Sign in to follow this  
Followers 0
CA3LE

"your speed test is inaccurate"

2 posts in this topic

your speed test is inaccurate, it said i got 7mbps - 14mbps download speed, however i use usenet everyday and download at 40mbps

 

Sorry, long email... hopefully you read it since I wrote it just for you.  Your connection may be faster if you take the time.

 

Hi Ross,

 
Don't be so quick to judge.  This isn't my first rodeo.  I've been testing bandwidth since 1996... (the name TestMy.net came about in 2001) I've dedicated myself to the subject and I've built TestMy.net myself line by line.  It IS the most accurate and compatible in browser speed test available and it's able to detect issues that slow you down that other speed tests fail to notice.
 
Ross, I hear this nearly every day.  Stick with me and I'll explain a few things.
 
First of all, I too use Usenet, every day.  Giganews to be exact. I also am able to pull faster speeds with Usenet than I get from my TestMy.net results.  There are various reasons for this.  First, your speed on Usenet doesn't instantly ramp up to full speed.  TMN is calculating your speed based on everything that happened from start to finish.  ...that doesn't account for your speed being THAT much slower, I know, but keep that in mind.  
 
Second, Usenet is multithreading the download.  Some providers right now are actually shaping their users bandwidth, limiting single thread transactions.  TMN reflects this... Usenet and Torrents will be unaffected because they are threading.  There is also TCP stack optimizations that can make a HUGE difference.  It looks like you're running windows.  This is more often the case with windows and fixing it is as easy as running TCP Optimizer, it's free with no install... it simply optimizes your TCP stack.  Upon reboot most people have drastically improved speeds.  Thousands of users over the years have emailed me and posted in my forum about the success they have with that.  When I ran windows that was the first thing I did on a fresh install.  Every time it made a huge difference.  Trust me, if your results on TMN are effected, you're effected elsewhere too.  If TCP is the issue, you will find relief and feel a difference in how snappy things are afterwords.
 
Just because you can achieve 40 Mbps doesn't mean your optimized for that speed.  TestMy.net is VERY good at picking up on TCP issues and provider bandwidth shaping because it's a linear, single thread transaction.  It's the ONLY speed test that works the way it does.  I also offer a multithread speed test now.  I encourage you to try this method, it's brand new and the public is barely starting to even notice that I offer it.  Personally.... here, I'll just show you my speed.
 
Without Multithread enabled
 
:::.. Download Speed Test Result Details ..:::
Download Connection Speed:: 46403 Kbps or 46.4 Mbps 
Download Speed Test Size:: 80.8 MB or 82688 kB or 84672512 bytes | Timed:: 14.609 seconds
Download Binary File Transfer Speed:: 5800 kB/s or 5.8 MB/s
Tested At:: http://TestMy.net Version 13
TiP Measurement Summary:: Min 23.65 Mbps | Middle Avg 50.84 Mbps | Max 51.28 Mbps | 44% Variance
TiP Data Points:: 24.89 Mbps, 47 Mbps, 51.28 Mbps, 51.19 Mbps, 50.94 Mbps, 51.11 Mbps, 51.19 Mbps, 51.06 Mbps, 50.98 Mbps, 51.28 Mbps, 51.19 Mbps, 51.02 Mbps, 51.15 Mbps, 50.98 Mbps, 50.72 Mbps, 50.81 Mbps, 51.02 Mbps, 51.28 Mbps, 23.65 Mbps
Test Time:: 2013-06-26 08:09:47 Local Time 
Location:: Phoenix, AZ US >> Destination:: San Jose, CA US
1MB Download in 0.18 Seconds - 1GB Download in ~3 Minutes - 829X faster than 56K
This test of exactly 82688 kB took 14.609 seconds to complete
Running at 253% of hosts average (Cox Communications http://testmy.net/hoststats/cox_communications
User Agent:: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/27.0.1453.93 Safari/537.36 [!]
 
