# Help with experiment

**URL:** <https://talk.jekyllrb.com/t/help-with-experiment/2654>\
**Category:** Help\
**Created:** [February 20, 2019, 1:25pm UTC](https://talk.jekyllrb.com/t/help-with-experiment/2654 "2019-02-20T13:25:07Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![ashmaroli](https://avatars.discourse-cdn.com/v4/letter/a/df705f/32.png) [@ashmaroli](https://talk.jekyllrb.com/u/ashmaroli)\
**Post date:** [February 20, 2019, 1:25pm UTC](https://talk.jekyllrb.com/t/help-with-experiment/2654/1 "2019-02-20T13:25:07Z")

</div>

Hello everyone! 👋

As part of optimizing Jekyll for v4.0, I’ve [opened a new pull request](https://github.com/jekyll/jekyll/pull/7542)

I’d love it if you guys here (especially those who have large sites using numerous includes) could use that PR’s branch to give a test-run and see if there’s a significant improvement in the build times between the PR branch and Jekyll’s `master` branch:

```ruby
gem "jekyll", git: repo, ref: "refs/pull/7542/head"

```

Thanks in advance.

Cheers,  
@ashmaroli

---

<div class="post-metadata">

**Author:** ![mmistakes](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.jekyllrb.com/mmistakes/32/2169_2.png) [@mmistakes](https://talk.jekyllrb.com/u/mmistakes)\
**Post date:** [February 20, 2019, 9:01pm UTC](https://talk.jekyllrb.com/t/help-with-experiment/2654/2 "2019-02-20T21:01:02Z")

</div>

There’s a significant gain for my rather large site. Knocked about 40 seconds off the build time!

| 3.8.5 | master | refs/pull/7542 |
| --- | --- | --- |
| 121.766s | 83.576s | 86.019s |

jekyll-seo-tag seems busted with these changes, but everything else looks good from a glance.

---

<div class="post-metadata">

**Author:** ![ashmaroli](https://avatars.discourse-cdn.com/v4/letter/a/df705f/32.png) [@ashmaroli](https://talk.jekyllrb.com/u/ashmaroli)\
**Post date:** [February 21, 2019, 2:30am UTC](https://talk.jekyllrb.com/t/help-with-experiment/2654/3 "2019-02-21T02:30:46Z")

</div>

Thank you for testing.  
But, I was looking for the comparison between the **`master`** branch and the PR branch since it has _already been established_ that the **`master`** branch is significantly faster than **`v3.8.5`**

_P.S. `jekyll-seo-tag` will be fixed around the same time Jekyll 4.0 is finalized._

---

<div class="post-metadata">

**Author:** ![mmistakes](https://yyz2.discourse-cdn.com/flex030/user_avatar/talk.jekyllrb.com/mmistakes/32/2169_2.png) [@mmistakes](https://talk.jekyllrb.com/u/mmistakes)\
**Post date:** [February 21, 2019, 1:11pm UTC](https://talk.jekyllrb.com/t/help-with-experiment/2654/4 "2019-02-21T13:11:31Z")

</div>

I completely missed the part about testing against `master`. I’ve amended my numbers above. Looks like the PR branch took about a 3 second hit compared to `master`.

---

<div class="post-metadata">

**Author:** ![ashmaroli](https://avatars.discourse-cdn.com/v4/letter/a/df705f/32.png) [@ashmaroli](https://talk.jekyllrb.com/u/ashmaroli)\
**Post date:** [February 21, 2019, 1:26pm UTC](https://talk.jekyllrb.com/t/help-with-experiment/2654/5 "2019-02-21T13:26:41Z")

</div>

Oh! That’s a big hit albeit unexpected though.  
Thank you very much Michael. 👍 👍