I'll split one with ya:-)

So, now it's AP and PFB WUs then, eh?

I'd say it's still AP and MB WUs. The MB WUs are just split in a new (better) way.

Gruß,
Gundolf

OK, we'll still refer to them as MB WUs then.

Anybody notice if the change has any effect on crunch times on WUs of a similar angle range?

---

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

Hope someone gets this - my processed files are not uploading back to SETI. Is there something wrong?

shurstuff

You computer has not successfully connected to the servers for the last 18 hours.

Are there any clues in the Boinc event log? Is you internet connection OK?

---

There are only two kinds of WUs here, for now: astropulse_v6 and setiathome_enhanced. Whether the SaH Enhanced work is split by a process named mb_splitter or pfb_splitter makes no difference to how users' hosts process the work.

That splitting divides the 2.5 MHz. bandwidth captured by one beam/polarity of the multibeam recorder at Arecibo by 256, producing 256 subbands each 9765.625 Hz. wide. It is just like tuning a radio or TV to a particular station, an excellent radio will not be at all affected by adjacent channels but a cheap one will. And the excellent radio will have flatter response within the bandwidth assigned to the station. The polyphase filter approach is an improvement to flatness of response and out-of-band rejection for the Enhanced work.

When we start doing data recorded at GBT, and/or shift to larger WUs for Enhanced, the division will probably change, but the basic process will remain the same.

I think a pfb_splitter is also faster than a mb_splitter. That's mainly based on the creation times of WUs seen at Beta where only one splitter is usually active. But Eric may choose tape files for Beta which need little replacement of RFI by shaped noise. Either kind of splitter probably is slowed significantly when a lot of the data has to be replaced while it's being loaded from the tape file.

                                                                  Joe

Never mind, must have been traffic gettin in the way, downloads completed.

Been getting the occasional project backoff on downloads for a couple of weeks now, but haven't been able to spot any particular patterns.

---

Grant
Darwin NT

did someone say FREE PIZZA!!! AND BEER!!!

+1 !!! :)

+1, I'd keep the Pizza, everyone else can have the beer, as long as it lasts that is...

Same here. Pizza and artificially sweetened iced tea (no lemon) for me. Maybe diet orange soda, but not too much even of that.

Reminds me of an old Beetle Bailey strip. Beetle and Sarge go into a restaurant and Sarge asks what they serve. The waitress says, "Pizza and beer, beer and pizza, and pizza soaked in beer." Sarge says, "Gee, I don't know. They all sound so good."

On-topic: I noticed a couple of stuck uploads last week, about the same time someone else noticed some. I kicked them and they went right through. I don't look very often, but I haven't noticed any other problems lately.

---

David
Sitting on my butt while others boldly go,
Waiting for a message from a small furry creature from Alpha Centauri.

Never mind, must have been traffic gettin in the way, downloads completed.

Been getting the occasional project backoff on downloads for a couple of weeks now, but haven't been able to spot any particular patterns.

I got an AP assigned last night and was watching the transfers tab and it took over 30 seconds for it to actually start receiving data, and when it did, I received it at 3MB/sec. I was actually waiting for it to time-out and retry, but then it started, finally.

---

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

Never mind, must have been traffic gettin in the way, downloads completed.

Been getting the occasional project backoff on downloads for a couple of weeks now, but haven't been able to spot any particular patterns.

I got an AP assigned last night and was watching the transfers tab and it took over 30 seconds for it to actually start receiving data, and when it did, I received it at 3MB/sec. I was actually waiting for it to time-out and retry, but then it started, finally.

Well, thankfully, that's the exception rather than the rule these days.
When I happen to be checking one of my rigs and some new tasks come through, I have to be pretty fast to click the downloads tab to watch them stream in.

---

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

I have noticed over the past few days that all my machines have one work unit that is in retry. When I hit the retry button they get downloaded fast. Maybe they go so fast now the server just loses track:)

---

[/quote]

Old James

Looks like some more shorties going through the system. And once again the splitters are unable to pick up the pace. Ready-to-send buffer rapidly shrinking.

---

Grant
Darwin NT

looking at the graphs of result creation for week there is a definate

dropoff on 26-apr when I first noticed all the mb_splitters were renamed pfb_splitter. Not a major dropoff but noticible.

I got an AP assigned last night and was watching the transfers tab and it took over 30 seconds for it to actually start receiving data, and when it did, I received it at 3MB/sec. I was actually waiting for it to time-out and retry, but then it started, finally.

The delay you notice in downloading is not specific to AP, it has been happening to MB as well. I first noticed it about two weeks ago.

---

time to PANIC

downloads have stopped!!!

---

rig needed a reboot. all good now:D

time to PANIC

downloads have stopped!!!

---

I got an AP assigned last night and was watching the transfers tab and it took over 30 seconds for it to actually start receiving data, and when it did, I received it at 3MB/sec. I was actually waiting for it to time-out and retry, but then it started, finally.

The delay you notice in downloading is not specific to AP, it has been happening to MB as well. I first noticed it about two weeks ago.

I figured it applied to everything, but AP is more noticeable to see what happens since the WU is much larger and takes longer to download. Pretty much one update interval for MB and the WU is gone, but at least with AP, it takes 3-5 UI update intervals before it finishes.

---

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

Almost out of CPU work. Most requests for work result in "Project has no tasks available" messages.
The splitters need fixing.

---

Grant
Darwin NT

same here

---

+ 1, the splitters our next bottleneck?

---

Maybe we are drinking the wrong bottles of Beer? Too many bottlenecks?

If we switch to Cans of Beer would that help?

Yes, Now switching to beer can, a 24 pack is a good start, I realy belive i need some, all my hosts are running out of GPU WU.

---