Re: [Jmol-users] Jmol in Crystallography

[Carroll, P. J <car...@sa...>]

Bob,

      I've been experimenting with my webpage load times to see what's going on. First, I wanted to see what the effect of the calculation of the unit cell contents is. So on my website, I used my 1158cell.html page, which is the slowest loading one in my Jmol examples. I changed the cells loaded from {555 555 0} to {444 666 0} to {333 777 0} and the results were a little surprising. Using the web developer tools you told me about, the load times of the three webpages above were 4.58 sec, 5.45 sec, and 11.53 sec, respectively. In each of these (and, in fact for all the pages tested) there seemed to be a pause at 3 or 4 sec and then, when the cell is finally displayed, the scan continues.

      I decided to look for the most egregious example I could find and next built a page using an old structure in a cubic space group that contains 72 molecules. I did the loads with {555 555 0), {444 666 0}, and {333 777 0} and got load times of 5.32 sec, 1.70 min, and 167.46 min!!! (That's right, almost three hours - I have repeated this several times).

     I then did these loads with the Jmol application (eliminating the effect of my server) and got times of 2 sec, 24 sec, and 8 min!

     Clearly the calculation of unit cell contents is the most important variable effecting load times.

     I'm a little surprised by these results, but I don't know if it's a big deal. For the sorts of packing diagrams I'm interested in, I would not use a structure like that cubic one and load times of 10 - 12 sec would be perfectly acceptable. 


Pat


________________________________________
From: Robert Hanson via Jmol-users 
Sent: Monday, December 1, 2025 11:04 PM
To: jmo...@li...
Cc: Robert Hanson
Subject: Re: [Jmol-users] Jmol in Crystallography

Nice to hear from you, Pat! Ah, yes....PARTS...

[image.png][image.png]

That was a nice addition. Thank you for pointing out the need for that and helping me get it right.

I wonder why your server is so slow:

[image.png]

Maybe because it is http and not https?

Are these buttons supposed to do something?

[image.png]

Wait, now I see it. I have to press "add" -- NICE!

[image.png]



Have you played with the new MODELKIT command? See

https://journals.iucr.org/j/issues/2024/05/00/dv5017/index.html

and

https://chemapps.stolaf.edu/jmol/jsmol/iucrdemo/

Wow, this goes on and on!!!

Fantastic. Is there an index?

Going to Calgary? https://www.iucr2026.org/
I'm hoping to make a couple of presentations there.

Bob