CA3LE

TestMy.net returning slower results than other speed tests on my rMBP

1 post in this topic

Hey Damon.  I've used your site for speed tests for numerous years.  And only your site really.  I got a new router today and was checking out some speeds.  And I decided to check a few other sources also.  They are all over the place for the varying speed test providers.  I've read your stuff on why many aren't accurate or can't really be trusted.  But what can be said about the below data?  

Testing setup:

Computer - 2013 MacBook Pro Retina (802.11ac Wi‑Fi wireless networking;3 IEEE 802.11a/b/g/n compatible)

Router - Apple Airport Extreme 802.11ac (latest version)
Distance from router - < 2 feet
Testing location (home) - NW suburbs of Atlanta (25 miles)
ISP: Comcast
Paid for speed: 105Mbps

 

Tests:
ATT.com internet speed test site  - 101.77, 109.05, 101.28, 101.18 (location not identified)
Speedtest.net - 107.63 (Atlanta), 113.25 (Atlanta), 113.80 (Atlanta), 109.08 (Chattanooga)
Speakeasy.net - 89.20 (Atlanta), 84.07 (NY), 73.18 (NY), 65.64 (Dallas), 79.43 (Atlanta)
Comcast iPV4 - 71.03 (Atlanta),  70.96 (Atlanta), 113.24 (Richmond, VA), 87.46 (Nashville)

Comcast iPV6 - 108.35 (Atlanta), 104.91 (Atlanta), 96.23 (Richmond, VA)

TestMy - 54.85, 54.9, 56.23, 47.7

 

** side note - why does Comcast offer iPV4 and iPV6 results?

 

Anyway, what does all this tell you?  Your numbers are significantly lower than all the rest.  Are all other sites are unreliable?  Am I really not getting anywhere near the advertised speed of 105Mbps?  

Thanks for your wisdom and advice!

Jack

 

Hi Jack,

 
Thanks for contacting me and using my site for so many years.
 
Let me start off by saying... I have the exact same provider as you, maxed out rMBP that's a couple months old.  I'm able to max out my connection but not under all circumstances.
 
This is also the most frequently asked question.  I like that you aren't like most people... most people attack me, many make assumptions that the test is to blame.  I can type till I'm blue in the fingers, people with that mentality usually don't get it.  Sometimes I doubt that they even read what I write.  Really sucks that it takes at least 20 people asking this question before I get one person who gets it.  The people who trust me... have faster connections than their friends.
 
Let me show you some of my personal results.  My logged results are confusing because sometimes I purposely limit my connection, it's part of my develop process... just know that the highest results are my normal daily speeds.  Rarely do I deviate.
 
http://testmy.net/stats/?&t=u&l=25&z=5&q=CA3LE  (showing only my laptop results, which is usually in my kitchen 10' from my Netgear Nighthawk with one wall between)
 
My iMac, same thing...
 
Multithread 119 Mbps >> http://testmy.net/db/EiaT6t1 - single thread 107 Mbps http://testmy.net/db/2L1ClGq
 
These are typical results for me.  I'm in Colorado Springs, CO -- over 700 miles from the main Dallas Texas server.  Testing at greater distance I start to drop.  I see between 65-80 Mbps to DC and Seattle... but if I multithread along the same routes I'm able to pull 109 Mbps >> http://testmy.net/db/YrdymVe - my Wife's rMBP and my mac mini are the same story.  All running 100% default configuration... there's nothing special that you have to do to get OSX to perform like mine does.
 
Other speed tests multithread... and I mean pretty much 100% of them.  This makes things look better... ideally you should be able to get similar results.  I'm able to.  When you perform better on a single thread you'll perform better everywhere.  I purposely make my tests harder... I feel that people have more to gain that way.  ... I'm the guy that tells you that you have food in your teeth.   :wink: -- some people don't like hearing the truth but I don't really care because I designed my tests to be useful first for myself.  An adjusted speed test that only aims to deliver the highest numbers is worthless for real technical troubleshooting.  Look at my TiP reading on the single thread test results, you can see the numbers are all in how you interpret them.  The final number output in your TMN results is unlike other speed tests because it's honest and unadjusted.  I don't show you the data through a funhouse mirror like the other guys do.
 
I highly recommend trying my multithread speed test.  Most people are able to achieve higher speeds with this test.  If you have a major difference between the two you should ask yourself why.  Also look at speed test mirrors if you haven't already.
 
