I tested both Speedtest.net and nperf.com and compared the results.

[SchumannFrequency]

I tested both Speedtest.net and nperf.com and compared the results. nPerf gets slightly higher download results, the upload speed is the same, but with latency I saw a significant difference. Then I looked up which one is the more reliable of the two and I found the following topics:

[RESOLU] nperf vs speedtest, pourquoi ces différences énormes ?

Bonjour, Je n'ai pas trouvé de réponse sur le site (en tout cas je ne l'ai pas vues), donc merci pour votre tolérance à mon premier message ;)

routeur4g.fr

original French text: Notre site de référence est nPerf, et je conseille d'utiliser toujours le même serveur (choix manuel) Wingo en Suisse 100Gbs. nPerf utilise plusieurs connexions simultanées pour son test.
Speedtest d'Ookla est pour moi moins qualitatif et souvent l'upload est bloqué chez moi.
Translation: Our reference site is nPerf, and I advise to always use the same server (manual choice) Wingo in Switzerland 100Gbs. nPerf uses multiple simultaneous connections for its test.
Ookla's Speedtest is less qualitative for me and the upload is often blocked for me.

original French text: Depuis 3 mois, je suis fibré (voir ma signature) , j'ai un forfait Orange 400Mb/s symétrique (en down et up ); avec Nperf, j'ai toujours autour de 400 dans les 2 sens et même des fois un peu plus! , mais avec Speedtest je n'ai jamais dépassé 300, souvent même en dessous de 200 ! mesures faites au mêmes instants ! j'ai même eu des valeurs très faibles inférieures a 10 alors que c'était faux.
Translation: For 3 months, I have been on fiber (see my signature), I have a symmetrical Orange 400Mb/s package (down and up); with Nperf, I always have around 400 in both directions and even sometimes a little more! , but with Speedtest I never exceeded 300, often even below 200! measurements taken at the same time! I even had very low values below 10 when it was wrong.

There are other topics that also report that there may be a difference between the two websites:

speedtest.mybroadband vs nperf.com

If I run a speedtest using http://speedtest.mybroadband.co.za/ I get: Ping 3ms Download 194 mbps If I run a speedtest using https://www.nperf.com/en/ I get: Ping 16ms Download 625 mbps Both sites showed similar upload speeds, around 95 mbps. nperf reports more than 5 times worse latency...

mybroadband.co.za

https://www.geekzone.co.nz/forums.asp?TopicId=191142

Not all speedtests are created equal – Fuzion

https://www.reddit.com/r/HomeNetworking/comments/6d17z5/_/dhz1upk Speedtest.net seems the most fishy/infiltrated by ISPs.

For this reason I'm going to post my nPerf result:

 

[boomheadshot8]

I tested both Speedtest.net and nperf.com and compared the results. nPerf gets slightly higher download results, the upload speed is the same, but with latency I saw a significant difference. Then I looked up which one is the more reliable of the two and I found the following topics:

[RESOLU] nperf vs speedtest, pourquoi ces différences énormes ?

Bonjour, Je n'ai pas trouvé de réponse sur le site (en tout cas je ne l'ai pas vues), donc merci pour votre tolérance à mon premier message ;)

routeur4g.fr

original French text: Notre site de référence est nPerf, et je conseille d'utiliser toujours le même serveur (choix manuel) Wingo en Suisse 100Gbs. nPerf utilise plusieurs connexions simultanées pour son test.
Speedtest d'Ookla est pour moi moins qualitatif et souvent l'upload est bloqué chez moi.
Translation: Our reference site is nPerf, and I advise to always use the same server (manual choice) Wingo in Switzerland 100Gbs. nPerf uses multiple simultaneous connections for its test.
Ookla's Speedtest is less qualitative for me and the upload is often blocked for me.

original French text: Depuis 3 mois, je suis fibré (voir ma signature) , j'ai un forfait Orange 400Mb/s symétrique (en down et up ); avec Nperf, j'ai toujours autour de 400 dans les 2 sens et même des fois un peu plus! , mais avec Speedtest je n'ai jamais dépassé 300, souvent même en dessous de 200 ! mesures faites au mêmes instants ! j'ai même eu des valeurs très faibles inférieures a 10 alors que c'était faux.
Translation: For 3 months, I have been on fiber (see my signature), I have a symmetrical Orange 400Mb/s package (down and up); with Nperf, I always have around 400 in both directions and even sometimes a little more! , but with Speedtest I never exceeded 300, often even below 200! measurements taken at the same time! I even had very low values below 10 when it was wrong.

