I just received my first four "mega" WUs. For CPU. I put other tasks on suspend and started doing these. Run time seems to be normal... *sigh*
So this is the new S@H crippling we will be plagued with?

---

I've deliberately kicked into action the fresh 5342s I've got & there's so little GPU load, despite all 4 threads going, that I can turn fans down to 0 & temps are still in teh 50s not the 70s. GPU fan speed near idling also at 35%. 4+ mins & still 0no % progress so far.

Aah here they go. Took 7 minutes to start!

I currently have 20 of them on my AMD machine and 22 on the i5.

---

got also 4 new tasks starting with 21jn11ac.5342...
start_ra = 5.08
start_dec = 15.97
end_ra = 5.11
end_dec = 17.97
true_angel_range = 0,430

and one normal but strange result is visible at server:
http://setiathome.berkeley.edu/result.php?resultid=2287096316
send: 28 Jan 2012 | 19:25:43 UTC
expiration date: 28 Jan 2012 | 19:38:27 UTC
can't find the result on the client by now.

---

Matthias

Mega-WUs seem to be getting commoner. I've got seven downloads stuck at the moment, and four of them are mega-WUs from the new 21jn11ac.5342. batch. But I did just report an older one, and it validated at the first attempt against a stock v6.10

Looks like it is not the whole 'tape'. I have this one that is of normal size.
http://setiathome.berkeley.edu/workunit.php?wuid=915319477

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

Mega-WUs seem to be getting commoner. I've got seven downloads stuck at the moment, and four of them are mega-WUs from the new 21jn11ac.5342. batch. But I did just report an older one, and it validated at the first attempt against a stock v6.10

Doesn't seem especially smart to send out 2.8 Meg MB files that doesn't require more calculations/crunching than the old 366K WU's.

I was wondering that myself, if the filesize is larger surely it should take longer to crunch? If not, oh boy the pipe is gonna have trouble if this is the new size of the tasks. Speaking of which, has anything like this happened before?

---

Member of the People Encouraging Niceness In Society club.

Yeah, but in the past AP work usually resuled in bandwidth bursts, not continued saturation.

Don't forget the shorties - at least three tapes (05oc11ai, 11oc11ah, 25oc11ab) have been producing shorties this morning.

And the bandwidth is hardly unusual - nowhere near ready for my little yellow friend to take his swim, yet.

Since last night most of my downloads have been VLARs or mid range. Very few shorties in sight in the last 18 hours.

And naturally an hour after i posted that the shorties came raining down.

---

Grant
Darwin NT

Is anyone else having connection problems,i report and upload but all my downloads are stuck saying "connect() failed".

Mega-WUs seem to be getting commoner. I've got seven downloads stuck at the moment, and four of them are mega-WUs from the new 21jn11ac.5342. batch. But I did just report an older one, and it validated at the first attempt against a stock v6.10

Doesn't seem especially smart to send out 2.8 Meg MB files that doesn't require more calculations/crunching than the old 366K WU's.

I was wondering that myself, if the filesize is larger surely it should take longer to crunch? If not, oh boy the pipe is gonna have trouble if this is the new size of the tasks. Speaking of which, has anything like this happened before?

There have been various glitches with work generation down the years, but no: nothing exactly like this has happened before (within my knowledge, at least).

And nobody I've spoken to, either here or in the various back-corridors at Lunatics, has suggested that these tasks were in any way expected. They're not Green Bank data: they're not related to the v7 developments in beta: and there's been nothing on the programming front (code or emails from staff) to prepare for them.

Everyone's best working guess is that the mega-WUs are a mistake or coding bug, and will disappear again as silently as they arrived. But we won't know for certain until the lab opens for business again next week, and we can try to coax an explanation out of the staff.

Mega-WUs seem to be getting commoner. I've got seven downloads stuck at the moment, and four of them are mega-WUs from the new 21jn11ac.5342. batch. But I did just report an older one, and it validated at the first attempt against a stock v6.10

I don't know where they are coming from unless they are breeding but the 13 I had have been processed and returned, some have been verified others are awaiting.

Since then another 39 all GPU have appeared, all are _0 or _1.

---

Kevin

Me I'm just getting HTTP errors and this:

1/28/2012 11:35:37 AM Project communication failed: attempting access to reference site

I can upload, report, but not download and so far I have 125 downloads that I can't get to go through.

Bump...

Now it's 162...

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

Me I'm just getting HTTP errors and this:

1/28/2012 11:35:37 AM Project communication failed: attempting access to reference site

I can upload, report, but not download and so far I have 125 downloads that I can't get to go through.

Bump...

Now it's 162...

No such problems at this end of the world.

Cheers.

---

Me I'm just getting HTTP errors and this:

1/28/2012 11:35:37 AM Project communication failed: attempting access to reference site

I can upload, report, but not download and so far I have 125 downloads that I can't get to go through.

Bump...

Now it's 162...

No such problems at this end of the world.

Cheers.

Well It's back to normal, I'm downloading now once again, someone must have hit the button on the remote control...

Spoke too soon, for a few minutes a window opened and I grabbed 261, I have 95 left to grab, maybe later they'll be done on their own.

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

Me I'm just getting HTTP errors and this:

1/28/2012 11:35:37 AM Project communication failed: attempting access to reference site

I can upload, report, but not download and so far I have 125 downloads that I can't get to go through.

Bump...

Now it's 162...

No such problems at this end of the world.

Cheers.

Well It's back to normal, I'm downloading now once again, someone must have hit the button on the remote control...

Spoke too soon, for a few minutes a window opened and I grabbed 261, I have 95 left to grab, maybe later they'll be done on their own.

Ah never mind, another hole opened up, for the moment the 6 gpus have enough...

---

The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's

I'm still getting a high proportion of download failures, but only on Campus as far as I've seen. Might have to tickle Networks tomorrow...

---

Trouble getting validated results through atm? RAC flatlining.

No comms problems here lately.

I have however seen the very long preprocessing times for those mystery jumbo WUs to start up and begin actual GPU work.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

Mega-WUs seem to be getting commoner. I've got seven downloads stuck at the moment, and four of them are mega-WUs from the new 21jn11ac.5342. batch. But I did just report an older one, and it validated at the first attempt against a stock v6.10

Doesn't seem especially smart to send out 2.8 Meg MB files that doesn't require more calculations/crunching than the old 366K WU's.

I was wondering that myself, if the filesize is larger surely it should take longer to crunch? If not, oh boy the pipe is gonna have trouble if this is the new size of the tasks. Speaking of which, has anything like this happened before?

There have been various glitches with work generation down the years, but no: nothing exactly like this has happened before (within my knowledge, at least).

And nobody I've spoken to, either here or in the various back-corridors at Lunatics, has suggested that these tasks were in any way expected. They're not Green Bank data: they're not related to the v7 developments in beta: and there's been nothing on the programming front (code or emails from staff) to prepare for them.

Everyone's best working guess is that the mega-WUs are a mistake or coding bug, and will disappear again as silently as they arrived. But we won't know for certain until the lab opens for business again next week, and we can try to coax an explanation out of the staff.

Thanks for the, er, info :) I do hope it is a temporary thing. Or something that's easily fixable.

---

Member of the People Encouraging Niceness In Society club.

And it looks like those 21jn11ac tasks are still being issued, even though that dataset still has not shown it's face on the server status page.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

And it looks like those 21jn11ac tasks are still being issued, even though that dataset still has not shown it's face on the server status page.

What splitter PID are you seeing as the second element in the task name? And are they mega, or ordinary WUs?