuntu 12 or 13; hopefully they'll help.

1) With BOINC open and the window active, your menu bar should look like this
B1.jpg


2) Move your mouse into the menu bar. File, View, Tools and Help will appear
B2.jpg


3) Click on VIEW > Advanced View...
B3.jpg


4) The Advanced View will now open. :) If you move your mouse back into the menu bar, all the Advanced View menu items will appear too.
B4.jpg


Hope this helps.
 
Wow, that's a lot of replies, and fast, too. I'll get back with you when I know more, but as of now, this is a big do-nothing, no-go.

Thanks, I'll be back. :)
 
Wow, that's a lot of replies, and fast, too. I'll get back with you when I know more, but as of now, this is a big do-nothing, no-go.

Thanks, I'll be back. :)

Where are we again? Oh yeah... [H] ;)
 
Now that I know how to get to BOINC Manager's advanced view and into Computer Preferences, I was able to get it running. Unfortunately, I guess the settings were too aggressive, and when I attempted to go back and change settings, the desktop was practically frozen. I'll work on it tomorrow some more and let you know how it's going.

Gilthanis, I had seen your guide before but had not scrolled down far enough to see the Linux instructions. I had thought it was just your walk through guide was just for Windows.
 
Last edited:
Keep in mind that if you leave the box unchecked that is for toggling run while in use, the default time it sits idle is 3 minutes before it will kick back on. If you want BOINC to pause longer, you can adjust that. Otherwise, tell BOINC to run while in use and let it go full bore. There is another override in the options up top under Activity. If you change your settings there, it bipasses the other preferences, so be careful using it....
 
Linden, I'm running BOINC on my 4P (Ubuntu 12.04 LTS) for Crunching and if it helps, here are the screenshots from my Preferences settings. They may not be totally optimized, but are working for me. Also, with WCG the team name is "HardOCP" as you will see from my screenshots. But with many other Crunching projects (ala PrimeGrid, etc) the name has to be "[H]ard|OCP" to join this team.

On the "Network Usage" screen-shot below, you will see that I have (3) days set for Min/Max Work Buffers. That is by choice and you can set those timeframes to however much additional work that you want to queue-up. I have them set for (3) days right now because I want to have a lot of work queued for the new Ebola project at WCG. If you set those settings to 0.0, then you will not queue-up more WUs than your current processors count can work.

Preferences-Processor-Usage_zps56b47e79.jpg


Preferences-Network-usage_zps97d170d7.jpg


Preferences-Disk-amp-Memory-usage_zpse2024b48.jpg


Preferences-Exclusive-Applications_zps01e20acd.jpg
 
Texinga, thanks for the graphics. They will be useful.Yesterday I did make some changes in preferences; then the computer became almost unusable as everything was almost locked in place. Not sure which setting i had made too aggressive. It will be later today when I can get to this again. I'll post and let you guys know how it's going. Gilthanis, I'll check that setting to mentioned also.
 
Linden, be careful of work units that are IO intensive as that will be a huge headache if you run multiples. Especially on multi socket rigs that can do 40+ work units at a time. I know that WCG has one app that is very IO heavy which is CEP2 and the UGM work units previously were. I don't know if they got the UGM work units straightened out yet, but that can make a rig like yours almost unusable if you are using spinners. RAM is a big deal too as work units that are hungry will eat that up quick. Rosetta work units want 512MB roughly each. CEP2 can ask for up to 1GB. I'm not aware of the other projects you attached, so let me know of them and I will give some advice in regards to those as well.
 
Success, it seems. I entered the command to see processing instances (
ps -eopid=,pcpu=,psr=,comm= | awk '{ if ($2 > 10) print }' | sort -n -k3,3) and it showed 42 cores running a "minirosetta" process. I assume that will increase to 48 cores when I'm not using the computer? tcp temp showed all cores to be running at the same temps they would if Folding Bigadv were engaged. So it looks like we have been successful. Oh yes, I have assigned the Rosetta project to credit results to HardOCP!

Alright, here's another issue, quite a bit different. I would like to crunch WCG, but I can't assign the project in BAM!. In BAM, at "Sign-up for projects," when I click on "Create account" (WCG) it returns with "no response from project." I went to WCG's site and registered there. My BAM/BOINC user name is Linden, the same as my H forums name. Unfortunately, the user name "Linden" was already taken at WCG. Additionally, my BAM password and WCG site password are different. In BAM, Sign-up for projects, WCG, if I click on "Find account," the return is "Incorrect password." Anyone know a workaround?


