On my WIN2K machine running SETI 4.09 & BOINC 4.25, half of my results get client error - taking much more CPU that others. The following show on my results:

54253635 13208686 18 Apr 2005 13:55:21 UTC 25 Apr 2005 7:37:25 UTC Over Client error Computing 81,968.23 70.69 ---

This is what shows on my messages under BOINC:
25/04/2005 10:30:20|SETI@home|Computation for result 23ja05aa.3021.10433.117328.79 finished
25/04/2005 10:30:20|SETI@home|Output file 23ja05aa.3021.10433.117328.79_3_0 for result 23ja05aa.3021.10433.117328.79_3 exceeds size limit.
25/04/2005 10:30:20|SETI@home|File size: 138621.000000 bytes. Limit: 65536.000000 bytes
25/04/2005 10:30:21|SETI@home|Unrecoverable error for result 23ja05aa.3021.10433.117328.79_3 ( 23ja05aa.3021.10433.117328.79_3_0 -131 )

It is very frustrating - literally half my results are in error, and they take up most of my donated SETI time, with WU taking over 24 hours (then thrown in the trash) compared to the successful ones taking 9 hours or so.

---

BESTEST!

Check the General Preferences on 'Your Account' page. Under 'Disk and Memory Usage' are three settings that control how much disk BOINC can use.

Other than those, do you have disk quotas enabled on youe disks?

---

> Check the General Preferences on 'Your Account' page. Under 'Disk and Memory
> Usage' are three settings that control how much disk BOINC can use.
>
> Other than those, do you have disk quotas enabled on youe disks?
>
>
>
Thanks! My settings were OK (I think) - .5 GB, leaving 1 GB and 10% - I've upped them to 1,1 & 25%. There is plenty of disk space on the machine for this (1.62 GB free out of a partition of 8), and no restrictions.

---

BESTEST!

I've had a lot of noisy Work Units lately. Noisy Work Units result in large result files, hence Seti will bail out on them sooner or later.

I'm just crunching Seti & having 50MB is more than enough for 4 days worth of work & that still leaves at least 20MB of unused BOINC allocated space.

---

Grant
Darwin NT

> >
> Thanks! My settings were OK (I think) - .5 GB, leaving 1 GB and 10% - I've
> upped them to 1,1 & 25%. There is plenty of disk space on the machine for
> this (1.62 GB free out of a partition of 8), and no restrictions.
>

Hope it helps... but I'm not convinced it will :(
The settings you had in your preferences sound like they were the defaults (don't remember exactly what the defaults should be though). These should have been sufficient for normal use.

On further thought...

The error you are getting seems to indicate that the output file that it is trying to create is 138621 bytes long. This seems very large for a result output. If this is the file that gets returned to Berkeley, I would expect it to be more in the range 8000 - 20000 bytes or so.

If I am correct, then it seems that your system is generating too much reportable data. There is another error that reports too many results for the available storage, which happens when you get a 'noisy' WU - full of spurious signals - but you're not getting this detected.

I guess it's possible that your system has some type of instability or memory/CPU fault. What is your hardware? Is it overclocked at all?

---

>
> I've had a lot of noisy Work Units lately. Noisy Work Units result in large
> result files, hence Seti will bail out on them sooner or later.
>
> I'm just crunching Seti & having 50MB is more than enough for 4 days worth
> of work & that still leaves at least 20MB of unused BOINC allocated space.
>
Thanks - but I should have mentioned that others who had the same WUs succeeded and had a fraction of my CPU time - something is causing my "puter" not to compute correctly!
For instance, with my last WU, three others succeeded two with 12,000 plus CPU and one with 15,000 plus and I had almost 82,000!!

---

BESTEST!

> > >
> > Thanks! My settings were OK (I think) - .5 GB, leaving 1 GB and 10% -
> I've
> > upped them to 1,1 & 25%. There is plenty of disk space on the machine
> for
> > this (1.62 GB free out of a partition of 8), and no restrictions.
> >
>
> Hope it helps... but I'm not convinced it will :(
> The settings you had in your preferences sound like they were the defaults
> (don't remember exactly what the defaults should be though). These should have
> been sufficient for normal use.
>
> On further thought...
>
> The error you are getting seems to indicate that the output file that it is
> trying to create is 138621 bytes long. This seems very large for a result
> output. If this is the file that gets returned to Berkeley, I would expect it
> to be more in the range 8000 - 20000 bytes or so.
>
> If I am correct, then it seems that your system is generating too much
> reportable data. There is another error that reports too many results for the
> available storage, which happens when you get a 'noisy' WU - full of spurious
> signals - but you're not getting this detected.
>
> I guess it's possible that your system has some type of instability or
> memory/CPU fault. What is your hardware? Is it overclocked at all?
>
>
>
You are being very helpful- thanks!
As per my recent answer to another helpful individual, it does not appear to be a noisy WU - however your second point deserves some thought.
The machine is an old well banged up Celeron think pad with an interesting history running W2KSP5 - with its original ME still extant on another partition. It is setup with regular keyboard, monitor, etc. and used primarily by my youngest - an active 17 year old - so who knows what else is running on it, although I keep it current and have the latest anti-virus, firewall and anti-spyware running. However it is stable, and has no other program problems, other than its BOINC screesaver wait-time being reset intermitently (see another thread).

---

BESTEST!