My TiddlyWiki is running very slow

454 views
Skip to first unread message

war...@sunyit.edu

unread,
Mar 26, 2018, 8:38:41 PM3/26/18
to TiddlyWiki
I am currently working on a TiddlyWiki project for my class at SunyPoly. I have imported 82,952 tiddlers using the xlsx spreadsheet utilities. 
With 82,952 tiddlers it runs very slow. The size of the whole project is around 45mb. Does anybody know if I could somehow speed up the TiddlyWiki, or have I pushed it to the limit. 

Project can be found here - http://wardjhtrumptweets.tiddlyspot.com/

Thanks, 

TonyM

unread,
Mar 26, 2018, 9:44:21 PM3/26/18
to TiddlyWiki
On a given computer with sufficient ram you may be served well by increasing the ram available to your browser, both firefox and chrome permit this, in fact they set a limit so they do not overwhelm the computer, however if you work mostly in the browser more ram helps. I also use an ssd drive.

Regards
Tony

Mark S.

unread,
Mar 26, 2018, 11:08:20 PM3/26/18
to TiddlyWiki
Well, I was able to find covfefe, so at least it works.

TW isn't a real, indexed database. So the bigger it gets, the slower it will run. Thirty megs has been about my practical upper limit and <10mb on devices.

Probably the thing to do is to split it by some sub-category.

-- Mark

Steven Schneider

unread,
Mar 29, 2018, 10:00:53 AM3/29/18
to TiddlyWiki
Jeremy, would this work better in Xememex? Might be worth a shot? //steve.

Jeremy Ruston

unread,
Mar 29, 2018, 11:15:21 AM3/29/18
to tiddl...@googlegroups.com
Hi Steve

Jeremy, would this work better in Xememex? Might be worth a shot? //steve.

Sadly not; as things stand, Xememex would still render it as one big HTML file.

There are a bunch of approaches that might make it work with reasonable performance:

* running the wiki entirely on the server, just shipping static output back to the browser. That’s something that’s supported by Xememex, but would need some setting up to get things working in this case.
* a hybrid mode where the browser loads most tiddlers lazily, and defers to the server for searching across tiddlers that haven’t yet been loaded
* a refinement to the TW file format making it possible to stream tiddlers from the HTML file after the main wiki has been initialised

This example stretches the envelope of TiddlyWiki quite impressively (I’d have laughed if you’d told me a few years ago that it would be even remotely practical to work with a wiki this size; I think the most important factor is actually how much better browsers have got over the last 5 years). There’s quite a lot of interest in pursuing further optimisation of the TW core (for example, the work that Evan Balster has been doing).

I don’t see anything obvious that you could do improve things immediately (that’s not to say that others might not have some good suggestions based on their experience).

I noticed that the wiki is running on 5.1.14; it’s definitely worth upgrading to the latest 5.1.15 (or even the prerelease) to get the latest speed improvements that have gone into the core.

I’m starting work on a wiki with 50,000+ tiddlers for a client, and intend to explore a few further optimisations which will hopefully also benefit this example.

Best wishes

Jeremy

--
You received this message because you are subscribed to the Google Groups "TiddlyWiki" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tiddlywiki+...@googlegroups.com.
To post to this group, send email to tiddl...@googlegroups.com.
Visit this group at https://groups.google.com/group/tiddlywiki.
To view this discussion on the web visit https://groups.google.com/d/msgid/tiddlywiki/8fc84ba1-9a1a-4a70-8e2c-e75a35ead451%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply all
Reply to author
Forward
0 new messages