World Community Grid

[RFGuy_KCCO]

Thanks for the idea about using app_config to restrict the number of running CEP tasks! That is a great solution that will allow me to continue crunching CEP tasks.

Yes, I made a rookie mistake when deselecting CEP tasks in WCG... It helps to also uncheck the "If no work available for the selected projects..." box. D'oh!

 

[Gilthanis]

https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,37686

Greetings all,

We are starting a new beta test for Android only FightAIDS@Home. This version has a few changes in it.

1. It will have the updated checkpointing for VINA to help prevent invalids on mid job checkpoints.
2. It has been compiled to only run on Android 4.1 and above.

This beta test will have 500 work units to be sent out. I have disabled the assimilator so the results will stay around in the database for us to inspect.

Thanks,
-Uplinger

 

[Gilthanis]

I have picked up 2 BETA's so far.

Edit: Make that 12

 

[texinga]

Let's see...Gil is either at work or maybe asleep.

Ok I'll post this small newsbyte.

Latest from Uplinger on the OutSmart Ebola project WUs progression.

http://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,37507_offset,310#483348

 

[Gilthanis]

Naw, just didn't read it yet. Too busy with 100 other projects... lol But yeah until they are ready, I don't get too excited. Sometimes those dates get pushed back quite a long time. But glad to hear they are finally making the progress. They really dropped the ball with that PR blitz...

 

[texinga]

Yeah, I know it lacks anything definite and there is really nothing to count upon in Uplinger's message. It has just been so long without any news that I figured sharing that bit of news might be something amid all the silence for that project.

For WCG "oldtimers", we know not to get overly concerned about project hiccups and even long delays. But for newbies, who got excited about helping out with the Ebola cure, the PR Blitz (with no work to do for months) can mean losing them all together. It is hard enough to even recruit people you know into Crunching much less hope they stay if the work peters-out.

 

[Gilthanis]

Exactly

 

[texinga]

The Alpha must have gone well because the Beta Units are flowing right now for Outsmart Ebola. Get 'em while they are here.

http://www.worldcommunitygrid.org/forums/wcg/viewthread_thread,37698

 

[tjmagneto]

I was checking my work queue and I have a few beta Outsmart Ebola Together tasks waiting to start.

 

[Gilthanis]

Nice catch guys.. I was in the middle of various support calls. So far I have gotten ~14 or so...


Edit: 79 so far

 

[texinga]

As mentioned in Uplinger's beta message, there was also a planned outage to do an upgrade on their systems. Looks like that started about 10pm Eastern yesterday and they are doing some testing now. Hope their upgrade goes well. This has been a great beta in terms of size and that it helps move the ball along toward working on the Ebola disease. I love Crunching for things like this...it's mah thang!

 

[Gilthanis]

BETA's still going out. I've had over 160 so far...

 

[Gilthanis]

I'm still curious as to what the new projects they hinted at back in January will be. However, like always they never give you any answers in advance...

 

[texinga]

Yeah, the only new (named) upcoming project that I recall reading about is the Phase II for Childhood Cancer project that is supposed to start-up in the future. I'd love to see more new projects at WCG.

On the OET Beta, I ended up processing 526 as "valid" and have another 25 "pending validation".

 

[AgrFan]

OET work is flowing again. New work will be available throughout the day. OET is set to low priority. Make sure to set your profile to OET only and deselect "If there's no work available ...".
https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,37742

 

[texinga]

Thank you AgriFan!! I'd stopped in here first today and read your post which sent me straight over to making sure my rigs were ready to get this work. All are now happily munching on OET WUs.

 

[Gilthanis]

Android devices are getting OET as well.

 

[Gilthanis]

I just got a resend. It is really odd to keep getting OET before all BETA were returned...lol

 

[texinga]

Thanks Gil for that heads-up. Running some OET on my Tegra Note 7 right now.

