Page load time extension results for popular websites

[SchumannFrequency]

Problem with most of the sites tested they are affected by your account, if you logged in etc.
But I did youtube subscriptions page in firefox.
1.07 secs.
I expect CPU, browser config and if you blocking trackers has more impact than the internet connection.

I always log out for the above results and I disable all extensions before doing the tests in the above results.

With the these configurations, I get the following result for the YouTube subscriptions page in Firefox.

 

[igormp]

I believe none of your results are valid, especially for the sites that load tons of dynamic content.

One simple example is how your empty subscriptions page loaded in almost the same time as the one with the full feed, with lots of thumbnails and whatnot.

Main problem is that it's just measuring the time to fetch and render the initial content the server throws you, but it totally ignores any subsequent ajax calls, which are often the slowest part of any modern website. If you take a look at the devtools' networking tab, you should see content taking way longer to download compared to what the extension points out.
The author itself says so, fwiw:

Not work's correctly on dynamic sites · Issue #8 · alex-vv/page-load-time

http://213.136.70.253/4chrome-demo/ Calculate loding time only for first request

github.com

I'm also not sure if login pages are that relevant to measure, since they don't contain much stuff at all.

 

[Launcestonian]

I have always used the same internet subscription since the beginning of these tests.
This is an 'unlimited' subscription, but among the slowest performance available in the country.
nPerf internet speed stats: download speed= +- 94 Mb/s upload speed= +- 14 Mb/s latency= +- 11.4 ms jitter= +- 2.5 ms

I would get higher results if I used a faster connection, but I have no idea how much difference it makes.
My computer is fairly compact. I can take it to an acquaintance who has fast internet one day to test what impact it has on the results.

Yes, my net config is similar to you, but the fact that a signal has to hop throw many different points of data connections because of geographical location on the earth is definitely a factor in the speed of loading web pages on any given device.

 

[igormp]

Yes, my net config is similar to you, but the fact that a signal has to hop throw many different points of data connections because of geographical location on the earth is definitely a factor in the speed of loading web pages on any given device.

Given how they're testing mostly big websites, CDNs and the like should make this mostly a non-issue.

 

[Launcestonian]

Given how they're testing mostly big websites, CDNs and the like should make this mostly a non-issue.

"mostly" is an interesting word. Relative of course, but CDNs work by caching frequently requested content at edge locations (proxies) closer to users.
You'll have to define "frequently requested"? Like how does the entire network know what I'm going to be doing when online? obviously the answer is unless they can read my mind, the do not know.

This TPU site is one of the slowest websites I use, not sure if that is unique to me but just saying.

 

[igormp]

"mostly" is an interesting word. Relative of course, but CDNs work by caching frequently requested content at edge locations (proxies) closer to users.
You'll have to define "frequently requested"? Like how does the entire network know what I'm going to be doing when online? obviously the answer is unless they can read my mind, the do not know.

This TPU site is one of the slowest websites I use, not sure if that is unique to me but just saying.

Yeah, stuff is relative for sure.

Have you tried measuring stuff with the network tab in dev tools to see what's the slowest part?

 

[chrcoluk]

I always log out for the above results and I disable all extensions before doing the tests in the above results.

With the these configurations, I get the following result for the YouTube subscriptions page in Firefox.

Getting 880ms now whilst logged in, Even zoom level, size of window affects it, as more window means more thumbnails to load in.

Its a nice extension, more convenient than going to the dev tools portion of browser.

 

[igormp]

Its a nice extension, more convenient than going to the dev tools portion of browser.

Those two show totally different results, as per my previous comment, so be aware of that.

 

[SchumannFrequency]

One simple example is how your empty subscriptions page loaded in almost the same time as the one with the full feed, with lots of thumbnails and whatnot.

Thumbnails are usually made in advance rather than on the fly.
You could store them in DynamoDB Accelerator (DAX) and then thumnail fetching is hardly going to add any latency.
If you are using CloudFront to make thumbnails available that you have stored in S3, the retrieval of the thumbnails is also normally going to be lightning fast.

 

[igormp]

Thumbnails are usually made in advance rather than on the fly.
You could store them in DynamoDB Accelerator (DAX) and then thumnail fetching is hardly going to add any latency.
If you are using CloudFront to make thumbnails available that you have stored in S3, the retrieval of the thumbnails is also normally going to be lightning fast.

It is fast, but the point is that those are subsequent requests that are not made within the initial requests, and thus is not measured by that extension.
Fetching the stuff is one thing, and can be fast, but there's also the rendering time for those in your DOM, and the overhead of all of the code underneath (it's not just fetching the images, but doing calls to know which images to fetch then getting those along with the metadata for the video itself).

But you don't need to care about what I said. The actual creator of the extension said so, as I posted before.

 

[SchumannFrequency]

The TPU forum is (probably) popular with the readers of this thread.
I tested the performance of the TPU forum using Chromium and Firefox on OpenBSD.
I was logged into the forum while I did these tests.

Firefox: 315 ms

Chromium: 337 ms

Internet speed/Software/Hardware: OpenBSD -current, Intel 12700KF, Realtek 1G LAN, 94MB/s (max) internet dl speed

Firefox version: 133.0
Chromium version: 131.0.6778.85