For the past several weeks I've had increasing numbers of work units which complete in less than 1 minute. As I write this, I have 16 work units 'ready to report'; 7 of these units completed in less than a minute.

Not a problem... I'm just curious whether others are seeing this and why it is happening.

---

<img src="http://www.setisynergy.com/images/stats/comb-130.jpg"></img>

You have two wu's on two different Machines...

The work Units were due in Semptember of 2004!

I think you missed the dealines....

---

Check the stderrout messages for those WU.

---

You have two wu's on two different Machines...

The work Units were due in Semptember of 2004!

I think you missed the dealines....

This answer does not relate to his question. Those are obviously some "lost" workunits that most of us that have been around since back then have on our computers and he is unable to delete those computers until they are gone.

For the past several weeks I've had increasing numbers of work units which complete in less than 1 minute. As I write this, I have 16 work units 'ready to report'; 7 of these units completed in less than a minute.

Not a problem... I'm just curious whether others are seeing this and why it is happening.

These are WU's with the fameaous "SETI@Home Informational message -9 result_overflow" message.
When a WU is crunched and "The number of results detected exceeds the storage space allocated" you get this kind of short CPU time.
These are the so called noisy WU's which can contain for example radio signals from earth source. AFAIK S@H Classic would crunch them up to the end but S@H/Boinc stops crunching them at a certain time.

These are the so called noisy WU's which can contain for example radio signals from earth source. AFAIK S@H Classic would crunch them up to the end but S@H/Boinc stops crunching them at a certain time.

That's the information I was looking for. Thanks.

---

<img src="http://www.setisynergy.com/images/stats/comb-130.jpg"></img>

S@H Classic would crunch them up to the end but S@H/Boinc stops crunching them at a certain time.

Classic did the same thing... that was one of the several ways people could manage to get higher WU counts ...

The dodges included (assuming you are interested)
1) WU Cloneing where partially processed WU were "shared" and completed on different machines and in some cases two people would share with each other also

2) Cherry selection, dropping WU with the "wrong" angle range and doing more of the "faster" work units, and trying to get more -9 types that could be processed faster

3) Returning the same WU multiple times

Countered by BOINC Assigning and tracking issued work so you cannot "share".

---