Badge hunters: About 11 days to get Computing for Sustainable Water

[OldChap]

According to posts about the project there are approximately 11 days of work left (plus mop up) before the current run comes to an end and the science is studied.

This is not to say there will not be any more work, just that they may decide to have a phase 2. (so a new badge)

Those of you who are close to a badge probably need to set your cache of work to ensure that you reach your target

Those of you who reach their target should probably return any remaining work in order to enable others to receive it and get their badges too

Others here may need to enlist some help if the time remaining is too short.

Oj feels that he is just about good to get his, so I have 8 threads available to give over a week of work (80 days or perhaps more if cache is included) if this will help anyone.

The Members in need list

Update as at: OCT 4th =*

Otis: Got Sapphire !!

Oj: Got Sapphire !!

Plan.B: Got Sapphire !!

Gamekiller Got Sapphire

Origin_Unknown Got Sapphire !!

bluestang Got Sapphire !!

Haschoiz Got Sapphire !!

werdwerdus Got Sapphire !!

Martijn Ran out of Work Units ....just 22 days short *

one-shot, Too late to the party



By request, here is a copy of post 70 (itself derived from This post) for ease of transferring rigs:

In windows I did this:

No new tasks>abort other WU's> let running tasks complete>go to projects>update.

open a Cmd window and....

cd C:\<whatever your path is to>\BOINC
boinccmd --project http://www.worldcommunitygrid.org detach
boinccmd --project_attach http://www.worldcommunitygrid.org 681752_3e76a26ed744c9a8ebf421de60569457. <<(this is werdwerdus' key)

with Linux I did this:

No new tasks>abort other WU's> let running tasks complete>go to projects>update.

open a console and....

cd BOINC

boinccmd --project_attach http://www.worldcommunitygrid.org 260583_c2c7974e75719159999d09ae26bd9221. <<(this is Martjin's key)

Yep edited the linux version and key for both

[Rob_B]

I have just send 230 WU's back for others to crunch.

And I have two 8 threads crunchers that can help others to get their badge.

[haschioz]

rob_b, what info do you need from me to let your crunchers crunch in my name? my name is haschi. id like ro get emerald i am at 180 day atm, so 3 crunchers should be able to pull it off.

[Rob_B]

Got to the WCG site. And under My Grid > My Profile. you will find your weak account key. Any one can use that key to crunch under your name (but we can't make any change to your account)

Post that key or PM it to me and I will get the 2 crunchers on your name.

[[XC] Oj101]

Thanks OC, I've got 171 days left to go to get to two years, and at my current rate of 14 cores x 11 days is 154 days plus the cache and then I've still got this from Dave's machine:





I'll make it

[OldChap]

Just a thought for all concerned...

There may be a whole bunch of kind hearted folk returning work back to the servers as they complete their badges, making the work available to those trying for badges. This will possibly play havoc with anybody with work pending validation because it is possible that work could be issued to a wingman, passed back not done, reissued, passed back etc. so maybe you will see some wu's taking a lot longer to validate. You will get there eventually though.

EDIT: Oj, I've been at no new tasks for a bit now so should be done in an hour or so. Good luck my friend.

[Gamekiller]

I've been running only this project for a while trying to hurry and get my 2 years in. I'm at 326 days atm... I don't think I can make it.

[[XC] Oj101]

CRAP!! You're right, OK I DO need more CPU power on here ASAP! My queue is slightly shorter than it was, just over 3,000 work units down from over 4,000. Considering it's set at eight days, that means I'm down to about six days of work left That means I need 171/6 = 28.5 cores to make my goal, add in Dave's pages of pending validation and call it 22 to 24 cores needed :/

[OldChap]

Is that 326 days to go Gk?

you seem to be running 12 cores so 120 days without a cache and how big that can be just now is a bit of an unknown.

If I add 8 cores that would be 200 days plus cache ..... we could use another few perhaps ... I will run the q6600 for a few days

Right then folks ...What to do with the weak key for the helper?

[OldChap]

I was already set up for yours Oj so it is running again.

I can still run other rigs for any of you guys if there are still wu's to be had.... just need to know what to do with any weak key I am given

[Rob_B]

This is what works for me for Windows.

Make a bat file with this line in it.
boinc --attach_project http://www.worldcommunitygrid.org WEAK KEY HERE

And put it in the BOINC folder that has the boinccmd.exe in it.
Detach from WCG in the Boinc manager and shut down Boinc. Now run the bat file and let it do its thing. You may have to run it as Admin on some systems.

