[chimerax-users] Memory problems with Orbital Viewer

[Tony Schaefer]

Michael,

I've released an updated version of SEQCROW to the toolshed (Tools -> More
Tools... on the menu). This just has a different method for determining
available memory, which will hopefully fix this. If memory is still an
issue, I can try to change how the orbitals are calculated to reduce memory
usage.

As a side note, ORCA 5.0.1 has a race condition bug, so it might give you
bogus energies if you run it in parallel. There's more details on the ORCA
forum, but the developers recommend updating to a newer version.

Best,

Tony

On Tue, Jan 24, 2023 at 12:50 AM Michael Patzschke <m.patzschke at hzdr.de>
wrote:

> Hello Tony,
>
> Thank you very much for your quick reply.Can I first say, that the orbital
> viewer is a very useful tool and that I appreciate your work and effort.
> Now to your questions. The message is copy and pated from the screen. It
> does indeed say 0.0GB. The orca version used for the calculation is 5.0.1,
> the settings are the standard settings in orbital viewer. Padding 3.0 Å,
> resolution 0.1 and IsoSurfaces 0.01.
>
> I just started ChimeraX (version 1.5) again and tried the same
> visualisation, now the message says "Estimated peak memory usage (0.1GB)”
> The molecule is acetonitrile.
>
> Kind regards,
> Michael
>
>
>
> On 24. Jan 2023, at 08:39, Tony Schaefer <tony.schaefer.chem at gmail.com>
> wrote:
>
> Michael,
>
> I develop the Orbital Viewer tool. Does it really say 0.0GB for both
> estimated peak memory usage and available memory? I've looked into it a
> bit, and it looks like the way I'm determining available memory is not the
> best. I'll release a patch for that soon. You might also have other
> applications open that are taking up memory.
>
> I do not expect the estimated memory usage to be 0 GB. That's calculated
> directly from the grid size, number of threads, and number of atoms. If
> this does say 0.0 GB, I'd be interested in your orbital settings, basis
> set, and ORCA version.
>
> As the message suggests, you can decrease the memory required by adjusting
> the resolution, padding, or number of threads on the 'orbital settings' tab
> of the Orbital Viewer tool. If you are able to load an orbital (with the
> resolution set to 0.75 A to be quick/safe), you can see how much extra
> padding you have by running the command `volume showOutlineBox true`. This
> will show the total volume where the orbital values were calculated, even
> though only the isosurface is shown. A smaller box (i.e. less padding) will
> require less memory at a given resolution.
>
> The warning is there because exceeding the available memory could crash
> ChimeraX or force you to restart your computer. It's not meant to be a
> nuisance, so I'd like to make sure it's working properly.
>
> Best,
>
> Tony
>
> On Mon, Jan 23, 2023 at 11:25 PM Michael Patzschke via ChimeraX-users <
> chimerax-users at cgl.ucsf.edu> wrote:
>
>> Dear colleagues,
>>
>> I am using ChimeraX mostly for small molecular complexes, typically up to
>> 100 atoms. I am running the code on a MacBook Pro with 16 GB of memory
>> under macOS 11.8.8. When I try to render molecular orbitals from an orca
>> calculation I get even for very small molecules the warning message:
>>
>> Estimated peak memory usage (0.0GB) is above or close to
>> the available memory (0.0GB).
>> Exceeding available memory might affect the stability of your
>> computer. You may attempt to continue, but it is recommended
>> that you lower your resolution, decrease padding, or use
>> fewer threads.
>>
>> Press "Ok" to continue or "Abort" to cancel
>>
>> Is there any way to fix this problem?
>>
>> Kind regards,
>> Michael Patzschke
>> _______________________________________________
>> ChimeraX-users mailing list
>> ChimeraX-users at cgl.ucsf.edu
>> Manage subscription:
>> https://www.rbvi.ucsf.edu/mailman/listinfo/chimerax-users
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.rbvi.ucsf.edu/pipermail/chimerax-users/attachments/20230124/04eafc7b/attachment.html>