Performance on large number of pages

[geedmo]

That's not really an issue.

If you are interested, I've been playing with Pico regarding to issue #97 and I made a few changes and a little performance check which gave me nice results. A brief summary in the repo https://github.com/geedmo/Pico

When you get chance, I would like to know your thoughts.

[Frodox]

Also, try to check out fork of Pico -- PhileCMS, may be you will like it :)

[geedmo]

PhileCMS is more complex and it goes beyond the philosophy of Pico. Just my opinion, thanks anyway.

[mistergraphx]

Hi @geedmo,

i just test your modifications, with 90 articles, and i've a better result without : as you say, the testing method is what it is ;-) but i'm astonished of these results.

---

Without optimisation :

Start: 05:20:41

End: 05:20:41

Processed: 90

Elapsed

0.23807597160339 seconds

---

With complete optimisation (page_content() & exerpt):

Start: 05:34:32

End: 05:34:32

Processed: 90

Elapsed

0.29801678657532 seconds

---

Without parse_content() only

Start: 05:24:01

End: 05:24:02

Processed: 90

Elapsed

0.30226588249207 seconds

---

[mistergraphx]

Hi, After doing more test

• exerpts recalculated each time we load a page,we lost 0.10 ms each time we load a page with this.

i've a flat file cms tool that i've made for me (named Little), the solution i've found is to write a temporary index.json file containing page list, page url, page.title, resume, md5 sum ... each page we just've to verify the index state and compare it to the the file system ... may it's quicker than parsing on every action ...

So i've test rendering the same article in Little and Pico (with cache enabled)

Pico with Twig : Elapsed: 0.1051 seconds "Little" with mustache.php : Elapsed: 0.095 sec

Playing with this, i've discover that the choice of the template engine is a perf issue too ! Twig is slower, due to he template inherance and more complex/logical layouts certainly.