 

[Linden]

I just joined HardOCP WCG under the name "LindenHOCP." My name on other projects will be the same as my forums name, "Linden." Unfortunately, at WCG, 'Linden' was already taken. But anyway, I'm a recent refugee from a well-known DC project. Heck, it was time to join an [H] team anyway! I don't know what mix of resources I'll add to WCG, but the potential is in my signature.

Let me know whatever nuggets of wisdom you have. I'm a newbie to both WCG and BOINC.

Regards, and I look forward to playing with you.

 

[Gilthanis]

Linden, here are the heads ups for WCG.
1. CEP2 is IO intensive and can use up to 1GB per work unit. This is the only sub project at WCG that is opt in as far as the number of work units allowed to be sent to each system. The uploads can be large compared to the others as well. Keep an eye on it and adjust your settings accordingly.

2. No GPU work right now, so hold tight and use GPU's elsewhere for now.

3. UGM had been a bit IO heavy and they were working on fixing that. I haven't kept up so others may chime in if events have changed.

4. ARM devices only run FAAH and OET currently.

5. FAAH has two different apps. You can 't control which you get. VINA can be up to 50% faster on Linux systems.

6. Stats on the WCG page are updated twice a day.

7. 1 BOINC point = 7 WCG points. This is due to not fixing the points from back before WCG used BOINC. (formerly used UD)

8. Badges are based on run time and not points like most other projects with badges.

9. BETA is opt in and you can't control which projects to do BETA work for. BETA's don't go out very often and when they do they are in short supply. This is the hardest badge to earn at WCG. They also limit you to 1 per thread/core at a time and only send them to reliable systems known to return work within 4 days. So, keep your cache less than 4 days to qualify.

 

[Linden]

The only hardware I'm crunching with are multi-core Opteron systems, so most your advice is not difficult for me.

CEP2 is IO intensive and can use up to 1GB per work unit. This is the only sub project at WCG that is opt in as far as the number of work units allowed to be sent to each system. The uploads can be large compared to the others as well. Keep an eye on it and adjust your settings accordingly.

How many work units will it assign per machine? One each per CPU core? How do I adjust settings for WCG downloads. Due to different user names and passwords between BAM and WCG account, I had to directly create an account at WCG, then attach it directly to the BOINC manager for each host. (I could not create or find my WCG account trough BAM's Sign Up for Projects. I either had to change my BAM name and password or select a different name at WCG, which I did.) OK, so that's the background why I can't change WCG options at BAM. I did not see anywhere at the WCG site or in BOINC manager where I can make WCG download settings and opt-in/outs.

 

[Gilthanis]

To change what sub projects you support, you have to do that at WCG. You log in, click My Contribution in top right, Click My projects on the left, first paragraph will have a link for device manager (click it), Click on device profiles at the left, Click Default (unless you set up others), Scroll down to "Project Specific Settings" and change those according to what you desire. The default is to only run one CEP2 unit per machine at a time.

You can set up different profiles and assign rigs to those profiles too. WCG allows Default, Home, Work, School.

Also, WCG uses a hybrid BOINC server and is constantly being patched and bridged, which is why it doesn't always work like other projects. Most other projects use the default BOINC websites built into the server. So, what you learn from WCG won't necessarily apply at all projects...

 

[Linden]

Thanks for the info. I think I'll just stay with default, where it is now. I ain't fancy enough yet to get that picky. You see, I'm like Grandpa. If I think too hard my head hurts. If my head hurts, I have to go to bed early. If I go to bed early, I can't drink as much. If I don't drink as much, the local brewery suffers and they have to lay off a twenty-something aspiring brewer. Since the pink-slipped chap is unemployed, he must turn to a life of crime to feed his five babies. It's just plain awful and unjust.

But I digress.

 

[Gilthanis]

Don't forget that if you can't drink as much, you won't take those pills to get rid of the headaches. That also trickles down to workers not feeding their kids.... It really is a vicious cycle.

 

[Jathanis]

Thanks for the info. I think I'll just stay with default, where it is now. I ain't fancy enough yet to get that picky. You see, I'm like Grandpa. If I think too hard my head hurts. If my head hurts, I have to go to bed early. If I go to bed early, I can't drink as much. If I don't drink as much, the local brewery suffers and they have to lay off a twenty-something aspiring brewer. Since the pink-slipped chap is unemployed, he must turn to a life of crime to feed his five babies. It's just plain awful and unjust.

