timeout on time linux cluster

I just happily discovered the "time" command on a linux cluster (see URL) for the computing
for life sciences cluster at san francisco state university.  i am doing some experiments running
a script called hybrid-select.pl that finds secondary structure of RNA but i will stop boring you
with those details.  in some literature, these types of folding algorithms are described as
being O(N^3) so i wanted to test this data.  the problem is, that the CCLS cluster is very
intolerant of terminals running for long enough for me to test my thoughts.  therefore i
tried to write a script that would save the output of time, to no avail.

here is a little illustrative unix command line output

[::begin clip::]

[kayve@ccls wet]$ ./first_test.bash &
[1] 644
[kayve@ccls wet]$
real    0m29.371s
user    0m29.222s
sys     0m0.073s

[1]+  Done                    ./first_test.bash
[kayve@ccls wet]$ cat data
cat: data: No such file or directory
[kayve@ccls wet]$ cat rundata
[kayve@ccls wet]$ cat data17804
seq17804.804 runtime data:

[kayve@ccls wet]$ ./first_test.bash &
[1] 6760
[kayve@ccls wet]$
real    0m29.631s
user    0m29.333s
sys     0m0.192s
Read from remote host ccls.research.sfsu.edu: Connection reset by peer
Connection to ccls.research.sfsu.edu closed.
[kayve@kv_bsd ~]$

[::end clip::]

i am trying to save that time data you see going to standard output above for a sequence
that has 804 base pairs (N=804) running for half a minute in real time. my goal is to push the
limit on a larger sequence to test the O(N^3) nature of the algorithm empircally.  this is
not possible with the way my terminals timeout.  i tried redirecting to stderr but that didn't
work.  here is another command line clip of my two scripts.  the filename convention i have
for the datafiles i generate from a genome wide 5'-UTR file with a total of over 32 000 sequences
is "seq(index).(length) so that below you will se my script running the hybrid-select.pl
on a filename seq17804.804 implying the data was taken from the 17804th sequence of the
32K sequences and the second script seq17611.7327 implying that the 17611th sequence
has 7327 base pairs.

[::clip::]

[kayve@ccls wet]$ cat first_test.bash ; cat time_test.bash
#! /bin/bash

time 2>rundata  hybrid-select.pl seq17804.804 > /dev/null
echo "seq17804.804 runtime data:" > data17804
cat rundata >> data17804
echo "${ran4}" >> data17804

#! /bin/bash

ran4=`time hybrid-select.pl seq17611.7327 > /dev/null`
echo "seq17611.7327 runtime data:" > data17611
echo "${ran4}" >> data17611

[kayve@ccls wet]$


[::end clip::]

an aside.. the >/dev/null is for getting rid of the output of the hybrid-select.pl script.