There are other topics that also report that there may be a difference between the two websites:

speedtest.mybroadband vs nperf.com

If I run a speedtest using http://speedtest.mybroadband.co.za/ I get: Ping 3ms Download 194 mbps If I run a speedtest using https://www.nperf.com/en/ I get: Ping 16ms Download 625 mbps Both sites showed similar upload speeds, around 95 mbps. nperf reports more than 5 times worse latency...

mybroadband.co.za

https://www.geekzone.co.nz/forums.asp?TopicId=191142

Not all speedtests are created equal – Fuzion

https://www.reddit.com/r/HomeNetworking/comments/6d17z5/_/dhz1upk Speedtest.net seems the most fishy/infiltrated by ISPs.

For this reason I'm going to post my nPerf result:
View attachment 256867

Maybe your CPU is limited, check
If you want a good test try IDM or something light (cmd) and go to http://bouygues.testdebit.info/ and also https://ip.lafibre.info/test-debit.php : mainly for Europe users

 

[SchumannFrequency]

Maybe your CPU is limited, check
If you want a good test try IDM or something light (cmd) and go to http://bouygues.testdebit.info/ and also https://ip.lafibre.info/test-debit.php : mainly for Europe users

That may explain the difference in download speed, but not the difference in latency.

There are two different people on router4g.fr who claim that Speedtest.net is significantly less accurate than nPerf in their experience. That applies to France.

But if you then read my link from geekzone.co.nz (New Zealand) you will see that the following is claimed:

1. In my experience, nPerf is closer to real-world performance (eg downloading a MS ISO) than OOKLA.
2. Speedtest.net servers are also massively variable between sites.

Speedtest.net also has significantly more adds.
And you have many other sources that claim that Speedtest.net is completely unreliable:

Why Do My Results Differ From Speedtest.net / Ookla Speed Tests?

The most common question that people ask us is why their results here differ from speedtest.net. I want to explain something that may help you understand why this is.According to the Ookla Wiki [updated link] the following is true about your tests taken at speedtest.net... The fastest 10% and slo...

testmy.net

1. i usually see very different results than here, when my internet is having issues speedtest.net doesn't usually detect them at all it will show my full 50/5 where as here i can see my service is really suffering and only receiving half that
2. Alot of times you cannot choose the closest server on speedtest.net. If i choose Tampa, which is 50 miles away, or Miami, or ANY of the close servers ; i get a 1/4th of my speed. But if i choose Atlanta GA , 900 miles away, or Huston TX , or any of Texas for that matter, i get top speeds Go figure.

https://www.reddit.com/r/Twitch/comments/3v3jk5 ISPs will often cheat and give priority to speedtest in particular

Then I'm going to look at mobile app ratings:
https://apps.apple.com/us/app/nperf-internet-speed-test/id916797048 Rating: 4.8
https://apps.apple.com/us/app/speedtest-by-ookla/id300704847 Rating: 4.1
https://play.google.com/store/apps/details?id=com.nperf.tester&hl=en&gl=US Rating: 4.7

Speedtest by Ookla - Apps on Google Play

From the makers of Speedtest.net. Take accurate speed tests anywhere!

play.google.com

Rating: 4.1

Then you also have this topic for example: https://mybroadband.co.za/forum/threads/speedtest-net-for-international-throughput-testing.935869/
I'm testing on a 500Mbps link. Speedtest.net never gets more that 100Mbps international. iperf and mikrotik get very close to full speed. All TCP.