But I digress.

You. I like you.

 

[fastgeek]

OET WUs are online? Have reached 10M goal on PG. Letting my boxes finish up those WU's and moving them back to WCG for this.... SSSLLLUUURRRRRRPPPPPPPP!

 

[texinga]

Heads up, there is another "major upgrade" planned for this Saturday at WCG. If the last one is still fresh in your mind, you may want to set your WU Cache accordingly (especially if you run with 0.0 Cache setting).

http://www.worldcommunitygrid.org/about_us/viewNewsArticle.do?articleId=421

 

[Gilthanis]

Or set up a backup project

 

[Linden]

Thanks for the heads up. Would a simple solution be just to go into the local BOINC manager and select "no new work" for WCG; to then, after the weekend, put the settings back to accept?

 

[Gilthanis]

I would let WCG continue pulling if that is what you want it to ultimately be doing. I would just add at least one other non-WCG project and set it to priority 0 so that if WCG can't pull work, it will try to pull from the other project...

 

[texinga]

In BOINC Manager, Tools, Computing Preferences, Network Usage tab, you can set the Minimum and Maximum Work Buffers set to x.x days each. It's personal choice, but if you have no additional buffer setup (i.e. 0.0 days), then when the Servers go off the air, and if they are no longer sending out new work, then you'll be dry when you finish the WUs in your queue. That could also be a good time to break-down the Rig for a good cleaning too.

Once the Servers are all back online, your completed work will upload and it will be back to business as usual.

 

[AgrFan]

The network outage is scheduled to last 8 hours. Increase your work cache to 1.0 days and you won't run out of work. Make sure to reset it when the network comes back up.

 

[texinga]

Linden, this may also be of use to you from the BOINC website regarding the BOINC Manager's "days of work" Preferences settings that we have been discussing.

• Minimum work buffer: This sets the minimum amount of work your BOINC is going to request.. It's the low water mark for BOINC, before it starts asking for work.
• Max additional work buffer: This sets the additional days worth of work you want to have.

http://boinc.berkeley.edu/wiki/Local_preferences

There is also a similar "days of work" Preference setting in your Device Profile(s) over on WCG's site, but it is a single entry and is titled "Cache [x.x] extra days of work".

AgrFan's advice would be a good answer to your question and would see you through the outage (even if the unexpected happens and it takes a bit longer than 8hrs). Naw, that wouldn't happen now would it...?

 

[Gilthanis]

Thanks to the incredible generosity of World Community Grid volunteers, the FightAIDS@Home project team has finished with an important stage of their project. The research team has refocused on analyzing their existing results and preparing for the end of this historic grid computing stage.

https://secure.worldcommunitygrid.org/about_us/viewNewsArticle.do?articleId=420

With your help, we have reached a new milestone: no new AutoDock (AD) or AD Vina docking experiments are currently being generated. Put another way: we’re done creating new work tasks. The AutoDock queue is now empty, and the AD Vina queue has more than a year's worth of jobs left. Most of our efforts have shifted towards analysis.

 

[texinga]

Good news and always appreciate the progress communications that we get from WCG.

 

[Nicolas_orleans]

Great news, thanks for sharing

 

[texinga]

Looks like the Network Upgrade work is done at WCG that was set to start at 10pm Eastern last night. All my rigs are functioning as usual (all uploads completed/reported) and plenty of work to do.

Good job by Uplinger and the Guys at WCG!
https://secure.worldcommunitygrid.org/about_us/viewNewsArticle.do?articleId=421

 

[Gilthanis]

For those that weren't paying attention and are interested, BOINC now runs on Android 5 (lollipop) and WCG has work units that also work.

 

[Gilthanis]

Apparently BETA work for UGM went out today. https://secure.worldcommunitygrid.org/forums/wcg/viewthread_thread,37907