Home Networking

Man I wish I was getting that. I’m paying for a gigabit and only get about 500 megabits

Those tools aren't perfect, that's how.

You lucky bastard

I have a 500Mbps connection.

Fast.com is usually wrong and says I get 800Mbps-1.0Gbps

Speedtest is much more accurate. Try that.

What is that upload

Also a device can be rated for a speed but go higher. They guarantee minimum speeds. Usually if it does go over it isn't by much.

In this case though it's just fast.com being wonky most likely.

Just collect the extra speed to use later

You are lucky our verizon landline gave us one mbps or less for a long time one step above dialup it sure was a slow go until we switched to new provider it was like the difference between day and night. verizon is way behind time getting any fiber out and overcharging for landline phone to make up for it we will see what the future brings

ISPs usually over-provision the link in order to make meeting the speed commitment easier.

Source: work in the industry.

On a side note...

You are usually provisioned for the 1.2 Gbps.

If you get you a modem and a router that has a 2.5 Gbps port or higher, as well as your computer having it as well, you will get speed tests close to the 1.2 Gbps as well

To be honest it’s a lie. It’s likely some buffering on the ISP side that is causing it or it’s just wrong. A 1G interface won’t go above 1G. These are more to make customers or end users feel good. I have a feeling some code is messed up and it’s using 1000 somewhere instead of 1024 and through error stacking you get this. Throwing darts at what it could be but it’s not that high to a single device.

Huh, interesting result on my work's wifi.... Never seen UL significantly faster than DL before.

its a bug like others said. sometimes i get 800kbps-1mbps on T-Mobile 2G using fast.com which is impossible. most of the time its accurate at 160-190kbps though.

Happens to me too in Canada with Telus FiberOptic. I pay for gigabit but it goes to 1.1 and sometimes spikes to 1.5

Pedantic question: When was the last time you had your implementation of Fast calibrated?

Shit's an approximation, yo.

I'm on 250mb (boosted to 500mb thanks to Volt) and I just hit 630mb on my wifi.

​

Superhub in modem mode, a good cat 6 cable, and an AX50 I bought 2nd hand for £40.

openspeedtest>

Compression

Probably due to overprovisioning. A lot ISPs set their limits higher than the rated speed in case of network congestion during peak times.

This is why I change the settings to run for longer. 60-90 seconds usually.

Also, you have a crap upload speed.

I wish this internet speed is possible in my third world country but it doesn't. Hopefully starlink can reach higher speeds while also accessible in my region.

Speed tests cheat, they don't make you download thousands of gigs of data to get a throughput. Which is exactly what you would need to do to accurately test your connection. They send a bit of data, see how fast it goes, do some calculations and assume.

ISP's can also cheat as they can give 'burst' service which temporarily gives you a bump in speed. This can be advertised to you as a perk, but I don't think many people care, or used to trick speed tests.

It's like when your boss walks by you and you work a bit harder to appear busy.

Fast.com is not really accurate. Use speedtest.net instead....

At the end of the day, it's an estimate of throughput using test data which is often not representative of real-world performance because real-world data is much more diverse than speed test data. It's not meant to be a 100% technically perfect bit-by-bit analysis because that level of detail is irrelevant if you're in the ballpark for "good enough for streaming"

If you would like a speed test that hides fewer details and gives you a better idea of what you can expect for different file sizes, etc. try cloud flare's speed test

speed.cloudflare.com

Is it fiber? I routinely see fiber go over the assignment speed by around 300Mbps.

Math bad.

It'll have peaks that happen here and there.