Well anyway, it's time to go and crank up more 48-core workstations, assignments for [H]ardOCP and HardOCP where available.:cool:
 
Well... if you used the same email address... it should still tie everything together via BAM!. However, the question is to just get your CPID to allign properly so that they know it is you. I would advise to manually connect to WCG on a machine that is running all other projects. This should sync the CPID with all projects. Once that is done, BAM! should be able to find you ok. Let's start there and then progress if we need.

As for you only seeing 42 cores being used.... How much RAM is in this system? You may have idle cores due to "waiting for memory". BOINC wont over commit, so if you limited the amount of RAM available to BOINC in your preferences like Texinga did in the second to last pic, you may be limiting the amount of work you can run at any given time. Remember that Rosetta will want ~ 512MB each...

Edit: as an example... If you set BOINC to use only 50% of RAM while in use like texinga had in the second to last pic, then a 32GB RAM rig will only give BOINC 16GB to use while you are active even if you told it to run always. Then when it goes idle again, it will bump back up to the idle setting (in his case 75% or 24GB).
 
Last edited:
Oh, thanks for the tips on RAM usage settings. I'm going to change that now, for two machines (yes, two machines crunching Rosetta for HardOCP). I'll try the WCG advice later, after ALL systems are heating the house (crunching).
 
FWIW I can confirm that doing too many CEP2 WUs at once isn't a good idea. Threw one of my big boxes at it and... well... it wasn't pretty. :p Think the safe limit might be around 10-15 WUs. Trying to do 30-40 at once? Not so much.
 
Three out of four systems are very productive now. I'm having problems with one system, the last one in my signature. For the life of me, I can't get more than four cores, to crunch simultaneously. I've checked settings over and over, but can't find anything wrong.

I'd like to completely remove all BOINC files, Manager, client, libraries - everything; to follow with reinstalling, but I don't know how. The system is running Ubuntu 12.04 LTS.
 
FWIW I can confirm that doing too many CEP2 WUs at once isn't a good idea. Threw one of my big boxes at it and... well... it wasn't pretty. :p Think the safe limit might be around 10-15 WUs.
I'm only three days into BOINC now, so my BOINC vocabulary isn't strong. "CEP2" work units? Each computer is only running Rosetta presently. Even with 40 cores occupied, the boxes currently crunching well are not overheating at all. They are quite well ventilated. Maybe I didn't understand you.

And while I've got this reply box open, does anyone know if Tear's optimized BOINC downloads work well Ubuntu 14.04 LTS? If I can't get my number 4 box crunching well, I may format the drive and start over.
 
I can't vouch for tear's optimized version as I don't run it.

CEP2 stands for Clean Energy Project Phase 2 which is a sub project at WCG. fastgeek may also want to clarify the hard disk he was running that many on as every hard drive is a bit different and some will not like more than 1 or 2. SSD's certainly fair better for IO projects.

On your fourth system, could you copy and paste the before mentioned Event Logs from your client starting from a clean BOINC Manager start? About the first 30 or so lines would suffice... You can get to it from the Advanced View, click Advanced, click Event log...
 
Yes, I'll do that. I just selected project reset on that computer a minute ago, so I might have to wait a bit.
 
After the reset, BOINC manager showed "No work available to process." So, the computer was only working a four Rosetta tasks/four computer cores because the Rosetta serve 'ran dry'?

Also, in an earlier attempt to troubleshoot, I went into the BIOS and noticed that the computer date and time were way off. I don't know if that's relevant or not, because I corrected the date and time, restarted, but the task count did not change.

