Badge hunting for both hpf2 and gfam. efficiency drive.

[Mumak]

My hosts were overloaded with GFAM tasks which I couldn't finish on time, so I have aborted about 150 WUs. Now I should maintain enough GFAM WUs to finish the target with my own hosts and could start building a cache for HPF2. Current run time returned is 32-36 days/day, so I believe the GFAM target could be reached on Sunday.
Will monitor how the situation evolves and adjust accordingly...

EDIT: OC, I posted this before I read your previous post, so that was my thought too Currently my rigs hold ~6 days of GFAM work, the rest is commited to HPF2, they all are set to buffer 10 days. I'll drop more GFAM WUs the closer I'm to the target and the more I'm sure I can reach it, so the HPF2 WUs fill-up before they dry out.

EDIT2: I have also disabled GFAM in my profile to avoid downloading further WUs (yep, I still got several resends), since I believe there's enough work buffered to reach the target. All machines on my account have currently 1200 GFAM WUs buffered ! That's plenty enough for ~100 more days needed

EDIT3: The current HPF2 buffer among all machines is ~1000 WUs and going up..

EDIT4: I think all helpers can abort most GFAM WUs in cache and leave ~2-3 days of GFAM work in their caches, so the HPF2 WUs get loaded. Then continuosly reduce the GFAM buffers depending on more precise status. This way GFAM will be finished ASAP and a full switch to HPF2 can be made, but all depends on how soon HPF2 new work will end... Due to the nature of HPF2, I believe there might be a lot of resends that we can catch even after the end date...

[Mumak]

GFAM update: 1:272
HPF2 status unchanged - all results returned are PV yet

[bluestang]

With the large caches of WUs we need to set, we are almost guaranteed to be running in high priority right? Or am I the only one having this issue?

My caches settings are now set for 8 min + 2 max add.

How does running in high priority affect getting new work?

[Mumak]

I don't seem to be having any problems running tasks in High Priority - I'm still getting new WUs.

[bluestang]

Ok, TestPC is now d/l'ing more work. I was able to get two batches of 30 WUs.

Still says quota 19 days, but I know have 84 WU's in cache @ 3.75 hrs each so about 13 days runtime/6.5 days of work.

[Yeroon]

Bocksie is having trouble receiving new wu. Its only looking for HPF2 tasks but comes back with 0.
Fresh non-GUI Debian (?), 7.0.27. Is this an issue of being a new machine? Or do I have some settings wrong? (Should be default boinc settings)

[Mumak]

Carrissa machine seems to return lots of errors...

[Mumak]

Bocksie is having trouble receiving new wu. Its only looking for HPF2 tasks but comes back with 0.
Fresh non-GUI Debian (?), 7.0.27. Is this an issue of being a new machine? Or do I have some settings wrong? (Should be default boinc settings)

Default profile is currently set to get HPF2 tasks only. Bocksie is the machine name? Did it get no tasks at all, if yes for how long? Maybe check the Even Log for more details.

[bluestang]

Carrissa machine seems to return lots of errors...

Did on my account as well, along with Snow Crash's rigs and with my 2 rigs. I want to say we probably had 200+ WU's error out in our quest.

They do it right away though, so no worries. Except, might affect resends.

[Yeroon]

It has not recieved any wu at all, since it fired up late last night.

07-Jun-2013 10:25:44 [World Community Grid] update requested by user
07-Jun-2013 10:25:44 [World Community Grid] Sending scheduler request: Requested by user.
07-Jun-2013 10:25:44 [World Community Grid] Requesting new tasks for CPU
07-Jun-2013 10:25:47 [World Community Grid] Scheduler request completed: got 0 new tasks
07-Jun-2013 10:25:47 [World Community Grid] No tasks sent
07-Jun-2013 10:25:47 [World Community Grid] No tasks are available for Human Proteome Folding - Phase 2
07-Jun-2013 10:25:47 [World Community Grid] No tasks are available for the applications you have selected.

[Mumak]

There should be HPF2 tasks available, I got the last one about 20 mins ago. I have no idea why is that, maybe others might know?

[Yeroon]

Swapped weak keys with my account, it DL's new talks from other projects, so switched it back to Mumaks so it can DL new wu if avaiable. None yet though. Does your "default" profile happen to have a core count limit that is possible to reach with this many crunchers? My only other guess.
Kyalami get 5 tasks when requested to see if new tasks were available, but through yojimbo's account.

