Re: problems running Structure with XP GUI?
- From: "Perplexed in Peoria" <jimmenegay@xxxxxxxxxxxxx>
- Date: Tue, 25 Jul 2006 16:59:15 -0400 (EDT)
"Perplexed in Peoria" <jimmenegay@xxxxxxxxxxxxx> wrote in message news:...
"Malcolm" <regniztar@xxxxxxxxxxxxxx> wrote in message news:ea1cf0$1g0u$1@xxxxxxxxxxxxxxxxxxxxxx
"evoeco@utexas" <johoma@xxxxxxxxx> wrote in message
news:e9s467$213t$1@xxxxxxxxxxxxxxxxxxxxxx
I've been running the latest build of Structure with a Windows XP GUII don't know about the specific program but I've a lot of experience with
on microsatellite data, but have stumbled across some weird issues:
1. Immediately after the initial install, the program seemed to run
well: 20k burn-in, followed by another 20k. Up to about 350 individuals
across 12 diploid loci, using five iterations per K.
2. The second and later jobs developed problems. On most of the
machines, K = 5 generated unknown errors in Windows, causing these runs
to be skipped. K = 6 ran OK, but no higher K values worked at all. On a
less-powerful Dell, these errors began with K = 1. With the most
powerful Dell, Structure quit at K = 5. I restarted the program and
reloaded the job, beginning the job at K = 5, which ran well. K = 6
went OK too, then the program crashed completely again on K = 7. I
again restarted the program and job, and it seemed to run OK from 7 on.
3. I've tried reinstalling and rebooting numerous times. The errors are
consistent and persistent.
Any ideas what's going on?
these types of problems.
Really you need the source of the program and a compiler, plus a programmer,
to get to the root of the malfunctions.
Assuming you don't have this, run a trivial job that will execute correctly.
Then systematically make it more complicated and more simple again until you
can nail down the exact conditions in which it fails. This may reveal the
problem. If not, you have something concrete to give to the program
designers.
My guess is that you are running out of some kind of resource; most
likely real memory, virtual memory, "handles", or work disk space.
You can check disk space availability by periodically checking
"properties' of the relevant drive using Windows Explorer. The
others can be monitored using TaskManager.
Incidentally, if you are running in a network configuration in which
you have your 'C:' drive mapped to a Lotus network drive, then that
by itself might explain your problem. Lotus, for some reason, doesn't
completely release closed work files until the process creating them
terminates. At least that was the case 4-5 years ago. Our work-file
intensive program worked fine if the C: drive was local, but some
customers ran into problems because their network policy was to put
the C: drives on the network. IIRC, the workaround was simply to
set the TEMP environment variable to point to a local hard drive when
running our program.
.
- Prev by Date: Re: UV Damage to DNA paper - Aug'05
- Next by Date: Re: On Evolution
- Previous by thread: Re: problems running Structure with XP GUI?
- Next by thread: Article: RNA interference is inherited
- Index(es):
Relevant Pages
|