Anyway, here's the log:
Sat 14 Feb 2015 07:30:04 AM AKST | | Starting BOINC client version 7.0.65 for x86_64-pc-linux-gnu
Sat 14 Feb 2015 07:30:04 AM AKST | | log flags: file_xfer, sched_ops, task
Sat 14 Feb 2015 07:30:04 AM AKST | | Libraries: libcurl/7.22.0 OpenSSL/1.0.1 zlib/1.2.3.4 libidn/1.23 librtmp/2.3
Sat 14 Feb 2015 07:30:04 AM AKST | | Data directory: /var/lib/boinc-client
Sat 14 Feb 2015 07:30:04 AM AKST | | Processor: 48 AuthenticAMD AMD Opteron(tm) Processor 6174 [Family 16 Model 9 Stepping 1]
Sat 14 Feb 2015 07:30:04 AM AKST | | Processor features: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid amd_dcm pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt nodeid_msr hw_pstate npt lbrv svm_lock nrip_save pausefilter
Sat 14 Feb 2015 07:30:04 AM AKST | | OS: Linux: 3.8.0-44-generic
Sat 14 Feb 2015 07:30:04 AM AKST | | Memory: 31.42 GB physical, 32.00 GB virtual
Sat 14 Feb 2015 07:30:04 AM AKST | | Disk: 41.71 GB total, 33.59 GB free
Sat 14 Feb 2015 07:30:04 AM AKST | | Local time is UTC -9 hours
Sat 14 Feb 2015 07:30:04 AM AKST | | No usable GPUs found
Sat 14 Feb 2015 07:30:04 AM AKST | | [error] bad value 1423963424.509297 of time stats last update; ignoring
Sat 14 Feb 2015 07:30:04 AM AKST | | Config: GUI RPCs allowed from:
Sat 14 Feb 2015 07:30:04 AM AKST | rosetta@home | URL http://boinc.bakerlab.org/rosetta/; Computer ID 2234972; resource share 100
Sat 14 Feb 2015 07:30:04 AM AKST | | General prefs: from http://bam.boincstats.com/ (last modified 13-Feb-2015 14:04:50)
Sat 14 Feb 2015 07:30:04 AM AKST | | Host location: none
Sat 14 Feb 2015 07:30:04 AM AKST | | General prefs: using your defaults
Sat 14 Feb 2015 07:30:04 AM AKST | | Reading preferences override file
Sat 14 Feb 2015 07:30:04 AM AKST | | Preferences:
Sat 14 Feb 2015 07:30:04 AM AKST | | max memory usage when active: 24128.44MB
Sat 14 Feb 2015 07:30:04 AM AKST | | max memory usage when idle: 32171.26MB
Sat 14 Feb 2015 07:30:04 AM AKST | | max disk usage: 0.51GB
Sat 14 Feb 2015 07:30:04 AM AKST | | don't use GPU while active
Sat 14 Feb 2015 07:30:04 AM AKST | | (to change preferences, visit a project web site or select Preferences in the Manager)
Sat 14 Feb 2015 07:30:04 AM AKST | | gui_rpc_auth.cfg is empty - no GUI RPC password protection
Sat 14 Feb 2015 07:30:04 AM AKST | | Not using a proxy
Sat 14 Feb 2015 04:30:36 PM AKST | rosetta@home | Restarting task rb_02_14_53552_99182_ab_stage0_h002___robetta_IGNORE_THE_REST_12_11_243632_6_0 using minirosetta version 352 in slot 0
Sat 14 Feb 2015 04:30:36 PM AKST | rosetta@home | Sending scheduler request: To fetch work.
Sat 14 Feb 2015 04:30:36 PM AKST | rosetta@home | Requesting new tasks for CPU
Sat 14 Feb 2015 04:31:15 PM AKST | rosetta@home | Restarting task 9ef0b6952d8e35d7bc387617f7c688e2_deNovoHeterodimer_15_02_20_38_06_globalDocking_2_SAVE_ALL_OUT_243449_12_0 using minirosetta version 352 in slot 1
Sat 14 Feb 2015 04:31:15 PM AKST | rosetta@home | Restarting task a08224fe1852490aabb5f82b1b6decd4_deNovoHeterodimer_15_02_20_02_20_globalDocking_6_SAVE_ALL_OUT_243160_11_0 using minirosetta version 352 in slot 2
Sat 14 Feb 2015 04:31:15 PM AKST | rosetta@home | Restarting task Ross3X3_SAVE_ALL_OUT_t178_001_243573_475_0 using minirosetta version 352 in slot 3
Sat 14 Feb 2015 04:31:15 PM AKST | rosetta@home | Restarting task Ross3X3_SAVE_ALL_OUT_t179_010_243575_483_0 using minirosetta version 352 in slot 4
Sat 14 Feb 2015 04:31:23 PM AKST | rosetta@home | Scheduler request completed: got 0 new tasks
Sat 14 Feb 2015 04:31:23 PM AKST | rosetta@home | No work sent
Sat 14 Feb 2015 04:34:13 PM AKST | | Fetching configuration file from http://bam.boincstats.com/get_project_config.php
Sat 14 Feb 2015 04:34:17 PM AKST | | Contacting account manager at http://bam.boincstats.com/
Sat 14 Feb 2015 04:34:18 PM AKST | | Account manager: BAM! User: 181430, Linden
Sat 14 Feb 2015 04:34:18 PM AKST | | Account manager: BAM! Host: 559734
Sat 14 Feb 2015 04:34:18 PM AKST | | Account manager: Number of BAM! connections for this host: 8
Sat 14 Feb 2015 04:34:18 PM AKST | | Account manager contact succeeded
Sat 14 Feb 2015 04:38:21 PM AKST | rosetta@home | update requested by user
Sat 14 Feb 2015 04:38:25 PM AKST | rosetta@home | Sending scheduler request: Requested by user.
Sat 14 Feb 2015 04:38:25 PM AKST | rosetta@home | Requesting new tasks for CPU
Sat 14 Feb 2015 04:38:27 PM AKST | rosetta@home | Scheduler request completed: got 0 new tasks
Sat 14 Feb 2015 04:38:27 PM AKST | rosetta@home | No work sent
Sat 14 Feb 2015 04:42:42 PM AKST | rosetta@home | Resetting project
Sat 14 Feb 2015 04:42:47 PM AKST | rosetta@home | Master file download succeeded
Sat 14 Feb 2015 04:42:52 PM AKST | rosetta@home | Sending scheduler request: To fetch work.
Sat 14 Feb 2015 04:42:52 PM AKST | rosetta@home | Requesting new tasks for CPU
Sat 14 Feb 2015 04:42:53 PM AKST | rosetta@home | Scheduler request completed: got 0 new tasks
Sat 14 Feb 2015 04:42:53 PM AKST | rosetta@home | No work sent
Sat 14 Feb 2015 04:48:53 PM AKST | | Fetching configuration file from http://bam.boincstats.com/get_project_config.php
Sat 14 Feb 2015 04:48:57 PM AKST | | Contacting account manager at http://bam.boincstats.com/
Sat 14 Feb 2015 04:48:59 PM AKST | | Account manager: BAM! User: 181430, Linden
Sat 14 Feb 2015 04:48:59 PM AKST | | Account manager: BAM! Host: 559734
Sat 14 Feb 2015 04:48:59 PM AKST | | Account manager: Number of BAM! connections for this host: 9
Sat 14 Feb 2015 04:48:59 PM AKST | | Account manager contact succeeded
 