[Mumak]

There's no such limit set, all the other machines are getting tasks successfully.

[Yeroon]

Yeah, my bad on the cores thing, its per machine, so I doubt that could cause my problem. Looks like I'll keep updating and see if i pick any WU up, unless someones got a better idea.

[Rob_B]

Could it be because you are running linux? Not sure if there is a difference in wu's between linux and windows on this project.

[Yeroon]

Don't think that's the issue as both Kyalami and Ra are running Linux, and I know at least one of them is able to receive wu (probably both though).

[OldChap]

Bocksie is having trouble receiving new wu. Its only looking for HPF2 tasks but comes back with 0.
Fresh non-GUI Debian (?), 7.0.27. Is this an issue of being a new machine? Or do I have some settings wrong? (Should be default boinc settings)

When I started up this linux rig I used an old ssd that was already installed with Mint 14 cinnamon then:
apt-get update
apt-get dist-upgrade
then
got 7.0.65 like this: http://www.xtremesystems.org/forums/...=1#post5024891 (don't forget to change all instances of BOINC version)

attached to wcg. then immediately

detached using cli like this:
cd to BOINCFOLDER then:
boinccmd --project http://www.worldcommunitygrid.org detach
boinccmd --project_attach http://www.worldcommunitygrid.org 829514_0d906f88e81e7441291e0a6b5bccca65

The above weak key is Mumak's from post 11 of this thread

Once it had dl'd the basics it pretty much started work straight away

Not sure if any of this will help but....

You may want to check tools>preferences> cpu and network settings just to be sure.

EDIT:
The rig in question has 180-200 wu's a day added for each of deadline days 14th, 15th, 16th and so far today has around 150 for the 17th so still getting work fine it seems.

[OldChap]

Update #8

Okay, currently I'm at: GO Fight Against Malaria 1:242:00:20:29


05/31/2013 0:044:22:21:28

06/01/2013 0:050:11:18:21

06/02/2013 0:054:19:19:03

06/03/2013 0:060:13:08:32

06/04/2013 0:053:16:57:37

06/05/2013 0:064:11:41:05

06/06/2013 0:072:13:45:45

06/07/2013 0:000:15:17:33

I have 102 PV WUs right now, so (102 tasks * ~4 hours)/24=17 days. 123 days left-17=106 more days with ~60 days/day of compute time = around 2 more days needed on GFAM for Sapphire. 123 * 6 WU per day = 738 WUs (4hr/WU) or 123 * ~6.8 WU per day = 835 WUs (3.5hr/WU) needed to reach goal. Of course I'm making the assumption of 4 hr/WU to 3.5 hr/WU.

I currently have 1094 GFAM WUs in queue among all devices. We need between 738 to 835 WUs and we have a surplus between 356 and 259 WUs.

I'm at HPF2 1:026:09:01:13 There are 3243 HPF2 WUs in queue among all rigs. I have 275 HPF2 PVs. I'm guessing an average of 5 hours per WU. 275 PVs *5 hours/24= 57 days. This places is us at 282 more days needed for Sapphire on HPF2. The WUs in queue give us 3243 * 5 / 24= 675 days in queue for HPF2.

I've been working a lot, so the updates haven't been as frequent. I hope this helps clear things up as to where we're at.

Excellent post One.shot!!! It contains ALL the pertinent information and calculations needed to show me that we don't need to do anything else for you..... and more especially we don't have to worry about having enough HPF2 work

Very happy with this.

[OldChap]

Originally Posted by yojimbo197
Human Proteome Folding - Phase 2 run time: 1:046:18:58:46

319 days, 6 hours still needed

I have my 3 rigs(10 cores) plus:
Aco2-508 8 cores PecosRiverM
Bco2-508 8 cores PecosRiverM <<< Thanks
a 12 cores - Fallwind
francine 8 cores - jeanguy2 <<< will be moving
ginette 12 cores - jeanguy2 <<< will be moving
Kyalami 3 cores - Yeroon
Home-PC 8 cores - Bluestang
cordell 4 cores - Bluestang

Not sure if you cannot post here PecosRiverM or if we have 2 different names for same guy but whichever...Thank you from me. My email is listed on WCG if you want to clarify

yojimbo: You should have 46 cores on this which gives you a maximum of around 550 days with 10 day cache + 2 days running. ~460 days with 8 day cache so all is looking great.

