seti is running out of work ....dotdotdot

[CryptokiD]

ok heres the deal. seti is basically running out of work. this isnt a hardware problem. its just that seti is processing qorkunits faster then they can get them from the arecebo tele.

as of the last 24 hours there has been little to no work available to download. so if you dont have youre work queue set to 10, now is the time to do so.

matt is working on a software blanking process so that they can split up and send out workunits from data that was previously contaminated via airforce radar and such. hes got it working but its a slow process and can only trickle in new workunits when we need the gardenhose to be fully open.

in the mean time you can get re issued workunits every once in a while from clients who quit seti, crashed, aborted, etc. but these are rather sparse.

too bad there isnt a way to queue to up workunits for more then 10 days. id love to queue up 25 days of work but even if that were possible, its too late now unfortulately because there is nothing left to download...

 

[Careface]

ok heres the deal. seti is basically running out of work. this isnt a hardware problem. its just that seti is processing qorkunits faster then they can get them from the arecebo tele.

as of the last 24 hours there has been little to no work available to download. so if you dont have youre work queue set to 10, now is the time to do so.

matt is working on a software blanking process so that they can split up and send out workunits from data that was previously contaminated via airforce radar and such. hes got it working but its a slow process and can only trickle in new workunits when we need the gardenhose to be fully open.

in the mean time you can get re issued workunits every once in a while from clients who quit seti, crashed, aborted, etc. but these are rather sparse.

too bad there isnt a way to queue to up workunits for more then 10 days. id love to queue up 25 days of work but even if that were possible, its too late now unfortulately because there is nothing left to download...

Yeah, we probably should have done this around 3 days ago >_<. The 'problem' is that the Arecibo telescope is currently undergoing maintenance, so no new work is being produced.

The problem with the blanked WUs that some people have reported is that they aren't running well on GPU systems. I personally havent tested it myself, but I've just set myself to allow new tasts (cache is down to around 3 days ).

From what I've read, the telescope will be back up early November, but we won't be seeing worked for 1-2 weeks after the telescope starts, as they need to fill up at least a decent amount of 50gb tapes to justify sending them back to Berkeley for splitting.

Sad times overall, tbh. I was hoping my 10 day cache would be enough, but it looks like I was wrong

Good luck getting WU guys, we need them!

EDIT: There is a way to queue up more than 10 days of WU; go to BOINC Manager -> Advanced -> Preferences -> Network, and set "Connect about every" to 10 days, and "Additional Work buffer" to 10 days; gives you an effective ~20day cache. Sadly, when I do this, I get around 2000 WU for my GPU, which makes the manager simply not possible to use as it lags so much, though, I have a feeling thats simply because I have a single core CPU

 

[Sir-Epix]

As long as you are crunching MB units you should be fine. The are currently splitting work units from 2006 and using the software radar blanket. I have had no problem getting these for my machines. Note you will not see this on the server status page as these work units are not seen by that script. Happy crunching!

 

[Rickdstik]

Even MB units have run dry at this point. I've got both my machines set to pick up AP/MB work available, and was lucky enough to pull down a AP unit on the work box that'll keep it busy for a couple days. The quad however, is already down to about 10 workunits before it goes idle again. Its too bad, I was just about to make a 4 GHz push with the Q6600 after moving in into a Bob Flay v2 case and giving it a pool to swim in. But given the current state, Ill probably leave it at 3.2 or maybe drop it down to 2.8 until we get more work to do.

 

[Mr.Guvernment]

this sounds like good news, finally caught up!!

now it will be done for science more so then points

think they need a couple new telescopes!

 

[muddocktor]

If your machines go idle, you can always go crunch a few wu's for the Rosetta team. And they are going to run a contest for the last half of the month with some pretty decent prizes. I swapped 3 of the 4 machines I still have running over to Rosie to help them out and pay them back for our guantlet fun.

 

[Duner]

I just got a bunch of work for all my CUDA rigs. Is the drought over? Either way, I'm crunching Rosetta for the rest of the month on my CPU's.

 

[Careface]

I just got a bunch of work for all my CUDA rigs. Is the drought over? Either way, I'm crunching Rosetta for the rest of the month on my CPU's.

Almost over. They have plenty of data from the telescope, but it takes time to get to Berkeley. With the improvements to the software radar blanking script, we are seeing some more work at the moment, but it wont be enough to cope with demand. However, the good news is that by the time that work runs out (I see 4 tapes are being split at the moment), hopefully Berkeley will have their shipment of new telescope data

So, long story short, it'll be slow for the next week or so, but after that, we're back in business.

 

[Jolly-Swagman]

Yeah havent had any work now for good few days either, and are set to 10 day cache.

So for now will have to give the Rossie team a helping hand with the Rossie RAC Harvest.

 

[tricknasty]

is seti still out of work? I finally installed it on my gaming rig to see what it will do and havent gotten any WU's in 3 days.

 

[QuietIce]

If you installed Tuesday (that's still maintenance day) then you probably haven't gotten any work. I've had comm problems the last couple of days but there's plenty of non-AstrroPulse work waiting for you once comms are back up ...