I'm noticing this

Sat 14 Feb 2015 07:30:04 AM AKST | | max disk usage: 0.51GB

Check your settings for how much hard drive BOINC is allowed to use....

Edit: Let's just say, that Rosetta currently takes up about 1.5GB space on my 32 thread system and it only has like 4 work units in progress at the moment... So, I'm assuming that rig doesn't have enough space allowed for BOINC per your settings. Let me know if it works after upping your limits quite a bit...

Edit2: And yes having the wrong date/time set in your system can have issues with BOINC pulling work, deadlines, etc... so make you it is set correctly.
 
Last edited:
I set max usage as 40GB, 55%. It's an 80GB HDD. "No work available to process" still shows after performing a synchronize.

I've gotta run. I'll be out for about three hours. Thanks for your help, so much! It's so pleasing to employ this hardware for something useful.
 
There is one more spot to check your setting and that is the "Leave at least" setting. Double check to make sure the HDD isn't full for some reason.

Rosetta's server status says there is plenty of work. http://boinc.bakerlab.org/rosetta/rah_status.php

However, since you reset the project you would have removed any work you had previously downloaded. So that No work available to process doesn't make much sense unless it is saying you haven't downloaded any. What are the event logs when you see that message? From the above Event log it is just saying none was sent. Not that there weren't any available. I'm guessing BOINC is calling for work since there is some space left (potentially) but Rosetta's servers aren't sending any work because it isn't seeing enough space for what work is there... Check that last setting for me and just set it to 0 for now to see if things change. You can always change it to something else. Again that was the "Leave at least" setting under disk and memory usage.

I'm also noticing the general preferences were being pulled from BAM instead of the client... that may be where the issue lies. Make the change via your client and that will override BAM preferences as it will be the latest preferences.
 
Ooops I guess I should have told you about the memory requirements of Roseta. Any way when you can not get the account manager to connect close the manager and run ( sudo service boinc-client stop ) then run ( sudo service boinc-client start ) If you are running Roseta let it run for a little while before opening the account manager again Roseta can be a bit tough in getting running again if not shut down properly, wait 5 or 10 minutes before starting the account manager.

As far as 14.04 iirc tears boinc did work on it but it has been a while since I ran it, so you know how these old fellas memory play tricks on them sometimes. :eek:

On the #4 machine did you check the settings you may have it limited to xx% of processor usage.

