~misfit~ wrote:

Alan Woodford wrote:

On Sat, 10 Jul 2004 17:04:00 GMT, Klaatu
<mutster8_nospam_@netscape.net> wrote:

Have a 2 to 3 day cache set for Boinc.  Noticed the cached WU's
start with about a 2 days supply but drops down to just one WU one
before the cache is built back up to 2 days worth.  Is anyone else
seeing this?  Is there an explanation or is it just a Boinc start up
issue, not enough units available?  Seems like the number of cached
units should remain relatively steady at whatever limits have been
set.  Didn't use a cache for classic so don't have any comparitive
info.  TIA

Not sure if it's the same problem, but...

My caches are filling with small numbers of units, with 26-28 hour
completion times (before they run), which are actually only taking 4-5
hours at most to process.
Both my machines have run the benchmarks in the last day or so, so
maybe something at Berkeley has gone awry?

Interesting you should mention this. I just noticed on a Celery 900 I'm
using that the cache looked a little empty. Then I noticed that one WU has
a 'time to completion' of 50 hours! The rest are 7 hours. That machine is
only turned on now and then so I have the cache set for 1-2 days. That one
unit is more than two days. LOL.
--
~misfit~

Trying Boinc on an old 700 AMD.  Normally starts out with 5 WU's, 4 ready
and one running.  Average WU time about 10 hours.  At the low point I have
just one WU ready ond one running.  Then Boinc goes back to 5 WU's.  Same
behavior last several days.  Running linux CLI with Boincgui.  If the WU's
are about the same length would expect to see 4 ready most of the time???