imshift on different platforms
Guest wrote on Nov 29, 2005
Greetings,
I have noticed some irregularities in running the iraf task "imshift"
after upgrading my desktop, running rh9, from PC-IRAF V2.12.1-EXPORT to
V2.12.2a-EXPORT. Using the same input parameters within "imshift" on
the same input file, the output file produced after this upgrade is
noticably different. (the input file is shifted differently) I have
reinstalled V2.12.1 back onto this desktop, and can reproduce my
original results (different than that produced on the V2.12.2a version).
I have further explored running this task with the same input parameters
(on the same input file) on different versions of iraf loaded on
different machines...and have found:
a) rh9 machine running v2.12.1-export; sun machine running v2.11.3; IRIX
machine running v2.11.3 all yield the same results
b) rh9 machine running v2.12.2a-export; fedora core 2 machine running
v2.12.2a-export yield the same results...which are different than that
found in a)
Have other users reported this anomaly when using v2.12.2a - and do you
have any idea what may be causing the problems described?
Thanks in advance for your help,
John Wisniewski
Post generated using Mail2Forum (http://www.mail2forum.com)
I have noticed some irregularities in running the iraf task "imshift"
after upgrading my desktop, running rh9, from PC-IRAF V2.12.1-EXPORT to
V2.12.2a-EXPORT. Using the same input parameters within "imshift" on
the same input file, the output file produced after this upgrade is
noticably different. (the input file is shifted differently) I have
reinstalled V2.12.1 back onto this desktop, and can reproduce my
original results (different than that produced on the V2.12.2a version).
I have further explored running this task with the same input parameters
(on the same input file) on different versions of iraf loaded on
different machines...and have found:
a) rh9 machine running v2.12.1-export; sun machine running v2.11.3; IRIX
machine running v2.11.3 all yield the same results
b) rh9 machine running v2.12.2a-export; fedora core 2 machine running
v2.12.2a-export yield the same results...which are different than that
found in a)
Have other users reported this anomaly when using v2.12.2a - and do you
have any idea what may be causing the problems described?
Thanks in advance for your help,
John Wisniewski
Post generated using Mail2Forum (http://www.mail2forum.com)
Guest wrote on Nov 29, 2005
Hi John,
What's the definition of "different" in this case? Following the
V2.12.1 release the task was modified to use double precision to fix a
bug where the specified shift was less than the precision of a real, but
otherwise the task is unchanged. Are the differences within the machine
precision or are we talking about major differences?
Regards,
Mike Fitzpatrick
Post generated using Mail2Forum (http://www.mail2forum.com)
What's the definition of "different" in this case? Following the
V2.12.1 release the task was modified to use double precision to fix a
bug where the specified shift was less than the precision of a real, but
otherwise the task is unchanged. Are the differences within the machine
precision or are we talking about major differences?
Regards,
Mike Fitzpatrick
Post generated using Mail2Forum (http://www.mail2forum.com)
Last post on Nov 29, 2005