Project

General

Profile

Bug #106

trajectory searching in 3.3.1

Added by Jochen Hub about 13 years ago. Updated about 13 years ago.

Status:
Closed
Priority:
High
Assignee:
Category:
mdrun
Target version:
Affected version - extra info:
Affected version:
Difficulty:
uncategorized
Close

Description

Hi,

searching for the first frame of a trajectory given with -b seems to be still
buggy. The trajecory test.xtc was created with trjconv 3.3.1 from a trajectory
created by mdrun 3.3. This was no problem. Now, however

trjconv -b 200 -f test.xtc -o x.xtc

cannot find the first frame:

Skipping frame 0 time 199.960

is printed again and again (which can be seen when sending output to a file).
This happens when using trjconv 3.3.1 and 3.3.

Regards,
Jochen

test.xtc (11.5 MB) test.xtc xtc file Jochen Hub, 09/18/2006 04:20 PM
test.pdb (6.46 KB) test.pdb corresponding structure Jochen Hub, 09/18/2006 04:22 PM
libxdrf-3.3.diff (943 Bytes) libxdrf-3.3.diff Patch for Gromacs-3.3 Filipe Maia, 09/28/2006 12:23 PM
libxdrf-3.3.1.diff (380 Bytes) libxdrf-3.3.1.diff Patch for Gromacs-3.3.1 Filipe Maia, 09/28/2006 12:24 PM
libxdrf-3.3.1.diff (7.99 KB) libxdrf-3.3.1.diff Updated patch to 3.3.1 Filipe Maia, 10/08/2006 02:56 AM
libxdrf-3.3.diff (9.86 KB) libxdrf-3.3.diff Updated patch to 3.3 Filipe Maia, 10/08/2006 03:11 AM

History

#1 Updated by Jochen Hub about 13 years ago

Created an attachment (id=74)
xtc file

#2 Updated by Jochen Hub about 13 years ago

Created an attachment (id=75)
corresponding structure

#3 Updated by Juergen Haas about 13 years ago

(In reply to comment #0)

Hi,

searching for the first frame of a trajectory given with -b seems to be still
buggy. The trajecory test.xtc was created with trjconv 3.3.1 from a trajectory
created by mdrun 3.3. This was no problem. Now, however

trjconv -b 200 -f test.xtc -o x.xtc

cannot find the first frame:

Skipping frame 0 time 199.960

is printed again and again (which can be seen when sending output to a file).
This happens when using trjconv 3.3.1 and 3.3.

Regards,
Jochen

Since this happens very often (almost every trajectory at least once) maybe we
could introduce a switch -searchM321 to allow searching frames with the old
style algorithm as a fallback option.
There might be a relationship between this error and #15/#36 ...

Juergen

#4 Updated by Filipe Maia about 13 years ago

Bug happened when the time seek binary searched ended up in a frame boundary.
Fixed in CVS.
Patches for 3.3 and 3.3.1 follow.

#5 Updated by Filipe Maia about 13 years ago

Created an attachment (id=76)
Patch for Gromacs-3.3

#6 Updated by Filipe Maia about 13 years ago

Created an attachment (id=77)
Patch for Gromacs-3.3.1

#7 Updated by Filipe Maia about 13 years ago

Created an attachment (id=94)
Updated patch

Fixes problems with magic numbers showing up outside of the header (which I
didn't know was possible).

#8 Updated by Filipe Maia about 13 years ago

Created an attachment (id=95)
Updated patch to 3.3

Also available in: Atom PDF