First thing I'd do is hook directly into the modem, test... directly into the router, test... that right there will tell you a lot.  Remove variables from the equation, add them back one at a time and when the speed drops... you have your answer.  The thing that's good about TestMy.net is that I designed it to be used that way... I designed it to have less variables between the results and I give you all the control over how the test is run.  This way, you can run the same exact test and see differences in results.  With TMN you know that nothing else has changed except the variable you're testing, which makes the troubleshooting process much easier.  Other speed tests are designed to produce the largest number possible.  If I could upload the information that I know about other tests into everyone's brains, I'd have no competition.  Sad fact, most people choose the tests that return higher numbers.  Sounds like you're a smart one and you know better, make sure you tell your friends about TMN too.  I rely completely on people like you to get the word out for me.
 
Just because the router is two feet from the laptop doesn't mean you'll get optimal signal.  You can actually get worse signal right up on top of the unit.  It needs a little space to radiate before it becomes more effective.  I'd also look for other electronics near by... whether they broadcast a signal or not is totally irrelevant.  Don't dismiss anything, even a vacuum cleaner and alarm clock emit radiation... and they can be right along the right wave lengths too.  Try to also position away from walls and shelves and if possible put it in the most centrally located spot in your home.  Even if it's further away from your main area, your entire home will benefit.  Another thing people don't think of is placement in the corner of their homes.  Think about it this way, your router is spreading the signal in 360 degrees, if you put it in the corner of your house you're throwing only 90 degrees towards your home, 270 degrees is lost in the wrong direction.  3 times more signal is going away from your house... which makes your connection weaker.
 
Hopefully this helps, feel free to email me any time,
- Damon - TestMy.net

 

Hey Damon.  Sorry I've been meaning to reply and thank you for your very extensive response.  I've been on the road for work.  You know how it goes.  But thank you.  The single thread / multithreading tests explanation made things make sense a lot more.  The funny thing was I tested the day after my email to you (singlethread) and my speeds were up quite a bit.  It may be because I moved a bit away from the router like you suggested.  Regardless, I can't really complain about these types of speeds.  I certainly remember connecting with a dial up modem!  :)

 

Oh and the suggestion of testing right from the router and modem is wise and I wish I could.  At the moment I don't have an adapter to connect ethernet on my rMBP.  What do you use?

 

And just fyi I've turned many people on to your site.  Even a few installers that have been brainwashed by Comcast and the sites that, like you said, return the fastest speed.  They were appreciative.  Comcast, probably not so much!

 

Thanks again man.  You've always been very cool and responsive.  I appreciate it.  I'll do my best to get your name out there more...

Jack

 

Hi Jack,

 
Of course right after you say, "You've always been very cool and responsive." I go and take 3 weeks to email you back.  My sincere apologies.  Makes me wonder just how many people that I've left hanging.  I get so many emails that it's become impossible to keep track of them all.  I'm going through my inbox and noticing that this is happening way too much lately.  So, I've decided to build a support ticket system.  I think a system is needed to keep track of it all and make sure that every single message is responded to.
 
"Oh and the suggestion of testing right from the router and modem is wise and I wish I could.  At the moment I don't have an adapter to connect ethernet on my rMBP.  What do you use?"
 
... first, look on your modem for a USB connection.  I think that most new modems don't have USB anymore, worth a look.  If you don't mind spending $10-$20 check amazon for "10/100/1000 USB Ethernet Adapter" (or 10/100 for cheaper ones.) -- it can be a good thing to have but not worth it for everyone.  If you have the late 2013 rMBP or better then you have 802.11ac, theoretically 300 Mbps faster than gigabit ethernet (on the right network).  If you have an earlier rMBP with b/g/n wifi and do any large internal network file transfers I would definitely get an adapter and directly connect to the router for those really big network transfers. Or if your home network is still b/g/n it can be good to have until you upgrade to ac.  But if you don't find yourself waiting for things like that right now, it probably isn't worth it for you.
 
A free option may be to just have a friend stop by who has ethernet.  Make sure they have a decent laptop first and get a baseline on their computer to compare to.  If they can score much higher directly connected then you know that your router/wifi is the issue.  If the numbers are similar then you can hook things back up and more confidently proceed with testing and tweaking knowing that the wifi is performing to par with the wired setup.
 
... you probably fixed this already.  If so, let me know what you figured out.  I'll make sure that information gets to others who are in similar situations.
 
- Happy Testing!
- Damon - TestMy.net

Share this post


Link to post
Share on other sites