[Rob_B]

Have a look at this post for more info on using the weak key. Also for Linux.

http://www.xtremesystems.org/forums/...=1#post5054807

[Gamekiller]

Is that 326 days to go Gk?

you seem to be running 12 cores so 120 days without a cache and how big that can be just now is a bit of an unknown.

If I add 8 cores that would be 200 days plus cache ..... we could use another few perhaps ... I will run the q6600 for a few days

That's my current run time on the project. So I'm almost half way there I guess. *nervous*

[OldChap]

I am thinking that those with a decent cache may benefit from dropping all overclocks therefore getting less work done in a day and making the cache last longer BUT You will get less new work in that cache so it is a case of picking the right time to do this.

[[XC] Oj101]

Thanks so much OC, I would've made it if there really were nine days left but I checked the log and my machine has been saying no new tasks for the last eight hours

[Otis11]

Set your cache to 10 days and uncheck the box for other WUs if this project is not available... then manually refresh until it's completely full.

It also helps if your set to report tasks immediately as it will then request to fill your 10 day cache every time a single WU completes. Given that each of my cores does 32 WUs a day, on an 8 core system that's checking 256 times a day! If there are WUs, you'll get them!

[[XC] Oj101]

Thanks Otis but I've had it configured like that for a while expect my cache is limited to eight days as with 8.5 or more I'd have it starting the work unit which was next in line, getting half way and swapping to a less important one. It would then do several others before the most important one eventually expired several days later, so eight days is the most I can do for cache. I've already got CFSW set as the only project, and it is set to connect every 0.1 hours. I've clicked Refresh till I'm blue in the face and gotten a grand total of TWO new work units. TWO. I do a work unit per thread every 29 to 32 minutes, so call it sixteen work units per hour. That means that in the last eight hours I've completed 128 work units and requested only two, and this must have been slowing down over a period of time. I checked a screen shot from before the weekend and had 4351 work units in the queue/running/ready to report, and right now I have 3379. Work is disappearing, and fast.

The last two work units are from batch 16031, the hundred or so before that are batch 15860 to 15870 so there has definitely without a doubt been a slow down of new work coming.

On another note, does anyone know how to set BOINC to download/upload more than two work units per project at a time? The delay between one download ending and the next beginning is longer than the time taken to download a work unit, so I think the total transfer time would be reduced if I could do four or eight (or more ) transfers simultaneously.

[Gamekiller]

I just bumped my cache up and grabbed a few thousand work units. ~1800 on my main rig alone. Some units popped into high priority mode which probably means I need to lower my cache a little or I could end up having it do like Oj said.

[[XC] Oj101]

I'm the only one not getting new work units? Worry not, I'm not aborting, but clicking abort is the only way I know to show the number of work units in the list.



As you can see, there were consistent downloads until about midday, and since then there has been one at 12:14:05, 13:49:23 and 18:5547. It is now 22:22:23. The list is once again a few work units shorter than the last time I posted, subtract two from the number shown as GPUGrid is also running.

[OldChap]

Waiting on a few to complete here. Is your default profile on sustainable water only Gk?

[OldChap]

Oj, Done 8, working on 8, but only 28 in cache so maybe they are being a little miserly in issuing to those with big caches just now

[Gamekiller]

Waiting on a few to complete here. Is your default profile on sustainable water only Gk?

Yes. CFSW is the only thing checked and it's set to not get work from other projects. I do have it set to allow betas (for the HCC gpu) so should I turn that off?

[Otis11]

Since my last post I've DLed 243 WUs.

Seriously, set boinc to report completed work IMMEDIATELY. Every time it connects, which can be many times an hour on multi core machines, it will request more WUs!

[Snow Crash]

I'm the only one not getting new work units? Worry not, I'm not aborting, but clicking abort is the only way I know to show the number of work units in the list.

While on the Tasks tab in BOINC Manager click the button that says "Show all tasks" (you may have clicked it once already when it said "Show active tasks"). As for only get 1 or 2 at a time, WCG will hand over everything it has to you (up to 32 in a single request) but the queue been put on a very slow trickle feed. This actually helps you as this means the 'set and forget' crowd is probably not getting any at all. Also, keep in mind that if a WU takes longer than your average it bumps all WU estimates so only a few seconds higher average can put you over you cache limit quite quickly.

+1 to what Otis said

[[XC] Oj101]

How? Connect every 0.0 hours? Or what?