Solving the Micromass Big Integral Challenge - Issues with Latex Rendering?

Click For Summary

Discussion Overview

The discussion revolves around performance issues related to LaTeX rendering in a specific forum thread on Physics Forums. Participants share their experiences with loading times and responsiveness across different browsers and devices, exploring potential causes and solutions.

Discussion Character

  • Technical explanation
  • Debate/contested

Main Points Raised

  • One participant reports significant delays when opening a specific thread, suggesting a possible issue with LaTeX rendering.
  • Another participant notes that threads with extensive LaTeX content tend to struggle due to client-side processing demands.
  • Some participants mention experiencing slower LaTeX rendering on mobile devices, particularly after a recent Firefox update.
  • A participant shares that they have not encountered the same issues on their desktop but acknowledges the resource limitations of mobile devices.
  • One participant suggests that the MathJax developer has indicated no immediate solutions are available, recommending the use of Opera for better JavaScript processing.
  • Another participant compares loading times across different devices and browsers, noting variability and specific issues with scrolling and rendering times.
  • Concerns are raised about browser cache settings and their impact on rendering times, with one participant reporting improvements after adjusting cache settings.
  • Participants express frustration with extraneous content on certain websites affecting loading times, suggesting a preference for browsers that allow quicker dismissal of such content.

Areas of Agreement / Disagreement

Participants generally agree that LaTeX rendering can cause performance issues, particularly in specific threads, but there is no consensus on the exact causes or solutions. Multiple competing views on browser performance and settings remain unresolved.

Contextual Notes

Limitations include varying experiences across different devices and browsers, the dependence on specific browser versions, and unresolved technical aspects of LaTeX rendering and caching mechanisms.

jim mcnamara
Mentor
Messages
4,789
Reaction score
3,852
Great thread!:
https://www.physicsforums.com/threads/micromass-big-integral-challenge.867904/#post-5450007

I have a 16GB i7 windows 7 box with Firefox v45.0.2. I have NO add-ons to Firefox.

Every time I open the one thread, it takes longer and longer - literally a minute the last time.
No other threads do that from PF or elsewhere. While freezing the PF window, everything else on the box is happy, the window simply becomes unresponsive for 60 seconds or longer.

Me, PF, or should I blame my dog? I'm guessing there is some sort of issue with Latex rendering. Dunno.
 
Physics news on Phys.org
Threads with a lot of latex struggle. Always have. There is a lot of processing going on in the background client side. I've sent the thread to the mathjax developer to see if there is anything we can do config wise.
 
I must say LaTeX rendering has seemed slower on my phone since the last Firefox update - I'd been thinking of starting a thread about it. Haven't seen it on my desktop though.
 
Hmm - we are doing what we can. In terms of the phone, those generally are pretty resource limited compared with most desktops. Yes?
 
Yes - but I don't remember it struggling so much until the last Firefox update. Could be bad luck with the threads I've been viewing of course. I've not done any formal testing.
 
The developer said nothing can be done at this point. I would suggest looking into Opera as the newest versions have tested even better than Chrome for javascript processing.
 
jim mcnamara said:
Every time I open the one thread, it takes longer and longer - literally a minute the last time.
I just looked at that thread using Firefox 45.0.2 on this 6-year-old iMac with an i3 processor and 4GB RAM. It takes maybe 3 seconds at most, for a page to render.

It does bog down seriously in Safari on my iPhone 5s. I have to wait a while before I can scroll the first page smoothly.
 
Last edited:
Well, I just tried that thread and it took 44 seconds to completely load. Used IE 11, Chrome, and Firefox. Times varied somewhat.

But - the (very good idea of) positioning cursors in the thread at what PF thinks is the last place you read is wherein lies the problem for me.
It takes about 8 seconds to get a rendering of post #nn. If I want to move up or down, at that point, forget it. That is what I meant by "completely".
If I attempt an up arrow or down arrow or whatever positioning I try, it hangs.
I would guess that is what @jtbell saw.

Is there some browser cache setting I can enable? my PC C: drive is a Crucial SSD and I/O is very good.
 
jim mcnamara said:
Is there some browser cache setting I can enable? my PC C: drive is a Crucial SSD and I/O is very good.
The equations are loaded on the fly and not cached. They are not images but css formated text.
 
  • #10
Tinked with some settings got the completed intial load time down by about 50% - Firefox 45.0.2 does in fact have caching for css rendering.
So, if I bump the cache somewhat, restart Firefox, then the page loads faster.

And this is not the only issue with Firefox, not related to PF. Sites like CNN and Christian Science Monitor spam a bunch of crud onto the main page so that you cannot see what's underneath and you have to wait until the extraneous junk loads to be able to dismiss it. With Chrome I can dismiss the overlaid junk early. So maybe I do need another browser like Opera.

Thanks for you help.
 
  • Like
Likes   Reactions: Greg Bernhardt

Similar threads

LaTeX Latex rendering (another thread)
  • · Replies 16 ·
Replies
16
Views
3K
Mathematica How to type mathematical equations using LaTeX
  • · Replies 3 ·
Replies
3
Views
2K
Challenge Micromass' big August challenge
  • · Replies 105 ·
4
Replies
105
Views
15K
  • Sticky
Physics Forums Global Guidelines
  • · Replies 2 ·
Replies
2
Views
506K