P.S. it is hard to leave them idles isn't it. :D

By the way there are some of the projects that do better on AMD than intel I listed a few of them in the forum before but I seem to have lost the thread Gilithanis might have it though it was in one of his threads.
 
Grandpa_01, I believe you are looking for the postings under the All Inclusive DC List in the Guides section....

However, there are not many projects that AMD's will outscore Intel. They may produce more science, but Intels typically score better due to BOINC flawed benchmark system. And I don't believe we have done a full exhaustive comparison on each project. The Commandos of days past had been short handed....lol

Edit: I should probably put a footnote on our findings with HT on the PrimeGrid sub projects. Some people find that SSD's level the playing field to where HT is beneficial in some rigs to be better than turning HT off. Heat is still greater of course, but I don't have access to many SSD's to actually do any comparisons.
 
fastgeek may also want to clarify the hard disk he was running that many on as every hard drive is a bit different and some will not like more than 1 or 2. SSD's certainly fair better for IO projects.

IIRC it was a RAID 5 or 6 using 6x2.5" 10k SAS drives. Had to send that box back, so can't verify for sure anymore. :(
 
There is one more spot to check your setting and that is the "Leave at least" setting. Double check to make sure the HDD isn't full for some reason.
Bingo. The drive wasn't full. It didn't even have much unused content, but it DID have a 34GB swap space. 34GB, what the....? In BOINC's computer preferences, I changed the "Leave at least" amount of disk space to minimum, 5GB. After doing that and restarting the system, within 10 minutes, all cores were processing. I may just reinstall from the ground up tomorrow on an HDD with more capacity.

As for the client not downloading more tasks, it resolved that problem without my intervention. I don't know how. Anyway, as it stands now, all four 4Ps are crunching Rosetta for HardOCP...and heating my Alaska home.

If you are running Roseta let it run for a little while before opening the account manager again Roseta can be a bit tough in getting running again if not shut down properly, wait 5 or 10 minutes before starting the account manager.
I followed your advice. Maybe that's why the client-server communications worked again. BTW, how does one shut down BOINC properly?

As far as 14.04 iirc tears boinc did work on it but it has been a while since I ran it, so you know how these old fellas memory play tricks on them sometimes. :eek:
I'm feeling adventurous. Tomorrow I'll reload from a freshly formatted HDD: Ubuntu 14.04 LTS, Tear's overclocking scripts, Tear's tools (TCP/OCNG and so forth), then Tear's optimized BOINC client. We'll see how it goes.

P.S. it is hard to leave them idles isn't it. :D
I was talking to my wife over a couple ales this evening, and I remarked that I was really happy to have this hardware working hard again for something worthwhile, a worthy cause. She said she's glad too, that I was grumpy all last week after turning off all the 4P beasts.
 
Glad to hear that you got it going. You will find that the Event Logs (especially posting from a fresh startup) are the easiest way to trouble shoot what is going on. BOINC lets you tweak and configure so much that it is easy to get lost in which option is breaking things. Now just wait until you decide to get even more adventurous and start delving into cc_config.xml and app_config.xml files.... lol But we will cross that bridge when the time comes.
 
And I about forgot, make sure to attach to WUProp. That project just monitors other BOINC projects and is non-CPU intensive. Basically, it is like an extra work unit that you can run alongside everything else that mostly sits in memory. The project is incredibly helpful and should be added to anyones list of BOINC references.
 
I'm in the same boat with Linden, just starting to convert over my folding boxes. I have 16GB of RAM on a OS from my USB memory stick. I was able to complete cripple the box with Einstein project. How can you know the memory requirements for a project and what's a project recommendation for 48 cores, 16GB of ram, running on a 8GB OS USB Memory stick?
 
bowlinra, a great way to find project requirements in an easy location is at a BOINC project called WUProp. They have a results link on the front page. Just click here: http://wuprop.boinc-af.org/results/ram.py

BOINC doesn't have many projects that have multi threaded apps which will be the best fit for systems with low memory per core ratios. However, there are a few projects that have low RAM needs such as enigma or SAT. WCG projects vary. CEP2 wants up to 1GB each but you can find the specs on their page here http://www.worldcommunitygrid.org/help/viewTopic.do?shortName=minimumreq
 
Sorry the last post was quick written as my little one isn't feeling well. If you visit my All Inclusive DC List in the Guides section here: http://hardforum.com/showthread.php?t=1801908 Post 6 has a list of projects that have multi threaded apps. But make sure you go to the project page to edit your project settings to only select the work you actually want to run as they may have non-multi threaded apps as well.
 
Back
Top