What are you thinking sending ASTROPULSE TO CRUNCH THAT TAKES 240 HOURS TO DO!??

LBJ

---

What are you thinking sending ASTROPULSE TO CRUNCH THAT TAKES 240 HOURS TO DO!??

LBJ

Why is it taking that long for you? My WU Astropulses take about 40-50 hrs.

There is also an optimized app.
Look in number crunching forums.

http://setiathome.berkeley.edu/forum_thread.php?id=31810

---

With each crime and every kindness we birth our future.

Looks like that is a v5 unit, so there is no optimized app yet.

---

What are you thinking sending ASTROPULSE TO CRUNCH THAT TAKES 240 HOURS TO DO!??

LBJ

You won't know how long it takes until it finishes. The initial estimate is an estimate.

---

Mine took 12.5 hours on a E4300 @ 2.4GHz and received 750 credits: http://setiathome.berkeley.edu/workunit.php?wuid=394288523. It did estimate the same 240 hours when it came.

Mine took 12.5 hours on a E4300 @ 2.4GHz and received 750 credits: http://setiathome.berkeley.edu/workunit.php?wuid=394288523. It did estimate the same 240 hours when it came.

We think he is probably talking about the new V5.03 tasks, which if the devs are correct will take 50% longer to process. And there is no optimised app, yet.

If you were to remove the optimised AP app and grab one of these new AP tasks, with the V5.03 app, it would probably take your computer ~75 hrs to complete.

1) Longer AP run will give bigger credits (at least it intended to be so). There are another projects with very long tasks too, no problem in just that fact.
2) Longer run increase probability of fault that can kill whole task, but shortening of task is impossible. Primary goal is to do more "science" with task and shortening task will reduce sensitivity.
3) Opt app already in RC1 stage, we just wait online validations. App passed already offline tests. Its speed will not very longer than prev release (although it will return much more valuable results than before).

Summary:
If you interesting in project goal try to get as stable host config as possible to be sure not to losing partially completed work.
In other case just set "no AP work" in web preferences.

I received a V5 AP. when i looked at it it said 131 hours to completion. 51 hours later it says 128 hours to completion and 29% done. Im running a n old P4 XP. depending on what happens when i get credit or not. ( client error on one of the wingmen, thats why i got it.) 51 hours for just 3 hours work it will take me a long time to crunch this one out . I might have to click on the no AP box. I didnt mind when the old AP took 4 or 5 days. but i dont want to risk getting a client error with that many hours invested in one WU either.

---

[/quote]

Old James

I received a V5 AP. when i looked at it it said 131 hours to completion. 51 hours later it says 128 hours to completion and 29% done. Im running a n old P4 XP. depending on what happens when i get credit or not. ( client error on one of the wingmen, thats why i got it.) 51 hours for just 3 hours work it will take me a long time to crunch this one out . I might have to click on the no AP box. I didnt mind when the old AP took 4 or 5 days. but i dont want to risk getting a client error with that many hours invested in one WU either.

The 51 hours and 29% progress forecast a total run around 175 hours so about 124 to go, the additional processing to look for repetitive pulses with negative dispersion has increased the crunch time about that much. The BOINC estimates use the Duration Correction Factor which currently apply to all applications for a project, so has adapted to the S@H Enhanced work the host has been doing.

By the time the host finishes that WU, there should be an optimized 5.03 available which would do the next WU in under 2 days on your host, you might consider that rather than not doing AP work in future.

                                                                Joe

Well, that's a lot. I'm crunching one (v.5.00) and the estimate is 80h. Right now, with nearly 4h i'm on 10%.
I'll finish this one to see how much credit will be granted. If not worth i'll stop crunching astropulse

Intel(R) Core(TM)2 Duo CPU T8300 @ 2.40GHz w/ 4Gb

---

With so big crunching time it's even more important to optimize host performance by installing opt versions of apps.
Moreover, opt AP app has additional measures to prevent work lose due to client restarts and OS crashes.

When will the opt app for v5 AP be available, it has taken me about 55hrs to crunch a v5 AP using stock

---

When will the opt app for v5 AP be available, it has taken me about 55hrs to crunch a v5 AP using stock

Cuirrently release delayed only by our stock app wingmans - still to few validations done for release.

Typical crunch time on my host (Q9450) 10-11 hours...

When will the opt app for v5 AP be available, it has taken me about 55hrs to crunch a v5 AP using stock

Cuirrently release delayed only by our stock app wingmans - still to few validations done for release.

Typical crunch time on my host (Q9450) 10-11 hours...

Thanks Raistmer for reply, maybe my Q6600 at 3.2Ghz will have similar time, that would be a result.

---

My AMD 940 knocked out about 20 AP's at around 11 hours each using the AP Op app.

---

In a rich man's house there is no place to spit but his face.
Diogenes Of Sinope

Among the 1000 of enhanced 6.08 I got 1 ap_v5 5.03 wu...
duration:
~14h 26min = 52,007 sec
415864032
still pending...
if we have more of this wu's validating against stock, the release can be done soon.
But at the moment we have to wait.

Among the 1000 of enhanced 6.08 I got 1 ap_v5 5.03 wu...
duration:
~14h 26min = 52,007 sec
415864032
still pending...
if we have more of this wu's validating against stock, the release can be done soon.
But at the moment we have to wait.

If that is the new typical credit claim, I think I'm in love..

---

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

Among the 1000 of enhanced 6.08 I got 1 ap_v5 5.03 wu...
duration:
~14h 26min = 52,007 sec
415864032
still pending...
if we have more of this wu's validating against stock, the release can be done soon.
But at the moment we have to wait.

If that is the new typical credit claim, I think I'm in love..

That 1,224.80 has actually come down by about 70 since the earliest results, claims may stabilize below 1200. There's probably too few results for the server-side credit multiplier to have a good median, though, so as some slower hosts finish it might grow again.

                                                                 Joe

Im at 95% with 12 hours to go on my V5 AP. Im running a P4 XP with stock version 5.2.13. I will upgrade to version 6.4.5 when this AP is done. Id upgrade to an op version for AP but i dont think i have the skill to do it.

---

[/quote]

Old James