With Multithread enabled
 
:::.. Download Speed Test Result Details ..:::
Download Connection Speed:: 85954 Kbps or 86 Mbps 
Download Speed Test Size:: 80.7 MB or 82668 kB or 84651996 bytes | Timed:: 7.879 seconds
Download Binary File Transfer Speed:: 10744 kB/s or 10.7 MB/s
Tested At:: http://TestMy.net Version 13
Multithread Test Utilizing:: west2.testmy.net
Test Time:: 2013-06-26 08:08:38 Local Time 
Location:: Phoenix, AZ US >> Destination:: Global Multithread
1MB Download in 0.1 Seconds - 1GB Download in ~2 Minutes - 1535X faster than 56K
This test of exactly 82668 kB took 7.879 seconds to complete
Running at 474% of hosts average (Cox Communications http://testmy.net/hoststats/cox_communications
User Agent:: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_8_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/27.0.1453.93 Safari/537.36 [!]
 
Both are testing from the same server, same amount of information, moments apart but the test is invoked differently.  Both results are right... put a meter on the line and you'll see.  Some faster connections need to multithread to pull their full speed.  Now, if I optimized my TCP stack just right I would probably be able to get those numbers more in line with each other... I leave settings default on my machines because it aids my development.  Having the issue myself showed me that I needed a multithread test to give people to compare the classic results to. ... so I recently built and released it.  If I always rocked optimal settings myself then I wouldn't be able to test the way the majority of my visitors test.  I have optimal setups I test with too, achieving hundreds of Mbps. (commercial connections, not my home connection... I wish!)
 
You can multithread to one or all of my servers... giving you the ability to get a global result from a single test.  No other speed test is able to perform this way.  Another feature that really sets TMN apart.
 
I also see that you're in Scotland, International routes and distance will drop your speeds as well.  I recommend testing off my server in Amsterdam.  Compare the results to what you get to my servers in the US (since you probably visit many sites hosted in the US).  Again, TCP optimization is important.  Because you're adjusting the maximum size of the packets.  Imagine it like this.  You have a bucket and a thimble.  The thimble is what you're working with now... when you increase your MTU you'll be working with a bucket.  You don't have to go back and fourth to the well as many times with the bucket.  At short distance it's less noticeable but testing all the way across the Atlantic... your speed is going to suffer with that thimble.  What also comes into play is that many consumer level international routes can be limited. (this is getting better over the years... it used to be a more widespread problem)  Multithreading will raise the speed in that situation as long as the route isn't being limited by congestion.
 
I hope that you find the issue, resolve it and email me back.  Like I said, I get this a lot.  Nearly everyone emails me back to say, "Sorry... I'm going to tell my friends about TestMy.net now..."   :razz:  -- I hope that in the end you'll see that TMN reports the truth and share my site with your friends.  I work day and night to make it better for you.  I don't advertise so I rely on my work to speak for itself and let the public decide if it's worth sharing.  So far, so good.
 
Cheers,
- Damon - TestMy.net
 
P.S.
 
I realized I forgot to give you a link to the Amsterdam server http://eu.testmy.net will default you to Amsterdam, NL.  You can also make a selection at http://testmy.net/mirror ... a link to that is at the top of all the pages.  See attachement.
 
post-2-0-11203500-1374961470_thumb.png

I recommend the official servers because I control them but if you look under the self hosted tab you'll see a server in Huerth, NRW Germany and London, GB.  You might want to try those too.  Any mirror listed is capable of at least 100 Mbps, my official servers are all connected with a minimum of 1000 Mbps... my main server in Dallas has dual 1000 Mbps uplinks to the Internet. ... all of my servers are connected through the Softlayer Network which has some of the deepest peering in the industry and a worldwide private network.  The cities where I host my servers are chosen based on their website hosting popularity.  My site in Texas hosts well over 100,000 servers each of which can have countless websites... that's not to mention all the other hosts in Dallas, that's JUST Softlayer's servers.  You're going to connect to the areas where my servers are if you're on the Internet... so those are the areas I have you test to.  Other speed tests try to eliminate routing factors to boost your score... I believe those are important variables.  You already know the speed you're supposed to get... I'm trying to show you the speed that you REALLY get.
 
