Looks like they're out of data to split, although even when there was data to split & all the splitters were running, not much was being produced.
Should be out of GPU work in an hour or 2.

Help may be on the way...
I just checked the Cricket graph again and it looks like some upload spikes have started. Should mean more datasets on the way to the servers.

Hopefully once they've got some Data to split, Eric will be albe to sort out why the splitters are so slow at actually splitting it.

Yup....first things first, I guess.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Prior to this weeks outage their splitting rate was as good as it's been since the PFB splitters came online. Not great, but at least enough to keep the ready-to-send buffer full.
Since the outage they've struggled to produce 10/s (20/s appears to be the minimum to meet demand these days).

---

Grant
Darwin NT

Yep, one of my (work time only)-crunchers only have some GPU-work left, ran totally out of cpu-WUs.

Would be a nice gesture from the staff to post just a short notice of whats going on...

---

Aloha, Uli

Yep, one of my (work time only)-crunchers only have some GPU-work left, ran totally out of cpu-WUs.

Would be a nice gesture from the staff to post just a short notice of whats going on...

Matt's posted quite a long one:

Technical News : Spring Cleaning

I wonder if the creation rate could be related to file 20jn12ac as it has been in its current state for way longer than is usual.

Cheers.

A question for those doing AP WUs-
Have there been a lot more errors than usual over the last 6 hours or so?
Since the splitters have been cranking out more work, the number of AP results returned per hour has climbed considerably and is holding at the much higher than normal level.
Other than the occasional sharp spike, the number received per hour is generally around 4,000. It's presently over 7,000 & has been for several hours now.


EDIT- the average turnaround time is usually around 60hrs, presently it's down to 24.

---

Grant
Darwin NT

A question for those doing AP WUs-
Have there been a lot more errors than usual over the last 6 hours or so?
Since the splitters have been cranking out more work, the number of AP results returned per hour has climbed considerably and is holding at the much higher than normal level.
Other than the occasional sharp spike, the number received per hour is generally around 4,000. It's presently over 7,000 & has been for several hours now.


EDIT- the average turnaround time is usually around 60hrs, presently it's down to 24.

Linux OpenCL AP apps were released almost 11 hours ago, along with new Windows OpenCL ATI AP apps, and the Windows Hybrid AP app.

Claggy

A question for those doing AP WUs-
Have there been a lot more errors than usual over the last 6 hours or so?
Since the splitters have been cranking out more work, the number of AP results returned per hour has climbed considerably and is holding at the much higher than normal level.
Other than the occasional sharp spike, the number received per hour is generally around 4,000. It's presently over 7,000 & has been for several hours now.


EDIT- the average turnaround time is usually around 60hrs, presently it's down to 24.

Linux OpenCL AP apps were released almost 11 hours ago.

That might do it.
I'm thinking if that is the case then the number returned per hour should settle down at the higher level, but the turn around time should head back up to around what it was previously as their caches fill.

---

Grant
Darwin NT

Maybe this has something to do with it.

Maybe this has something to do with it.

Yes, the newly released Brook app for older cards is having issues.
One for Raistmer to investigate and debug.

Apps go through beta, but some problems only crop up with main's larger and more varied user base.

At least that one should be fairly easy to diagnose and fix, not like the old DLL problem we've so far failed to solve.

---

A person who won't read has no advantage over one who can't read. (Mark Twain)

A question for those doing AP WUs-
Have there been a lot more errors than usual over the last 6 hours or so?.....

Not here, AP errors on my 2 AP machines are 2 and 1 respectively, AP Inconclusives are also at "normal" levels (8 & 6)

T.A.

To lend some balance to the Multiverse:

Oh Waily Waily Waily! My Rac has gotten huger than
it's ever been "Recent average credit 31,490.38".
The Dread Lord is toying with my sanity!

Smashing Panic Button Now!

I have noticed I'm getting a lot of not _0 or _1 tasks in the past 24 hours.

Such as..

wu1266348384
wu1266363116
wu1266367979
wu1266475608
wu1266597583

Just to name a few. I would say they're all related.. looks to be all ATI errors.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)

Once again, the PFB splitters have fallen over.
Server status page shows green, but they're not producing any work (1.5/s doesn't count) even though there's plenty of data ready to be split. Ready-to-send buffer continues to shrink.

---

Grant
Darwin NT

I wonder if the creation rate could be related to file 20jn12ac as it has been in its current state for way longer than is usual.

I see that this file is still stuck in the splitters. :-(

Cheers.

Server status page frozen as of 1650 UTC; that's about 8 hours now.
Didn't anyone notice?

Waiting for the Marines....

---

Server status page frozen as of 1650 UTC; that's about 8 hours now.
Didn't anyone notice?

Waiting for the Marines....

That's strange, it says [As of 22 Jun 2013, 0:50:04 UTC] here.

Claggy

Server status updating normally at present.

---

"Time is simply the mechanism that keeps everything from happening all at once."

I wonder if the creation rate could be related to file 20jn12ac as it has been in its current state for way longer than is usual.

That file is still in SSP.

---

"Please keep Your signature under four lines so Internet traffic doesn't go up too much"

- In 1992 when I had my first e-mail address -

I wonder if the creation rate could be related to file 20jn12ac as it has been in its current state for way longer than is usual.

That file is still in SSP.

I sent a message to Eric asking about it, but have not received a response as of yet.

---

"Time is simply the mechanism that keeps everything from happening all at once."