After cycling through my nodes something odd caught my eye. On one machine I noticed under the work tab that one "fullproduction" WU was reported as "preempted". I also noticed that it had skipped half-way through the queue and started 2 new wu's.

I havent suspended or abort anything as everything seems to be running, but I thought it very odd that it would skip 30 or so queued WU's and start a new one, ignoring the older.

Is this due to some project priority or emphasis that has been placed on these newer wu's (fullatom blah blah). Is this to be considered normal or has something run amok?

TIA.

Sigma

---

It may have had an earlier deadline. The queue is not processed FIFO it is processed by the earliest deadline. Even in round robin mode this effects which result from a particular project is processed.

---

BOINC WIKI

BOINCing since 2002/12/8

Interesting KK. Thanks for the info. I'm seeing this behavior now on many nodes. It seems that the newer WU's being distributed have much shorter deadlines than those given out just a couple of days ago. I still find it a bit odd that it would preempt a WU currently being processed.

Oh well...Que sera, sera. I wont waste any more cycles on this topic then. :D

Thanks again for the info.

Sigma

There is a thread all about shorter deadlines ---


I've just noticed that all WUs we've downloaded in the last 24 hours have 7 day deadlines on them, but when I looked last year, WUs had 4 week (or 1 month) deadlines.

Is this a permanent change?

It's not a big deal for us in the most part, except on a couple of PCs which are only switched on a few hours a day.

---

I think what you are seeing is a result of trying to sort out the bandwidth problem. The WU`s appear to be the same size (MB wise) but taking 4 times as long as they are doing 4 x the amount of work on one machine. Hence it would also follow that the work would be needed in 1/4 of the time scale.

I maybe wrong !

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

I think what you are seeing is a result of trying to sort out the bandwidth problem. The WU`s appear to be the same size (MB wise) but taking 4 times as long as they are doing 4 x the amount of work on one machine. Hence it would also follow that the work would be needed in 1/4 of the time scale.

I maybe wrong !

I'm seeing the short deadline ones take about half the time. Driving the DCT way down.

---

Proudly Banned from Predictator@Home and now Cosmology@home as well. Added SETI to the list today. Temporary ban only - so need to work harder :)



"You can't fix stupid" (Ron White)

I`ll check my other machines but on here I have all 6 hour plus estimates and 21-02 deadlines.

It maybe that they had to get rid of the old before the new type came in. I`m guessing.

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

I`ll check my other machines but on here I have all 6 hour plus estimates and 21-02 deadlines.

It maybe that they had to get rid of the old before the new type came in. I`m guessing.

Remember that the time to completion (estimate)that you see in BOINC Manager is adjusted by the DCT for that particular PC. You can see the DCT on the "My Computer" screen.

---

Proudly Banned from Predictator@Home and now Cosmology@home as well. Added SETI to the list today. Temporary ban only - so need to work harder :)



"You can't fix stupid" (Ron White)

I`ve a DCT ? ( Duration Correction Factor ) of 3.44 on this but checking the time on the current unit it seem`s quite accurate to the estimate.

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

I`ve a DCT ? ( Duration Correction Factor ) of 3.44 on this but checking the time on the current unit it seem`s quite accurate to the estimate.

Dude! If your DCT is 3.44, you've got a seriously slow computer! That means on average that the PC is taking 3.44 times longer to crunch the WU than the Rosetta estimate. I hope for your sake that you got the wrong number!

---

Proudly Banned from Predictator@Home and now Cosmology@home as well. Added SETI to the list today. Temporary ban only - so need to work harder :)



"You can't fix stupid" (Ron White)

Goes off to check cos this aint slow.....it`s a 3GHz P4 Northwood OC`ed to 3300MHz......

There go...

Result duration correction factor 3.109464


But take into account I do not run 24/7 .....about 16 hours a day would that make a difference ?

Saying that my two AMD`s are about 1.9 correction factor !!

If the factoring is taking into account my shutdown time I can see this would alter it as it may appear to take 11 hours over 1 normal 3 hour WU every night.

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

Goes off to check cos this aint slow.....it`s a 3GHz P4 Northwood OC`ed to 3300MHz......

There go...

Result duration correction factor 3.109464


But take into account I do not run 24/7 .....about 16 hours a day would that make a difference ?

Saying that my two AMD`s are about 1.9 correction factor !!

If the factoring is taking into account my shutdown time I can see this would alter it as it may appear to take 11 hours over 1 normal 3 hour WU every night.

That's weird - I thought it was based on CPU time, not wall clock hours.

Time to dive into the WIKI

---

Proudly Banned from Predictator@Home and now Cosmology@home as well. Added SETI to the list today. Temporary ban only - so need to work harder :)



"You can't fix stupid" (Ron White)

It is wierd cos if true then the two AMD`s should be similar...they are now down to 1.4 DCF.....strange.....

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

P4 3GHz

% of time BOINC client is running 71.8565 %
While BOINC running, % of time host has an Internet connection 93.0627 %
While BOINC running, % of time work is allowed 99.9833 %
Average CPU efficiency 0.909067
Result duration correction factor 3.109464


AMD 64 3000+

% of time BOINC client is running 80.228 %
While BOINC running, % of time host has an Internet connection 76.6875 %
While BOINC running, % of time work is allowed 99.9897 %
Average CPU efficiency 0.871907
Result duration correction factor 1.512877

AMD 64 3000+

% of time BOINC client is running 67.8654 %
While BOINC running, % of time host has an Internet connection 75.4095 %
While BOINC running, % of time work is allowed 99.989 %
Average CPU efficiency 0.964855
Result duration correction factor 1.498211

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

The varying DCFs are most likely due to the mix of work that specific host has had. Plus the DCF rises quickly but drops more slowly. One underestimated workunit can change the as much or more than 10 underestimated workunits.

---

BOINC WIKI

BOINCing since 2002/12/8

Thanks Keck...

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.