How many wu's in cache overall? I would be looking for about 1900 to be really sure.

[OldChap]

My hosts were overloaded with GFAM tasks which I couldn't finish on time, so I have aborted about 150 WUs. Now I should maintain enough GFAM WUs to finish the target with my own hosts and could start building a cache for HPF2. Current run time returned is 32-36 days/day, so I believe the GFAM target could be reached on Sunday.
Will monitor how the situation evolves and adjust accordingly...

EDIT: OC, I posted this before I read your previous post, so that was my thought too Currently my rigs hold ~6 days of GFAM work, the rest is commited to HPF2, they all are set to buffer 10 days. I'll drop more GFAM WUs the closer I'm to the target and the more I'm sure I can reach it, so the HPF2 WUs fill-up before they dry out.

EDIT2: I have also disabled GFAM in my profile to avoid downloading further WUs (yep, I still got several resends), since I believe there's enough work buffered to reach the target. All machines on my account have currently 1200 GFAM WUs buffered ! That's plenty enough for ~100 more days needed

EDIT3: The current HPF2 buffer among all machines is ~1000 WUs and going up..

EDIT4: I think all helpers can abort most GFAM WUs in cache and leave ~2-3 days of GFAM work in their caches, so the HPF2 WUs get loaded. Then continuosly reduce the GFAM buffers depending on more precise status. This way GFAM will be finished ASAP and a full switch to HPF2 can be made, but all depends on how soon HPF2 new work will end... Due to the nature of HPF2, I believe there might be a lot of resends that we can catch even after the end date...

I only dumped about 25 wu's. We need to be sure not to get this bit wrong as much as we need to get HPF2 work. I feel we have another 24 hours before we REALLY need to know how much work is already in cache for HPF2 so no more dumping please, at least not yet.

Take a look at One.shot's post and maybe give us as much of that sort of info as you possibly can for the next couple of days. WU's in HPF2 cache total interests me greatly.

I will try to get number of cores info updated by this time tomorrow.

[Mumak]

I have updated the list of helper machines here: http://www.xtremesystems.org/forums/...=1#post5192178

As of now, there are 90 GFAM PVs, very roughly that should give ~15 days of work, so let's take 10 to be safe.
94 days of work is needed for GFAM target, -PVs = 84 days needed.
Current rate is ~30 days of work/day

Currently there are 2530 HPF2 tasks in caches among all machines including mine.

[OldChap]

So, already 420 days @ 4 hours average per wu.

I am still running GFAM so I have no point of reference for HPF2 average runtime but I am thinking 4 hours is a bit on the low side.

The very least you will get is emerald

save me the pain guys... EDITED

simracingmedia2 - Rob - 8 cores
Kyalami - Yeroon - 3 cores
Quad - OC - 4 cores

TestPC - bluestang - 2 cores
ginette - jeanguy2 - 12 cores
francine - jeanguy2 - 8 cores
d -fallwind - 8 cores
e -fallwind - 8 cores
f -fallwind - 8 cores
Carrissa - fallwind - 12 cores
ra - Yeroon's bro - 6 cores (8 really but not full time)
MainToy - PecosRiverM - 8 cores
Daughter - PecosRiverM - 4 cores
another - PecosRiverM - 4 cores

...am I right so far? 91 cores + another machine and was it 20 cores of Mumak's own?

35 cores running GFAM now able to get 7 days would mean something in the order of 245 days

64 cores able to run now and get 8 days cache too could be 640 days

Mumak: that cache of yours could be growing pretty damn fast ...I predict >5000 wu's in total.

2 years @ 4 hours per wu needs 4380 wu's evenly spread in time across all the rigs

[bluestang]

TestPC - bluestang - 2 cores

Carrissa - fallwind - 12 cores

[fallwind]

Carrissa=12 cores. d, e and f are 8 cores each. All have been set to 7 days cache.

[Mumak]

@OC: yes, I have 20 cores.
I have checked returned HPF2 results and it seem the run time ranges between ~3 to ~5.7 hours. If you need a more exact mean value based on results returned so far, I can calc that...
Current total HPF2 cache is 3165 WUs and I'm sure it will grow..