VOO bride-t-elle les vitesses sur speedtest.net ? https://forum.voo.be/conversations/...702b52af49?commentId=5e53e5884cf00d702b2506f1
I did a speetest on speedtest.net on several different servers I am at 60 constant (apart from a few months or even a week ago I was at 120)
I did the Nperf test I got my 150mb/s

https://gathering.tweakers.net/forum/list_messages/1802201
1. Now we probably all know speedtest.net (from Ookla), but I regularly hear that the values you get from them are very questionable and do not give a good idea of the real speed you get from your ISP.
The speed test that you can do on your ISP's server is often not reliable, because it always gives perfect scores on their own servers.
2. At speedtest.net a server is automatically chosen, but beyond distance (based on geo-IP :X ) is not looked at. Not all servers are equally reliable. Plus you don't automatically have the same server twice in a row.

ADSL-BC • Consulter le sujet - Fiabilité des test de vitesses de download/upload ?

Q. How can the download speed drop like this, even lower than the upload speed?
A. nPerf might be more reliable than speedtest

Is speedtest.net reliable? https://www.reddit.com/r/networking/comments/39a78t I have a 100 Mbit/sec connection at work, that I run. Previously, I could peg out my home connection at 30Mbit. After the "upgrade", any transfers start out in the 60Mbit range, and over the course of about 15-20 seconds, tapers off to about 30, many times much less.
In effect, what they have going is a burst rate, like old Frame Relay circuits. To me this is just dishonesty on their part. Don't sell a 30 Mbit circuit as 60 when all you can sustain is 30. I would rather have my 30 back, as something about their algorithms messes with connections past the burst. You'll get that burst, then it tail drops or something and really fucks up TCP connections for a minute.
Back to the speedtest.net issue. The testing phase lasts far longer than my normal burst time, so it's obvious that they have built in exceptions for the speedtest.net nodes. Not sure if they do this at the edge, or in the cable modem, but it is damned dishonest.

Finally, I see that Dutch hardware websites also rely on nPerf instead of speedtest.net:

Hoe snel is jouw internetverbinding? Test het met de Hardware.Info Internet Snelheidstest!

Hoe snel is jouw internetverbiding? Samen met nPerf hebben we een handige tool ontwikkeld waarmee je dat snel met één druk op de knop kunt testen. Hiero...

nl.hardware.info

My general conclusions about speedtest.net are the following:
1. The servers they use are less reliable than the ones nPerf uses, maybe more network traffic saturation, maybe the servers are simply less qualitative, etc.
2. Speedtest.net's algorithm is (usually) less accurate than nPerf's
3. Speedtest.net's mobile apps get significantly lower ratings than nPerf.
4. speedtest.net has more ads, which is undesirable
5. Ookla doesn't offer any privacy: https://www.speedtest.net/about/privacy
6. As Rex9 points out in the above reddit comment, although the speed test is long it only gives the short boost speeds, which is very misleading to customers.

I would personally use speedtests other than speedtest.net to measure my speeds. And I would always prefer nPerf over speedtest.net as their results are more accurate and more reliable for most people in the world, and they have better apps, better privacy, and fewer ads.

 

[P4-630]

You could also check www.fast.com or www.speedtest.nl

 

[I hit the lottery]

Nperf had multiple throttled sites for me. as you can see here, this is what your graph will look like if you have a fast connection and they're trying to throttle you:

I had good luck with speedtest here is results from all over my country:

and just to make me feel good. here is a server i always get a good connect to, even thought verizon is not my provier.

And yes my upload isnt great, but when you're getting gigabit though coaxial...the copper, the big shared line, it is what it is, but for what i pay, the download is smoking, and the pings arent bad 10-50ms through my country.

 

[Det0x]

Not really sure what i'm supposed to be testing here..

Home server:

Warsaw server 2000km away. (by road 2650km)
(~1240miles and ~1650miles for those from us)

 

[Athlonite]

Both speedof.me and testmy.net choose none 10Gb servers so can't get an accurate reading for my 2Gb/2Gb connection as for nPerf even to local NZ 10Gb servers I can't get upload speeds anywhere near what I should be

I get better out of speedtest though

 

[dgianstefani]

I'd prefer half the latency over 50% more speed.