Cheers,
- D

 

Gio Bacareza and iceb like this

Share this post


Link to post
Share on other sites

Update: Here's my updated scores.  Updated and optimized network configuration and different provider in a different city.  The computer and it's configuration were untouched from the first set of tests, still default OSX settings.

 

Multithread Enabled

 

:::.. Download Speed Test Result Details ..:::
Download Connection Speed:: 107267 Kbps or 107.3 Mbps 
Download Speed Test Size:: 200.9 MB or 205692 kB or 210628710 bytes
Download Binary File Transfer Speed:: 13408 kB/s or 13.4 MB/s
Tested At:: http://TestMy.net Version 13
Multithread Test Utilizing:: google.testmy.net
Test Time:: 2014-08-07 11:33:45 Local Time 
Client Location:: Colorado Springs, CO US http://testmy.net/city/colorado_springs_
Target:: Global Multithread http://testmy.net/multithread
Client Host:: Comcast Cable http://testmy.net/hoststats/comcast_cable
Compare:: 16% slower than client avg, 271% faster than host avg, 240% faster than city avg, 484% faster than country avg, 929% faster than world index 
1MB Download in 0.08 Seconds - 1GB Download in ~82 Seconds - 1915X faster than 56K
This test of exactly 205692 kB took 15.709 seconds to complete
User Agent:: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.77.4 (KHTML, like Gecko) Version/7.0.5 Safari/537.77.4 [!]
 
Classic Download Test
 
:::.. Download Speed Test Result Details ..:::
Download Connection Speed:: 103437 Kbps or 103.4 Mbps 
Download Speed Test Size:: 200 MB or 204800 kB or 209715200 bytes
Download Binary File Transfer Speed:: 12930 kB/s or 12.9 MB/s
Tested At:: http://TestMy.net Version 13
TiP Measurement Summary:: Min 41.45 Mbps | Middle Avg 111.35 Mbps | Max 113.96 Mbps | 5% Variance
TiP Data Points:: 88.15 Mbps, 112.01 Mbps, 112.68 Mbps, 113.79 Mbps, 113.96 Mbps, 113.36 Mbps, 110.05 Mbps, 109.73 Mbps, 112.85 Mbps, 113.44 Mbps, 111.93 Mbps, 107.55 Mbps, 107.93 Mbps, 105.66 Mbps, 106.79 Mbps, 113.87 Mbps, 113.79 Mbps, 113.53 Mbps, 41.45 Mbps
Test Time:: 2014-08-07 11:36:04 Local Time 
Client Location:: Colorado Springs, CO US http://testmy.net/city/colorado_springs_
Target:: Dallas, TX US http://dallas.testmy.net
Client Host:: Comcast Cable http://testmy.net/hoststats/comcast_cable
Compare:: 29% faster than client avg, 251% faster than host avg, 219% faster than city avg, 423% faster than country avg, 885% faster than world index 
1MB Download in 0.08 Seconds - 1GB Download in ~82 Seconds - 1847X faster than 56K
This test of exactly 204800 kB took 16.226 seconds to complete
User Agent:: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_4) AppleWebKit/537.77.4 (KHTML, like Gecko) Version/7.0.5 Safari/537.77.4 [!]

 

That's what you want to see, both numbers in line with each other.  This not only shows that your Internet is performing well... your computer has to be running right also.  I have devices that simply can't perform as well in multithread... most likely being held back by hardware limitations, that test is much more hardware intensive.

 

When your computer and network are running as well as mine do online services just work right.  Although, sometimes I'll still have an issue with a streaming service, run a test here and helps illustrate that it's not affecting my whole connection.  Also, it helps having a baseline speed test on a great connection and computer.  Then when you test slower on a different computer on the same network you know that it's not a fault of the connection.  Could still be a wifi issue, that's why you test wired if you can.

mudmanc4 likes this

Share this post


Link to post
Share on other sites
Sign in to follow this  
Followers 0