QuteMol: Trobleshooting Guide

Problem: QuteMol refuses to start!

Most probably, the video card is not up to the task.
The most advanced feature required by QuteMol is support for FrameBufferObjects (FBO).
You can check this list of video card models that are supposed to support FBO.
Sometimes, the card is capable of supporting FBO but its driver it is not. Updating the video drivers could fix the problem.
No, we don't plan to implement a fall back strategy for machines without FBO support: video card capabilities are increasing fast, so we think it is not worth it (at least not with an high priority).


Problem: the "custom" sliders occasinally stop responding

On same older card, this problem is related to the maximum size of the fragment program, so when the combination of settings results in a too complex per-pixel processing, it will just stop working. This has been reported on a ATI Radeon 9800 PRO, and there is no solution yet.


Problem: an error message occurs when I try to save a snapshot

On same platform, this problem is related to the maximum size of the screen buffer. You can edit manually the file "qutemol.cfg" (which is created at the 1st execution in the same directory where you installed qutemol) and set SNAP_SIZE to 512. Restart QuteMol for changes to take place.
However, screenshots will be smaller.


Problem: the "global illumination" term looks totally wrong!

There is a bug on some NVidia card per-fragment program. Newest generation cards and drivers do not show this, and sometimes the bugs heals updating video drivers. However for some model that is not the case.
Here is how the "porin" molecule (included in the installer) can look like as a result of this bugs.
To fix this, you can edit the "qutemol.cfg" (which is created at the 1st execution in the same directory where you installed qutemol) and set "doubleSM" to 0 and "NVIDIA_PATCH" to 1. This somewhat worsen both computation times and rendering quality (as far as ambient occlusion is concerned), so don't do this unless needed to fix the bug.
Restart QuteMol for changes to take place.
This BUG has been reported on some 6xxx generation NVidia cards.


Problem: on larger molecules, the "global illumination" terms seems somewhat wrong.

When a molecule exceeds some 32K atoms, the texture required to achieve the same quality exceeds the default max texture size: the effect is that the molecule looks a lot darker in some parts, as if lighted from a direction.
Currently, this must be fixed manually: edit the file "qutemol.cfg" (which is created at the 1st execution in the same directory where you installed qutemol) and set "TSIZE" (1024 by default) to 2048 (however, this could slow down AO computation on some machine). Restart QuteMol for changes to take place. Unpdate from version 0,3 on, texture size is automatically increase by the system when the size of the molecule requires it; if you encounter problem you can lower the max texture size "MAX_TSIZE".
This will hopefully be automatic on newer version.


Problem: the "global illumination" computation is too slow / the frame rate is too slow!

On older machines, you might want to trade speed for visual quality. You can do that by changing the first 6 parameters inside the file "qutemol.cfg" (which is created at the 1st execution in the same directory where you installed qutemol).
Restart QuteMol for changes to take place.
As a general rule, larger values means higher quality rendering, and smaller values make for faster rendering times. All values but "N_VIEW_DIR" must be powers of 2. If you want to undo your changes, simply delete "qutemol.cfg" (it will be recreated at the next execution).


Last updated on 26/09/06
[back]